<%BANNER%>

The Role of Birds and Microsites in the Regeneration of South-Temperate Rainforest

xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101122_AAAABL INGEST_TIME 2010-11-23T02:21:12Z PACKAGE UFE0012043_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 8423998 DFID F20101122_AAASOO ORIGIN DEPOSITOR PATH milleson_m_Page_030.tif GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
7daa82a3455b2ae410ca031a46db1f48
SHA-1
456d52545c62cd2fc1bd4a88bd14db3eb2ad282a
1053954 F20101122_AAASPC milleson_m_Page_045.tif
33bd92fffcad6530de342ec3064b8cd1
fb3359db8ce0fc7c2d312a629488ef42da98dea8
F20101122_AAASOP milleson_m_Page_031.tif
0ce6d59f65d33d4ac918fca51746bacb
dd26d7ae7b1f3c135b158f8fccc1fb27c3702a27
F20101122_AAASPD milleson_m_Page_047.tif
24c95d03992dfa629c2a958f84193b40
5b525a3e9e3ad72d0b6678ec9b3a87ea42dfd334
F20101122_AAASOQ milleson_m_Page_032.tif
48df3813fed4c8f81d0dd4facc562c80
9a498433811629af717044106626838be821d8e7
F20101122_AAASPE milleson_m_Page_048.tif
f366728007f0b010553937d8b2754ad5
56323d95745f6cce7c3d1ef6334f5e3558f384ee
F20101122_AAASOR milleson_m_Page_033.tif
addce434aa028f2d101e8c37ec9c6b44
08abd42b774da171d03f84457b1e16cb7f11ddb1
F20101122_AAASPF milleson_m_Page_049.tif
5c863f99263960fe9594e86ebfb0ccee
bea011fafde7444f9ada3bf6bd4a2fcacd5d62d8
25271604 F20101122_AAASOS milleson_m_Page_034.tif
fc1828ed7183b214867a68aa7bb4e46b
488ddedf4767e0081b1b0c89567d97dd8ee5f32b
F20101122_AAASPG milleson_m_Page_050.tif
f96ba5bb4eec9ea80a3761dd48265e8a
d6277ed756c79a4dfe5c89bb2c3e69b55b49f583
F20101122_AAASOT milleson_m_Page_035.tif
11ce4dfa4b8e6e4584b6ddc905717b49
7a2e4ce1b0016d9da6d588efab2f5d103219c1e1
F20101122_AAASPH milleson_m_Page_051.tif
f216cdbbe82503c60c4fd5b81918d8db
0ef2876fd8b3fe739967e926129647995b7ddef9
F20101122_AAASOU milleson_m_Page_036.tif
5dee16902b0b1330ab3290849177df21
9cdbb3b13b35e723b69fb9004b3584bf2b6aab82
F20101122_AAASPI milleson_m_Page_052.tif
060cf378b1649001eef8524976275160
15947a146d0e0de56e1a059efb9c931bd12f09c4
F20101122_AAASOV milleson_m_Page_037.tif
ea8d9a349a561902e2e70454b0ca2e96
85268d6e72243a7bb418094bc440c9c73348f883
F20101122_AAASPJ milleson_m_Page_053.tif
8566911f000912c00cf0c27d4712862f
c662485a32bab573afc1b4420a5d294e6e811dfb
F20101122_AAASOW milleson_m_Page_038.tif
a2fbd5b3a0123cf96497ab5b171271f3
586bc3e417194c81cb6588c53b10e6ff93506ffd
F20101122_AAASPK milleson_m_Page_054.tif
afa9ebe26f3bb21008df565b9fa603b7
d9923e4fe513585b69d898ab4f000eef1b1c5ab0
F20101122_AAASOX milleson_m_Page_039.tif
1858f44420de15e7bb7b8923241fa171
5c541b97e61ec9b6feb36d8810609350acc65382
F20101122_AAASQA milleson_m_Page_070.tif
f557d3a815dc9705696e134cf0eba49d
6b46665b40c4429e9bde6f1387f86494d3e106d4
F20101122_AAASPL milleson_m_Page_055.tif
ccbdedfa9b3a46288d8be50ef30d74df
fb8f757b7f2a9a3b7419338d9391e2d5a4161b9b
F20101122_AAASOY milleson_m_Page_040.tif
480151f5d3505be8c5bef5bd60739861
352aab4870bf89d756c3b0104158874f1eee0f2c
F20101122_AAASQB milleson_m_Page_071.tif
d7213a5d0600714ef0bf1305136d92e8
77870f7c12fb44893ab141a21292309504e788ff
F20101122_AAASPM milleson_m_Page_056.tif
22609d523d177694fb499f1b7b0f2efa
7a71cb5e428d8abffd8766cfc0b1883945b9b190
F20101122_AAASOZ milleson_m_Page_041.tif
f9dc3d7f9f2466c074da0d78bad43b6b
3c045708deccc9aba31ba7bf017f8ad7dc0eca50
F20101122_AAASPN milleson_m_Page_057.tif
27d310656e15f96801928b3be37d5016
ce9bda848bfb77ca7949fcbfcfc0c8504941686f
F20101122_AAASQC milleson_m_Page_072.tif
c89eb998493efd6f117bbb8834984c21
2bc2070455f3a262185bf5156863a76121915d65
F20101122_AAASPO milleson_m_Page_058.tif
5d76f190e12ec5f039bc3757cabf7d4f
f9de2fedc05ee198f53c56f1e1970c589093fcb8
F20101122_AAASQD milleson_m_Page_073.tif
09a78ad624e8db87e63e1eea45d78bd2
4fd7b5756ef074a2c6180ae5a1922b5876810ef1
F20101122_AAASPP milleson_m_Page_059.tif
1a3fc80b65c976d0d4d9aec07975dc48
bcef27bf3fd4c9f2e91d886cfaff9d37887c31bc
F20101122_AAASQE milleson_m_Page_074.tif
68be37ce57409bca63e1338817250216
e6b008fd2af9b9a496c5ce95e453feb29c316b8f
F20101122_AAASPQ milleson_m_Page_060.tif
aa00cf8c63282a8550002215ca807257
ad873d44ced4002a01e8f57e79f58205b05a7a97
F20101122_AAASQF milleson_m_Page_075.tif
2a64e2a54ed2e38e83a3eed350ee01e3
7fd65d6376455caecfe182ec12ea4a11c7b32fd9
F20101122_AAASPR milleson_m_Page_061.tif
2a102d387d17a272f2dd9b5f5434d64f
116d3353768a327ab52e5eb9699fa25d25ea45dc
F20101122_AAASQG milleson_m_Page_076.tif
edf612219cbd821c5c12b97bd90639bd
4e9dfe6405c2c4aa56a2d5c6b5be7109b01dd444
F20101122_AAASPS milleson_m_Page_062.tif
8f074cf12adb88b054bfe35a66d580da
ddc395e58cdc57bc393149826bf5a09dc8742211
F20101122_AAASQH milleson_m_Page_077.tif
ef94c44ff484046770d6a2a7499e8f4d
562c68893bc8f8b7943ed81abcf39c6fc80271a4
F20101122_AAASPT milleson_m_Page_063.tif
b7bef49aceab0a47d300289787e45b3d
f98098e504dcf2b396ebd917fd897bc67fb7098f
F20101122_AAASQI milleson_m_Page_078.tif
aad35a566cbc7fdb9d1e48c6b9e46355
89b3fe58f5fc1c90e0163a4faf7ebbf7e1ab4177
F20101122_AAASPU milleson_m_Page_064.tif
b1973f88c5058574e01b7ea23ec06e80
5c560d158630561fd04e6dacc2142f6240d79ce9
F20101122_AAASQJ milleson_m_Page_079.tif
f4c309ec1af978b2539072d9386f1852
9791c2611ddfc5011c71458b32aa3c5f13169c8e
F20101122_AAASPV milleson_m_Page_065.tif
c8771268d65ab4997d7ba5b6d407b0bd
4ff6ed092678864641d3c007a88c7c3f71799415
F20101122_AAASQK milleson_m_Page_081.tif
32506ef1f047222d65842bea531697c2
aa30f999465c3f2421af693b4d52dbab16633d5c
F20101122_AAASPW milleson_m_Page_066.tif
08e7d88b9ec7741da431afe1e2d4fddf
4d2f8762777720a9ad65d42b83b8a8302f7a1f1e
F20101122_AAASRA milleson_m_Page_100.tif
b828bc4d32be96a2d1adcfb0a3ddeddb
3ec776c0f67e027784e9ee538c22136053b4cffd
F20101122_AAASQL milleson_m_Page_083.tif
03fd5517b5d053363e55b8c7ec10834c
9d1cafe07bdc75007423ce563ed18d8ab27d95b7
F20101122_AAASPX milleson_m_Page_067.tif
15001f89ce0802680ae1b3be69f4c551
cdb11b5d8a779b7c34c50043a3bcf466067f1b70
F20101122_AAASRB milleson_m_Page_101.tif
37c5187d6d3a8844d627d8197b851aee
c4266d030362bd1648765d2ae8dcb7563f63d34f
F20101122_AAASQM milleson_m_Page_084.tif
2dd11d26b89ae577b28c298c368cacea
275689737d5d6304b048a4cd52680226feeb1aed
F20101122_AAASPY milleson_m_Page_068.tif
27c663d7228a5d06dfed735a04fb7fd0
747524388711a64d048307b48ed9dd091435acae
F20101122_AAASRC milleson_m_Page_102.tif
cd2e02f26dde4c1423e8e4bde53e52f5
eb7e2713bac5e95c79f61f83346135577655b44f
F20101122_AAASQN milleson_m_Page_085.tif
b297dbee33917bcefc22563fc0f7bbca
c6573cc510116a809f5225288449a758105b51af
F20101122_AAASPZ milleson_m_Page_069.tif
c7909b4b0494046f3eca5d0d6c541af3
cc99492b65000461570da8d1bd31f85de9f5803a
F20101122_AAASQO milleson_m_Page_086.tif
28b92202b3b978f204ac780d06da7d6e
d94b386887e90d0d20747b17e9fe0181031e1d7e
F20101122_AAASRD milleson_m_Page_104.tif
e721008ec7ade80fdb8b5bebe2af66dd
315ac87db1fa6be1d1ae39602a5cb0f75e14a6b1
F20101122_AAASQP milleson_m_Page_087.tif
f82345e436b5bb338d1261c73453f566
255705e5104ce593c8ac60ae3c878ef5e97bfab3
F20101122_AAASRE milleson_m_Page_105.tif
fbd05b99d49dccae8ab179cb531404d4
9d8480afdb02de6c2140cb8c6bb0aff0ffe96e31
F20101122_AAASQQ milleson_m_Page_088.tif
eeb16e347e53800d152cbc8b1a883b6d
b8bb20ce223378b1966cb29bf44a978ecc46a9f8
F20101122_AAASRF milleson_m_Page_106.tif
4036e8ca2257e62f288e60f90b66ebe8
5934210cf21c68f98ffd38b0dd940b283b2935d7
F20101122_AAASQR milleson_m_Page_089.tif
d7a68f829e2cb7dac8c2bddeddbb9599
e8bfe16d739420a97cf06789674ba9c0d72747f2
8039 F20101122_AAASRG milleson_m_Page_001.pro
7c4a18927d1c80e77a0ea9cefc0cf037
7ec9729476dc8c57c5b49e894f1d1df676cb705c
F20101122_AAASQS milleson_m_Page_090.tif
d0eea1be850f6c6595e41f0565beafca
74b58f02d324845b6e9f7cbebb660bc14e1189ea
1280 F20101122_AAASRH milleson_m_Page_002.pro
077ca571884190df0efe91e678ea7f58
045f5297b41e4e7b23762fdb9f10da78f03c8391
F20101122_AAASQT milleson_m_Page_091.tif
cf3944280cd32ba3548ceff6bc499c54
448ce941b54051138138a710c02d8bc0b15b2510
10659 F20101122_AAASRI milleson_m_Page_003.pro
12364e9e29fd3e009d15f18dbba544f1
b2a8f8246dc7fbe201372024195126ef5ec0077a
F20101122_AAASQU milleson_m_Page_092.tif
0ac721dd428e404580866078d03f4738
71f437bf01e825fc927c1475973775116f6119c2
78520 F20101122_AAASRJ milleson_m_Page_004.pro
b890264da2b2d6aee82ed675eab9923f
fcd14f43be678b2336e4ee296bf5ffbcd30b6779
F20101122_AAASQV milleson_m_Page_095.tif
a1e8817c7b288d4f26ac77b55b916af4
de95ae6c33723eb7e2573a0aa1d3e653e0a4e152
110950 F20101122_AAASRK milleson_m_Page_005.pro
9baaa0cb4acca658c274850e013f5cf8
c323f64144c1d4c56b102fbdabdadfcee306752e
F20101122_AAASQW milleson_m_Page_096.tif
33022a72ebfec5821ad316d6756b8f53
1167b03ea3399945cf396e2a32a38e6fe166c223
24962 F20101122_AAASRL milleson_m_Page_006.pro
8cc694200cb30d3bd51b790bc205323a
4af2c83f63f52ad686da6c55b6e459025486f210
F20101122_AAASQX milleson_m_Page_097.tif
73dd91c9475ed36a90b7d6c5838a5106
e0770256ffca8cef2d94cf14f97cfca8c95c7b8c
51984 F20101122_AAASSA milleson_m_Page_021.pro
bf8cdd0e261bf4d71e2c1fbf324ff38b
b541042883be4c8f24a670b7ba386650989ff043
15463 F20101122_AAASRM milleson_m_Page_007.pro
2f0b35c902954c8b0c0006930aa1e6c9
a65ab4bf9943aae06bf051a30c5913242c7ef863
F20101122_AAASQY milleson_m_Page_098.tif
90c01411cdc3dc488f92610d892b8f34
7d7c39a11318005f9367dbba4cf15e0049047b2c
51720 F20101122_AAASSB milleson_m_Page_022.pro
d455c848980fb61a08e492254d4b836b
81d65281fdf543213d58e2dade171989b82bd569
63480 F20101122_AAASRN milleson_m_Page_008.pro
96f311bc6febb58f21356f8c742146b8
f3b29232a8406f8c6ed3a45f7e72c473a22d9de2
F20101122_AAASQZ milleson_m_Page_099.tif
74843f0c0f5e215362945fb08b7299b9
71b867985e1eb2fe1fe5a76ba0256c2ceee99ed2
50573 F20101122_AAASSC milleson_m_Page_023.pro
a50b1dd64410ebfee2962d60094433bf
e5b5392734ef714e1cfb1e351dca6bbac4b24bad
45121 F20101122_AAASSD milleson_m_Page_024.pro
399602537e3649873dc181211d831296
31b2abfe0a49a789a740a36974fccc372e8586bf
82796 F20101122_AAASRO milleson_m_Page_009.pro
6ddc3905e090c26debaabb23bf376066
284052fd7b08adfc59a226b6291cf3767eba2733
34515 F20101122_AAASRP milleson_m_Page_010.pro
802f63ed93e608800dbb72678e5c7fd6
d2f817f04d0c1332d34d7f07fbe27ca541fd4df3
51443 F20101122_AAASSE milleson_m_Page_025.pro
c2708493dbf08304f75c79277c0304fc
5dcc15bea7e98590b9ec98fee9d70f9a5afb54ef
38447 F20101122_AAASRQ milleson_m_Page_011.pro
942bc3d9df7a4d3d7fae611908eebc6d
c4ace9feb93265eefa2a5829831803a11789e4cb
46642 F20101122_AAASSF milleson_m_Page_026.pro
9e96b40ff07e6da7fbb8ad3481ea2a37
6de1e4dcf9d44920b92e9f9d6555db7a676d9a99
46503 F20101122_AAASRR milleson_m_Page_012.pro
73c2b7acd2930690cfcaf2df9196d68e
aef7d5773be7dc93f7df5aa706bc1abe25b7e058
50605 F20101122_AAASSG milleson_m_Page_027.pro
f3e777bd2ae5de42d95197ee05058134
808a97e26e0759acae62d21f812e56fd8fd2d21e
43099 F20101122_AAASRS milleson_m_Page_013.pro
d812019c7b217b3ec0c25b04cdb7b826
81eb91846909dc599c8d7edfe849d0579eb4a6fb
50366 F20101122_AAASSH milleson_m_Page_028.pro
33f8465b3c7b134c40320399e87c2e7e
ba66e90a2565cfde3a5dbd0838879fe5342e8881
52819 F20101122_AAASRT milleson_m_Page_014.pro
a4fbcdf8cd4db851bd4a21e4411689fc
6093e4efb9f6311da1f1751eafe7e54e51ede368
44971 F20101122_AAASSI milleson_m_Page_029.pro
d55ae684977129c95d911d4da0ea8a7e
fb2ed5a0868afdba123b46eb225841b3aecda8e7
51491 F20101122_AAASRU milleson_m_Page_015.pro
ca681684107bff0bbe2da115b024eb53
38d3cd14a706ede598eefd176476b495ba3e07b5
25618 F20101122_AAASSJ milleson_m_Page_030.pro
094345ffe463d1ee665252952090be9a
34638d8af4d732f072d9a282eddf4278963aefab
50838 F20101122_AAASRV milleson_m_Page_016.pro
7ffa45e1baedb8cb2db28372167e85ae
a8b8e929ae77b7aeb8e0ac6a6626346030e081dd
30853 F20101122_AAASSK milleson_m_Page_031.pro
cbb13a0680677f6c7968684d9a728e83
a729edbe29139389e773ba6d2bd5706e40f4a877
49805 F20101122_AAASRW milleson_m_Page_017.pro
5e66a1b84b41807351f19fbaa201077b
e4c623dd9d4e680b49cd7964cc348e3eb54e7462
46965 F20101122_AAASTA milleson_m_Page_047.pro
c8fc7abe915319f8b36590f7b382d14c
9e1f5bcf0d7981a805ec424b3b0584fc1d876ab2
25458 F20101122_AAASSL milleson_m_Page_032.pro
f129e9bef93274e73778ecb144117e30
fcf7a4eabd5a59baf0028d9c44b48d4c716be394
51814 F20101122_AAASRX milleson_m_Page_018.pro
a2770e59bb80c5eb6e679d61d8e04a5a
c311b27cb23d2328ef4fac3e27bc93e090f29aa4
49742 F20101122_AAASTB milleson_m_Page_048.pro
7132734825741034e4dd270f30f08dd2
ddd5538394d94cfd47ab74ae91c30fb9f3eb03e6
27547 F20101122_AAASSM milleson_m_Page_033.pro
80fc3a9559ac10b3639e5367af416a59
af0329ad4b772bfa22918e7c79ab2e9019e8ef1b
14391 F20101122_AAASRY milleson_m_Page_019.pro
3085c2a5ca938ffa491e3ae6c94e20d7
90ec98c9b0438dcf02f1281ed68e7b79b660c538
46987 F20101122_AAASTC milleson_m_Page_049.pro
22db684d3d1c71330522584e61cbca2f
cbeefeac2b5829c1ee0de31462efc01fbfd0d6e5
32965 F20101122_AAASSN milleson_m_Page_034.pro
774735d7de36cc8ae5ae7e75436ba393
0502f07ccb22207ce1cbd8bd98ffcb387007ea94
53833 F20101122_AAASRZ milleson_m_Page_020.pro
56b25325978fb1d563fe96d247c9ea6f
d8a435579b41cfbf92b3f9827b45fc60ad693a47
48413 F20101122_AAASTD milleson_m_Page_050.pro
c8efd074eda5150f4e8b38e56c3bf98e
31aeca795febb7dcef2bedbece8235dc85d24fff
35674 F20101122_AAASSO milleson_m_Page_035.pro
cd7670e43c7bfd5c8ac6f98f680bf998
6e665f8799de6eac376f9e8827a69bba03726ba4
52888 F20101122_AAASTE milleson_m_Page_051.pro
30699b56db34c1e784bd917e86b29986
85441f89d33f7ef8f1aef2a2df7556325090a7a6
50846 F20101122_AAASSP milleson_m_Page_036.pro
dbfbb0ec1cccd8c69df34ec7d22a2187
24599c908ff403395a971bcc0f96e14e58495c53
33515 F20101122_AAASSQ milleson_m_Page_037.pro
e7d5c6e50bebcf22132b50aafbca8cca
ae0375c94c06c88f7c4a1dac59313265c6606252
51820 F20101122_AAASTF milleson_m_Page_052.pro
d3735ecacc91ecc60f808b032e0247de
b3c8d14159a3a02ea42be982bbef7156e62bf72b
52509 F20101122_AAASSR milleson_m_Page_038.pro
5b897c9ef6930c5afc82038cbddd224e
73b8d86d5a80010a61e31feda7750259364bfaf8
48556 F20101122_AAASTG milleson_m_Page_053.pro
b037ff7edc981e186f07c385842f8177
8227e2e89b215e21d7410a1f5a7a9a19bbea3255
41558 F20101122_AAASSS milleson_m_Page_039.pro
e67cab10e0d87059622cf6c14b195e54
8fcea874fbe1e10bd61e9f02b0c6e4600850eca3
46165 F20101122_AAASTH milleson_m_Page_054.pro
05a98f77f6a3452f24dadf988bcf42ad
a20bd0b78d8fa7165597dadcd2570cf6769a6c1f
42273 F20101122_AAASST milleson_m_Page_040.pro
dee4ee10d1215dd5cd1e76fe2320db0c
a0af0b9c5d7ea31c17c2844af1001f791a9cfe41
45009 F20101122_AAASTI milleson_m_Page_056.pro
1ab1817d8cf5916b75576d2e69402246
84f9ec87b662d2fb47661e1309366a95079a2f92
50356 F20101122_AAASSU milleson_m_Page_041.pro
fe7c0365c9dbbbae5a7aa39bb60e5ad6
ea9d58a85a2e6714fe06da3be65d08f07684079b
27704 F20101122_AAASTJ milleson_m_Page_057.pro
3e5109b9f48d2f4476370f96463e8601
472018efbf605c27fd68246d860a79b1dfa4d8ff
52659 F20101122_AAASSV milleson_m_Page_042.pro
dc87544ba4687e51f4353eb1ac50be06
0b6ed902089dbcee606596638da00648aa4c1b60
36762 F20101122_AAASTK milleson_m_Page_058.pro
93f74d92deba8f49f2a530af404e56a7
8e4f741fde0c7d7138c3096466c06e39e71ff847
51675 F20101122_AAASSW milleson_m_Page_043.pro
7995a98d586da83c9e26892ff05ccc47
27d889417d282a8797148d133999d5c958011a4e
28325 F20101122_AAASTL milleson_m_Page_059.pro
146e00a79840e511140477de7f9a9e46
d0b8e75c269b1f832ca702daa36e815dc09708d1
51471 F20101122_AAASSX milleson_m_Page_044.pro
86282e7887838a3d6a62afec3e1be85e
278a53b0ac6a7e43e13986ad63c65386093ab93f
52620 F20101122_AAASUA milleson_m_Page_076.pro
0d715afae9622f5c3e02346000dac5d3
821135675d3252e681e652cc003cacb856759432
42295 F20101122_AAASTM milleson_m_Page_060.pro
377805c5348b2e7773b8a4a891f62809
6542b9f6fec0fcaf22cbb3cf5479171e077fd78a
49860 F20101122_AAASSY milleson_m_Page_045.pro
35b1beca44ec934b1896898df92476ec
95c78fa9504497688fb5ad60e3c2702ac55ee8c0
43675 F20101122_AAASUB milleson_m_Page_077.pro
7e20deba7b6bcc12403e8bffffade1f6
157e544fb1bc33367ed29bde9f84eac39b45d4ac
30296 F20101122_AAASTN milleson_m_Page_061.pro
2610c8a7f673aecd590306c86063d418
98b63c7aac71ba0c002bb1f6ee8191ce41a616a0
21471 F20101122_AAASSZ milleson_m_Page_046.pro
43ef253db255ef7e5c41854248bc812d
3208e69300eb1ac259574ed4dcaf48de1165c613
37539 F20101122_AAASUC milleson_m_Page_079.pro
9e72ed8281a627ea444c78a0d197774d
acfa6834374dacadb8766560eac5de8b8941ea86
34057 F20101122_AAASTO milleson_m_Page_062.pro
1eba7d11f7743017f8cd52229e7851e3
8ba7e16a5759cbf6d4008b082d81b963016ee40e
48124 F20101122_AAASUD milleson_m_Page_080.pro
946f89ab3054c9643d1ed609d3e3cd2b
2ff73de79953c02e3061d5ad62ce4789619067da
17274 F20101122_AAASTP milleson_m_Page_063.pro
8e27b2231e8a59d5d55029d09d972301
f7572feb78ba8f198fd3a63d67c7e338d8128700
37605 F20101122_AAASUE milleson_m_Page_082.pro
b0e6fe2a77e3172dc9ce28281feba281
f4f80627fd2d931f382d49996a074b39e636e0df
17099 F20101122_AAASTQ milleson_m_Page_064.pro
71ce4a812fa6327e808d42705746dc7c
98ec18aa71f92ab3a2587343388fb3b318c63fc7
3173 F20101122_AAASUF milleson_m_Page_083.pro
6f6c99856fcecab147a8c8eb1df6fe90
c2c526ee6f94b60c6ea80f78069432d0635c0591
36395 F20101122_AAASTR milleson_m_Page_065.pro
f070cee88d7f5b7bc651600296910bea
b285dfcc29db9975d1f7e0608075cc7fca1d2da7
23780 F20101122_AAATAA milleson_m_Page_021.QC.jpg
aacee3ea0436a54bf778e5db7bbbca3c
5981546c6eec0feaed579a636862d991eb332a01
38248 F20101122_AAASTS milleson_m_Page_066.pro
909a6420ff1a046e7bf366362429b75e
87fcb2891082e33e25d51ea93b74c0b2c1acf191
6572 F20101122_AAATAB milleson_m_Page_021thm.jpg
d310ae9d7fca42a6b2b79f8b8c3fdd40
f3d2d2e2a76da04ff035fcd2f254968ac858df9a
5101 F20101122_AAASUG milleson_m_Page_084.pro
8cdc79abe4adc5264a0b63fa10f8af30
ee8646f5f1be1d39903994b049dac8c9b47f43bd
28802 F20101122_AAASTT milleson_m_Page_068.pro
f1b69aecdae2e9aa22cb27049fd896c9
18d9a5e531d636549c2747033439642a03f0d332
23501 F20101122_AAATAC milleson_m_Page_022.QC.jpg
868ab4b68bf1c2ce6fbb1727253c7a4f
829146bb1927ee3857edbadbcfbeaa3994994583
334 F20101122_AAASUH milleson_m_Page_086.pro
7b974c741c516b94fd1c96cc33010fe9
b6ada18f3fe5637bb49c1539c49c78fede92eff1
36500 F20101122_AAASTU milleson_m_Page_069.pro
daeca694205a4c1569ce8f814fe6a815
33671ca6179d620212c372db987a69141f834b7b
6460 F20101122_AAATAD milleson_m_Page_022thm.jpg
e4bb15eb4a56e609cf8ce2295db4a841
899692c07c57c384bff683c6c8f1704f36bbaad9
42173 F20101122_AAASUI milleson_m_Page_088.pro
87f10dc8377fcad2c7ece4e0933dbc2d
b47fab9afcb2eb1487d3382d27b519bd1f8f7775
35292 F20101122_AAASTV milleson_m_Page_070.pro
c9ee8558c76c9dff77dc4540f1470752
e93583a245649dceccef84cbec17bf7b51e56ac9
23246 F20101122_AAATAE milleson_m_Page_023.QC.jpg
00b0f219a9c21788a38035cab38393a0
1cad7050f1ebaa2406aa1e8c8a80a801dcb6ab9b
37590 F20101122_AAASUJ milleson_m_Page_089.pro
cb8d8b88017934d92bb560209246c1fd
aea6507ccfb847829ee9d0aeafe94bb8066f81ff
20802 F20101122_AAASTW milleson_m_Page_071.pro
6d1b4209424825b3416af8ca82341079
71b819c7634ad67e25aed4168ef336ad964e4dfc
6508 F20101122_AAATAF milleson_m_Page_023thm.jpg
e8eee91679d6c498a44fe6c69d9f8d54
ffa9f9e377bf61477c902fc7715e39cfb49a51ca
45206 F20101122_AAASUK milleson_m_Page_092.pro
038bbd86d87c588217daf993ed6ffc10
5d86f6ee75d43b98cc565f667408bea1c507e971
45290 F20101122_AAASTX milleson_m_Page_072.pro
606018ba49deee69260aed6d90313167
73a44f1463c1151bd5645865e8d940de68654bf3
22041 F20101122_AAATAG milleson_m_Page_024.QC.jpg
5d6274fdb3564251f50788db26a10892
179371043f81590216b96440e0ed10d6106cdfad
472 F20101122_AAASVA milleson_m_Page_003.txt
369588a98d7ef48d6f44b7f5ec069b4d
b4d39be1691997a7ef561fa122f1910d8474a4a1
36558 F20101122_AAASUL milleson_m_Page_093.pro
1a60f7d93d8a694e5111fb4a26646c48
f236fa5fdc14e6695c6aa3b72c657bce157136a2
40374 F20101122_AAASTY milleson_m_Page_073.pro
fd360f2f570fffa9749b7940a7a8fbb2
d003a69a7942d2cfc4d805e96f966c6364607ebc
6099 F20101122_AAATAH milleson_m_Page_024thm.jpg
023da22c634e6b4433fe8355df981112
7e8412c65dd1d3a94041272422bb268b53d16fee
3319 F20101122_AAASVB milleson_m_Page_004.txt
c1f99faabc23e4ba8bd2039a76ec1ef0
f2566fee8cd5da753670f32d493127e566afef92
59106 F20101122_AAASUM milleson_m_Page_094.pro
3cb8e5226837ee860fd4a39617adb2bc
35c1d377e99ada9bbb47d698a78aa9d679c6f85f
54129 F20101122_AAASTZ milleson_m_Page_074.pro
600cb28e95afab00e912c1e599108ca0
5474dda396f1394fc23229b866c6d845e658931e
23928 F20101122_AAATAI milleson_m_Page_025.QC.jpg
1cd46e7195717ad41a1835b497d9769b
d1768a5232c746b49203439e56cc6bec6509f9fa
1036 F20101122_AAASVC milleson_m_Page_006.txt
808fde93d61aecfc60137eaf5a50a66e
d88c19cb0bcb45a796a4ce7deb0a6b181a3e7f51
35494 F20101122_AAASUN milleson_m_Page_095.pro
4dd3da3de939ca1bd143cbba5550e5a4
c30c199607546c07e63501a2ca48d8374757793c
21513 F20101122_AAATAJ milleson_m_Page_026.QC.jpg
e9b57f34288c0ab6fd97f87c1ada9a65
a9c23737a29bce228b989b2da6e87e8ad779d472
710 F20101122_AAASVD milleson_m_Page_007.txt
895aab94c8f29139ef84b0f1691ac25c
b955204c76c24a4f12ef0b9cab0070af646689be
56647 F20101122_AAASUO milleson_m_Page_096.pro
295b081968bfd9f1ef2178d5d739c85d
3c11ea1248136752a846eac309e5cd0fe01e0efa
6276 F20101122_AAATAK milleson_m_Page_026thm.jpg
27045214c90f42c03cdc54804c5fd922
2718e583296648a693048a696cb6cc43c7a45ff6
2596 F20101122_AAASVE milleson_m_Page_008.txt
630333ba39319cca8200cda44bdf296f
56a584ff787ce14263dd328806791fea5ac20bd2
64189 F20101122_AAASUP milleson_m_Page_098.pro
ab68124f52323a5f5e5d473c0599a000
41e0d338b66e001d47d28a194a12803eeceba0a0
23349 F20101122_AAATAL milleson_m_Page_027.QC.jpg
969f16dfade0c4327a4fc778566736e2
3ff79f7c422dec7234fb7e1c0a51ade6c27ae7a6
3350 F20101122_AAASVF milleson_m_Page_009.txt
5c8fc5e12e6f71fe531c9111b0e804c9
3c5bb321dd7bc0ee9aa6beb3a9b4fe9325cbf1c7
62329 F20101122_AAASUQ milleson_m_Page_099.pro
722537890665fc8454113da3469ce0cd
c138dc99a887755761fe17a33f30ccec97f1b644
6536 F20101122_AAATAM milleson_m_Page_027thm.jpg
d022a38bf7738d8fd0903e557b5298c7
75db6925242ea94bfc60ac7ae50f4abed0530e68
1416 F20101122_AAASVG milleson_m_Page_010.txt
acc2e032578c519be0fb7bbc3d154108
76a1174364d85293be3e368ff41188399b790da8
56690 F20101122_AAASUR milleson_m_Page_100.pro
7a25fcc06602afd58fd75cf676f5fa36
9d20ee952e111ad10be06134b35a02d22770e599
17488 F20101122_AAATBA milleson_m_Page_035.QC.jpg
5b8574b8344cbe751896ff775fb8f961
47e849c22ee071c9b084cf1a77e71d6bcc8bf3ff
23422 F20101122_AAATAN milleson_m_Page_028.QC.jpg
951305733e028d5363a412710d72a8b9
8b1e53d024af60194f9b08319e5daae981822a2b
59644 F20101122_AAASUS milleson_m_Page_101.pro
8bf2aa81ed517a065591040e6c55bacd
f7f26ac754b4e8a18d9509032422ac49a928a148
5256 F20101122_AAATBB milleson_m_Page_035thm.jpg
04250116e6bd1d8544bc013641945119
3eeb4ffb7411294e840ab789abb6e61dd9a977a9
6603 F20101122_AAATAO milleson_m_Page_028thm.jpg
7eb8bdaac56c5a04fcca1abaf3244067
db110b3c99b95bbb36f506521e03fe689dbf0554
1837 F20101122_AAASVH milleson_m_Page_012.txt
adf6a9004ab9714717e80f36484d7bf7
86f9ccb3d9f92efa4077f771f727436a4516acb9
64056 F20101122_AAASUT milleson_m_Page_102.pro
fc62bcccadb4dfa15998bd413b5aca66
8e2a39c23d8502caeec7f59299c24f7334f77de8
23616 F20101122_AAATBC milleson_m_Page_036.QC.jpg
9630a30b93d625e3563f4c1bd39a1c4d
886c80f9586a2db328ac3e55867620dc804ac594
21800 F20101122_AAATAP milleson_m_Page_029.QC.jpg
be45f711cab4dbd293ae119ed11134dc
ad3fe52fddd667096fd458eae347031d4107810f
1777 F20101122_AAASVI milleson_m_Page_013.txt
71732240ebc469c51a41d5c3adaf0510
b6625a0b9bf0537c88832e689200b9c56670ea1e
61497 F20101122_AAASUU milleson_m_Page_103.pro
78630f0e3101fcfe68166aa69ab9a47c
4585c739a1c7b57b0ce3b3bc9832918291e94fb3
6503 F20101122_AAATBD milleson_m_Page_036thm.jpg
bda4f27d834e805e36f18ce205701ec2
02e584a58a52905ea28cd3b1f64144f8ccecac70
5887 F20101122_AAATAQ milleson_m_Page_029thm.jpg
c963712729919b79d54cee13e57125f4
9ae580418f2e4fcd111ed229b648b5219cdb6bcc
2070 F20101122_AAASVJ milleson_m_Page_014.txt
3648873b43366cbe8b29d5f5fbf78bb4
403911cd568e29e24addb738e3f2d3ff09c01469
60440 F20101122_AAASUV milleson_m_Page_104.pro
f6f2099043b7b12db052e67a3e9b71b8
0ec9b050506387bc11cfd30648d96c372264b270
5208 F20101122_AAATBE milleson_m_Page_037thm.jpg
0d76ee5c929ad5c31ea9bbfd7d15ed71
e65bca2199eba02c8b362648af35770852c902b4
2027 F20101122_AAASVK milleson_m_Page_015.txt
e83e7673e376b29fafc1adb864eeba87
f9f925c72219cebad452f126caf142a0543f30e9
39333 F20101122_AAASUW milleson_m_Page_105.pro
1344a5fc45995b3a0ec45acc5b0aec8c
313173c0f3fef7785aba582c79ddcdfc72be79ce
24364 F20101122_AAATBF milleson_m_Page_038.QC.jpg
9e7b1de7a227de61d1291d88af800fd2
ccef6a6c4d1060848f6a67cff0d6777c9eaf4095
1112 F20101122_AAASWA milleson_m_Page_032.txt
c0522777013b96a32297f66d6fb337c0
37fd6e7c9c2e10a403eb093043a26941d453574f
14413 F20101122_AAATAR milleson_m_Page_030.QC.jpg
78cfbfc66de0dd012c7ad3059e395521
5648cbdca3c1bf389a0b1ff990064b574f40c06b
2001 F20101122_AAASVL milleson_m_Page_016.txt
0033e1f791e4f4302f883a472285dbf9
deba1cb3c3849d50e4c608533adc07b8c0fce073
26937 F20101122_AAASUX milleson_m_Page_106.pro
08c87600b7bc97886bf5ddf62653f6b0
26a3b6ce99259efa33e2e3f375a9d2863b9527dd
6719 F20101122_AAATBG milleson_m_Page_038thm.jpg
650be3bd1deb7103edfe06406d0c49be
e0788a3729fc45bb38604bfd632e465d03036538
1265 F20101122_AAASWB milleson_m_Page_033.txt
44cf99266d34267270166a3b882ebb2e
c4b468452a29d0b4a8a195fd3a25291c1ca02ac6
4215 F20101122_AAATAS milleson_m_Page_030thm.jpg
ffe5ab3d0d120eea6baafde4a8833430
55320e1d35e09548822788b374ea2f9a82d31560
2039 F20101122_AAASVM milleson_m_Page_018.txt
f006b931e14165d2e0a2b21407fac2d2
fbdb4fa61f379b59570af9649e106ce6106fa402
467 F20101122_AAASUY milleson_m_Page_001.txt
2cb9c33995256a8dcf413cefe6ce0677
333f5276c1bb6176f5203d55c3d93348fe3cd0d0
19783 F20101122_AAATBH milleson_m_Page_039.QC.jpg
60a851972dac2679fb240e6f5d03de97
aa44df46c8c36ada6d4dbe0198a1407636069db2
1455 F20101122_AAASWC milleson_m_Page_034.txt
a68b2a98edf21e7d89181a82e11d55fe
d5fc7536e308c16e7f23eccb9d4ae554b1eec41c
16898 F20101122_AAATAT milleson_m_Page_031.QC.jpg
88ab4d8e5a0e96cc5b68a38054f5f73b
27b5f5b9cfa89e8caad7dec2c6baa91d7a3133ae
648 F20101122_AAASVN milleson_m_Page_019.txt
74d33faa51734b3351317591ba598c60
2b0043973ea5bf4311457bf7b6371b35dc3403e0
117 F20101122_AAASUZ milleson_m_Page_002.txt
227dee5e3459d9d98e4727c446e34179
7a3ec3cb9c46ce4b93d80cb2d501c3ca3d5c8ee7
5493 F20101122_AAATBI milleson_m_Page_039thm.jpg
b0834babb4b0d8310f642d37dcf300cc
60837db42a3d6894ec19f1a73c47b8a0fea9d511
1607 F20101122_AAASWD milleson_m_Page_035.txt
bab0becefdc0360908e55a6c422ab085
6b8c887d2465488a6f3f1ab22f3121af96284977
4730 F20101122_AAATAU milleson_m_Page_031thm.jpg
4189dfa0ab885c50773aa1ff21036521
a0b09e4ac6ad2be094b255a02b64d526f8f98166
2113 F20101122_AAASVO milleson_m_Page_020.txt
6fd8c61ec3329f4620b4525e333799d5
104a16dc43a59bcccd9fa46c7029a7e049cd7ec5
21089 F20101122_AAATBJ milleson_m_Page_040.QC.jpg
6e19b87b703c0b6fd88dcbb8da457f21
c3eff25a27f624caa2b92793755aebe2afb6b301
1997 F20101122_AAASWE milleson_m_Page_036.txt
71e719efd6b15302dafb2634a26747b4
f50c03471a9e9799fabccbbd9b59dcc5f0449fd6
14173 F20101122_AAATAV milleson_m_Page_032.QC.jpg
7c200c4c9027d5a25c6e0efc2a3ce97e
9ef4c57539cbee5ac4eb76d129f59f16d068e458
2069 F20101122_AAASVP milleson_m_Page_021.txt
72fb765e9d7d0f3ef60d82816761ce42
e13ce2a44b32e707370013c2e66b5e817056a081
5886 F20101122_AAATBK milleson_m_Page_040thm.jpg
128566a6226c1aba6a0e546b78968274
051ecfb1821182b9ff8f27572eda8829f0a1134f
1469 F20101122_AAASWF milleson_m_Page_037.txt
10d9ab5cf0cedc4713d98bd26c4b55a8
2ac94f2558320111270c9c2f6c4f7c3c2b9cb256
4387 F20101122_AAATAW milleson_m_Page_032thm.jpg
7c56278a2347ac8bdc7024a198d111f3
8d54b2971ed388625babd77a9a0a0438352c8329
2031 F20101122_AAASVQ milleson_m_Page_022.txt
ecd112af161bf7681f1be3c98aee9f48
4f076d29bb0cc9feed41665418e7224ce594b43c
23828 F20101122_AAATBL milleson_m_Page_041.QC.jpg
a3c500d50eba85c965e38d91b2cfebe6
1210e0ac197bdd520966b6bce597d3bea6655c9e
2057 F20101122_AAASWG milleson_m_Page_038.txt
d0f0ac8c315b950e75aa372b20942c7c
26a8b7555223a279db7744d15371dbbb6a3d48f7
4504 F20101122_AAATAX milleson_m_Page_033thm.jpg
1c5b9fd671c244923afb9d921afd2b21
f5824cc3ee1c80f0970515200965489bfb611324
1994 F20101122_AAASVR milleson_m_Page_023.txt
ec386b06cbaed623f980dde987767573
0652736838c410cb248d87873d2053d9f62e69f4
6208 F20101122_AAATCA milleson_m_Page_049thm.jpg
e5a53a1c5d510134cfa3f35a1e92dd41
2cc3c81e7cf34d49b349dcde5eb408a155b9f599
24113 F20101122_AAATBM milleson_m_Page_042.QC.jpg
9302c6b453c3a83348e489f7db14b44a
82e64e9511c2fce62aff15083fb5bd88a69e5b23
1655 F20101122_AAASWH milleson_m_Page_039.txt
7005411929b61b27973b6358c75e57e8
0c74d45be8991526d861908a51a7421f1debfcaf
18286 F20101122_AAATAY milleson_m_Page_034.QC.jpg
1db9fd6ee18d4ea80f92ff62f9b1956d
b6f97a044a01a04dcd192477dd556e3692ad4b08
1824 F20101122_AAASVS milleson_m_Page_024.txt
ded9aa421d24369c2b2afef2041a305e
70c21c525096853faf776f612754ddbe78c1b4f8
22067 F20101122_AAATCB milleson_m_Page_050.QC.jpg
a7d290b54b0cafc7fe9e869102c8ee7f
853d1a05af405f153a49468277659cb9161dd850
6527 F20101122_AAATBN milleson_m_Page_042thm.jpg
3520359aa72551115ee123d88c8b4798
c462b071cedd0eaba9ed82d5ce5bbfa292cd044d
5591 F20101122_AAATAZ milleson_m_Page_034thm.jpg
fc145dce791d0b46d53c47c09eaf7140
33be60ffc38aa5ca4f45baf0bcd75f6935e6944b
2015 F20101122_AAASVT milleson_m_Page_025.txt
13e122b2f6aa4548c2ffe7066b61bc62
91629fd98a7c19989be1c4258ab9f54193b19d41
6164 F20101122_AAATCC milleson_m_Page_050thm.jpg
3ced6d7c9be7ef8aa3df5ebe0752c80d
5094bccd8cf7dc78bc16fad4c32bdf35b72e3222
6482 F20101122_AAATBO milleson_m_Page_043thm.jpg
076a8a1a6bcc8322aacd287d51252b78
b5455a20b50ef142376fbecb21ce9d56800d5d39
1776 F20101122_AAASWI milleson_m_Page_040.txt
4380dc5f758cbcfe1dfd4d9cf382d6bf
70ee15237a9806f4bdbc516204c7a40dbd2b237e
1844 F20101122_AAASVU milleson_m_Page_026.txt
0494b9280b4048aca2717921046dd4f7
3d8c4e3a630802f3e55d6ae08a31e603591d6343
24046 F20101122_AAATCD milleson_m_Page_051.QC.jpg
ccc1a237b2002de62e10959667657d4d
c681cce3ddbc6c8e571a6fa114d8e2ab96d8207a
23148 F20101122_AAATBP milleson_m_Page_044.QC.jpg
7b25fd829e00800d29a6997ae13bb264
6fe6c0bcca84b252fda31db9fcb1c5ebf570a9ca
1983 F20101122_AAASWJ milleson_m_Page_041.txt
2b6a40e66015416a525112a53f4fa1ef
f748d004167c7032edcf109d99f69e235fc919db
1992 F20101122_AAASVV milleson_m_Page_027.txt
9f8d0d9a0392868f6caaafbc425e1ef0
77e56a7dda0d26ae9bbe70d1709a93564dc1a5b1
6591 F20101122_AAATCE milleson_m_Page_051thm.jpg
c54b677752f82c2601a90e95fa3d4753
4db2bedf7714f2664a656c9cb8aaee1fec31dd4b
6715 F20101122_AAATBQ milleson_m_Page_044thm.jpg
49279a4de9102f24ae6f3f19f62fa939
04f82c20f05e1194bb125b995b4fe19fb77f4ffe
2064 F20101122_AAASWK milleson_m_Page_042.txt
dec5a9b52f34418e1065c1bcf36ce3ef
ddd335531780b843f3017ffb68cd6ceaf5fcc6bd
1980 F20101122_AAASVW milleson_m_Page_028.txt
724203e51f5ce27b120d11c22f792bad
bd2c289be95610584844019ab588eb1b959ea052
23821 F20101122_AAATCF milleson_m_Page_052.QC.jpg
de4b2284ec98968cb8a4c897c78afa98
f657d71afab902d94ea70843eef23f2f5f67bb03
22425 F20101122_AAATBR milleson_m_Page_045.QC.jpg
415346b0c85717c0a5259151bf8749cd
10b131717a775fc14f978c60c1cc3056cf3d4a0a
1815 F20101122_AAASVX milleson_m_Page_029.txt
ed16608b3d68c8408cfd974aede95765
0f8c29c5ffaa2b38726477263262327715c48384
6512 F20101122_AAATCG milleson_m_Page_052thm.jpg
aafc92e995947495f52373afb08beeb1
6367591a8e6bb3f9191ce80f4f1a6d09ff05e76d
1332 F20101122_AAASXA milleson_m_Page_059.txt
d4e5edd4bdc15b2e9675db27f4a3e219
2d3164a12465c2391e29af9636706f58830e63a1
2025 F20101122_AAASWL milleson_m_Page_044.txt
8ef4fede9becc28256e7e10e3fafc69e
64561b84d09a1c3a02a7d444caa29057e4d21667
1126 F20101122_AAASVY milleson_m_Page_030.txt
b06553fd82d29b7dbcced26391fb64a1
6bd3432801bdf76a2bbcbd59094fd29fb4af2c4b
22361 F20101122_AAATCH milleson_m_Page_053.QC.jpg
34158814c7d3c96755ff0029234a60e2
60210ad0477400c78968a3690a980e03ce38a9cf
F20101122_AAASXB milleson_m_Page_061.txt
427369a46584fdc89af2ab4f13336e3d
2d00a787b154897010481b45eef71fc8f3565d0c
6203 F20101122_AAATBS milleson_m_Page_045thm.jpg
305871a83b5df121ff924ac3d7ed4009
f183f6c15b525a1e4230088b46aed296600db8cd
1988 F20101122_AAASWM milleson_m_Page_045.txt
2b319877366d8505196ee62190cd5be4
c61b6edfcd6d01297b8ffb217d1f9bb6ab08cf4b
1342 F20101122_AAASVZ milleson_m_Page_031.txt
8586dcbfbc59e548fdec1a67a45b0930
d3ef55ea6c90e1782a1bc9041bcfaf7b06105909
F20101122_AAATCI milleson_m_Page_053thm.jpg
915de4d33cad067ab325aae181a3f6ce
81904c89a715288f52b6bb645bff9e504538f2cb
1581 F20101122_AAASXC milleson_m_Page_062.txt
32236d0d4a327b490dafef7be4a747e6
e580e0af98cffa5117512b8655dd833b33b10253
16453 F20101122_AAATBT milleson_m_Page_046.QC.jpg
5e2dd3d8b7123c63556fbdda2142a044
548afdc0a1ca1e2501ea7f0cc0959287b891cdf0
863 F20101122_AAASWN milleson_m_Page_046.txt
be290c3495ee749045b30b49e9fb5d21
fc0e19310005426e8ffae1899fccf464d80d6f3d
21159 F20101122_AAATCJ milleson_m_Page_054.QC.jpg
44c9b95f5bd91177e13a7267e506f548
ea7dcce0e4e4b2f16cfc3bcbf5e381c3a20c8cfb
774 F20101122_AAASXD milleson_m_Page_063.txt
6bc815594d487a986263a67ee9f1df88
1322ea0ea65831e4717e2cc4495fec956612ec4e
5048 F20101122_AAATBU milleson_m_Page_046thm.jpg
b3d2624f5d820f03a3734dc7d18b5aa2
75b41bc9373663af9a082e0ede300dc59d825f91
1901 F20101122_AAASWO milleson_m_Page_047.txt
9d5d32a8c2b1e8b1df777ee31351c035
56dbec2e61be512e9e145471e23c46e1b8fa3856
6095 F20101122_AAATCK milleson_m_Page_054thm.jpg
9a790c6bdf24e51f861bc5e5a1b6e73f
df15a04e83ef8c12564beafb245ae8f779a101f2
789 F20101122_AAASXE milleson_m_Page_064.txt
7830510001b0d51cd4c6fc306d6477b2
af32fdb7f92b8745202143049c749057c7229263
21358 F20101122_AAATBV milleson_m_Page_047.QC.jpg
d31c4154920216f1b3e498a371e42d0d
0bf371f394b993677467f5d18661985268bfb835
1960 F20101122_AAASWP milleson_m_Page_048.txt
265a9c88969c7d783c829525f2c0338e
4746fa26b4b682d1310b3f684cab06b27cc038a4
20031 F20101122_AAATCL milleson_m_Page_055.QC.jpg
99384cdae6e2d145d75abac9cb4f6f93
019aafebb8d66ce3e27056029691291a2f3ffcd2
1727 F20101122_AAASXF milleson_m_Page_065.txt
5cff9c58d65d3aa24350c62d7bf80d47
e1e4dcba70d4c8fe7ecf167ba7d059bd582057c8
6252 F20101122_AAATBW milleson_m_Page_047thm.jpg
76dec1fb4eddc76c94155e1d3674221b
a27003b15074118cc5be04500ee7426b94656e6f
1883 F20101122_AAASWQ milleson_m_Page_049.txt
e3e6eae4cd00200451509c76374b4f17
88cd41b6cc86341ae3be9f228097d88b2f0c591e
14338 F20101122_AAATDA milleson_m_Page_063.QC.jpg
17451cbd38f6f4a1e04a5d412271262f
1a1295292d0a4e3f160d385e152a3d261b5aaab1
20095 F20101122_AAATCM milleson_m_Page_056.QC.jpg
b00fa83b28f95cf1bc87409c3b27b55a
bccf73d4b701e5dfdba2087461ce2de91aadc114
1773 F20101122_AAASXG milleson_m_Page_066.txt
ca0cc5b2154bd0fca95954d4591e22af
6483d203d6918039ca75943c762370e562b9b6d6
23082 F20101122_AAATBX milleson_m_Page_048.QC.jpg
051adcf58908107b34beb25e281d557d
495c67d5d697ee222035ba1bb8213d6b9b5f699d
1938 F20101122_AAASWR milleson_m_Page_050.txt
2a1d710ae9d3151833f7dca80a9c0879
7a35d6d7330d214a61791414f0c038a61dbabb24
14486 F20101122_AAATDB milleson_m_Page_064.QC.jpg
9fa6a14171f6837f4f6f40a786c28102
414f550f2f0f6c49afb537881f0a33e43796777f
5867 F20101122_AAATCN milleson_m_Page_056thm.jpg
98fc878e9b33358e503967b8ffbc157d
e2e2161753a8c4e4598d9750651231e89cf4ccce
1862 F20101122_AAASXH milleson_m_Page_067.txt
5b242148b5499713636451c7f96b53eb
70a28d79068133c93a6a30e09452fd4ce141cb83
6279 F20101122_AAATBY milleson_m_Page_048thm.jpg
d91d87ab4b42891c546f527a794cd99e
a82c51ef3f0efce901cbf58e17e60e2097adac9b
2073 F20101122_AAASWS milleson_m_Page_051.txt
d03734b9cb4451cca527b005c36b55f5
c8d9531d8e0df7f2593e2410623cd33d847c281c
4417 F20101122_AAATDC milleson_m_Page_064thm.jpg
14a7a0fb4d1e10d07c0ab5ae6811c95a
7deaa2bffbf2c1fec50950ff16fbc21b64382279
13706 F20101122_AAATCO milleson_m_Page_057.QC.jpg
0209ae658aa7bd3dc5548f8d2d140adc
81cd217b50988692ad006a32c3ecf59bd272877d
1388 F20101122_AAASXI milleson_m_Page_068.txt
3d0cee69465fee1e57d8c66f9714748f
7bf2aeb4173f287f07d4ca2e269c7b7767aecec5
21585 F20101122_AAATBZ milleson_m_Page_049.QC.jpg
2800afa62a6ca8ef6bb350fbd82f8e90
280f6850e499f6e99cf2ad1b24b2f6ecd0ea51fe
2088 F20101122_AAASWT milleson_m_Page_052.txt
d41c252b076e67e8cbaf2f5565888c80
42ae61de23f8884d05ce5d3fb1200fbde62d5b97
17942 F20101122_AAATDD milleson_m_Page_065.QC.jpg
a47a2b12bbce3432a7b5172172e454cc
6a0e594b7b069076500f74483aee5175a84bc37f
4537 F20101122_AAATCP milleson_m_Page_057thm.jpg
5d518e2fe7663cd3945c99dbfdea72cd
bace1d1bde94407163cfa6772c798ce5704d5f3e
1951 F20101122_AAASWU milleson_m_Page_053.txt
ff9eecb3cc455b4aa1e3e9f8ba1734d6
eaeb1bd6763bcaee85a203342f85ad8a971f4c6d
5145 F20101122_AAATDE milleson_m_Page_065thm.jpg
c93ed83f750eda2aa1a05f3dd9ce2893
e1b5ff60da19531eaf702400f319856269242172
17192 F20101122_AAATCQ milleson_m_Page_058.QC.jpg
e0b4b1fd3976eaea15f5f4a75ac6c4e5
6455b89ddde6d78bfa2371332406e6a896c6ca71
1823 F20101122_AAASXJ milleson_m_Page_069.txt
7a09f5be7f388e860efc3204c0013a5d
21515f3d6224c3a6705a145562b69e65d17b0f7f
1876 F20101122_AAASWV milleson_m_Page_054.txt
7077d8c0742cff3a07908af8065409af
2ce171fe5b7b606b113de56acd84e4cc441cae4d
19001 F20101122_AAATDF milleson_m_Page_066.QC.jpg
cd8dc941e629f5358d75efbff8367f61
8b5184bfd38832ad36ebc89054b9ad656aadc042
5384 F20101122_AAATCR milleson_m_Page_058thm.jpg
7f6446d24b276adbd40b4a5ed5977ea0
da79ae52c8cf2b5990eebc3bbf4f5fd1b82689f7
1698 F20101122_AAASXK milleson_m_Page_070.txt
0aa5fab64d367b01cdd844bebc7c9c63
574a9dfc73f0eebbfe5bb5a52acce3a606053cb3
1553 F20101122_AAASWW milleson_m_Page_055.txt
c8eedeeee529ad2216c380e51d163679
f1062612b3643a44c7946eb6f673bf94cf917e08
5316 F20101122_AAATDG milleson_m_Page_066thm.jpg
0c0748b351275fc49b0769181fa67098
e32bd41ee0d6597d2099f51c1b0327be1de6d1ce
1629 F20101122_AAASYA milleson_m_Page_088.txt
0eb397cfbc1eda16085d7cfe1112ea53
bac348341947c8ceca5359a31cebcc99ee86607e
14683 F20101122_AAATCS milleson_m_Page_059.QC.jpg
80541e0fb1c682c220ae5fd0a20897e4
1417dac358369c2602b40ca0d04e56c91b8e11a3
977 F20101122_AAASXL milleson_m_Page_071.txt
e29b1e25d40331af2ade5cb4511c4907
d762f21c38b9dea3df91cdc942aceea61bbad141
2185 F20101122_AAASWX milleson_m_Page_056.txt
1441f638909f417b9d586d68dcbd8e30
fcd076f37815f321888d9723657c3ea6db9d1b20
5363 F20101122_AAATDH milleson_m_Page_067thm.jpg
41084eab3d9989a15678b2dad32dd1d4
06aca388adab6dcd49fd4e77c07461fe412b192d
1480 F20101122_AAASYB milleson_m_Page_089.txt
c32441f1b2bb418e885f800c406aced7
7f11cb262d5b140e739396f28452f8eca54d2ec6
2128 F20101122_AAASXM milleson_m_Page_072.txt
c8fbd62888e7e5c5de5e4461076f1b6d
93251243def4114d9cd0579134c7b16297273fa1
1357 F20101122_AAASWY milleson_m_Page_057.txt
fd59b95023f9415365beea3c9c643fcb
aad2bca76728c3924e2b17e168886dfa47eb4f6a
16561 F20101122_AAATDI milleson_m_Page_068.QC.jpg
ddaf6fef7dbe586bb0a4bee1587ce470
03ede55d164da518912fc37b1f8e19ec4c8f269b
1313 F20101122_AAASYC milleson_m_Page_090.txt
54e6f4de79c981dfebbbf4d6efed5417
b01c8d2c9b805e406240f1bdc532db7340e37af6
4475 F20101122_AAATCT milleson_m_Page_059thm.jpg
a1d3e81c97d12ff58c4d608e23183748
22d89eb31803bdf13244bfcd916e02aab1fbafad
2117 F20101122_AAASXN milleson_m_Page_074.txt
5194c0c0a52849c8c1d66e74add6c143
04a5b90ffeec4510669ccf4656da43926d481f8e
1787 F20101122_AAASWZ milleson_m_Page_058.txt
af4bc56bf5656e1b8c1f4de36dfc88ba
d0480d57e5ebfc2a380f959e711752b9b380fa5b
1345 F20101122_AAASYD milleson_m_Page_091.txt
c78418fd1970f867a4d63a3901633aa4
74f3aa55a9f44315f5fcf17d16345701e88f79f1
20439 F20101122_AAATCU milleson_m_Page_060.QC.jpg
846f7359318cdbce9f4a740af853fa4f
b2afd517ef0e26007aca4f3b77e4b59466c3f348
1985 F20101122_AAASXO milleson_m_Page_075.txt
12dab333daaa2feb5508a936abaa162e
1691fabf11d649daff18b6280a850011bd73316b
4728 F20101122_AAATDJ milleson_m_Page_068thm.jpg
2189afb40c77f090dc8ec302ad685752
b2d5447a87a395cc2eeb4107f2174efa31bf6a47
1435 F20101122_AAASYE milleson_m_Page_093.txt
f5e1a7844e411883ec70831bdd733cc2
bc48331018bd16b98ca6156616e484093ac2a5ed
6002 F20101122_AAATCV milleson_m_Page_060thm.jpg
c538e2f62e6a5f899dc952a6fd765aa6
5a61facf7013b8887c85cc0c8b7c583aae168c57
F20101122_AAASXP milleson_m_Page_076.txt
e5a37e0845f8c7aaa0d875060a08b194
f8bc558599ec7f8757ab1c198fda145d2f91c45c
18436 F20101122_AAATDK milleson_m_Page_069.QC.jpg
8273b6931709480d1943852a32f0a951
2d5140e2f9f59aa6a32b7b8e4e9af499bed8e39c
2282 F20101122_AAASYF milleson_m_Page_094.txt
b63a019cc6af9116369b15780d74ddd4
5595d30e19c3cda64cc15e41d407eb4ca3359bb8
16902 F20101122_AAATCW milleson_m_Page_061.QC.jpg
fe31f4c469eebd04c8693b47436340fb
9ba23d2685024ac4aacbefdcba599bafb1d95689
1984 F20101122_AAASXQ milleson_m_Page_077.txt
5ff1b5700edb393aa471f84ea7fa15de
60acb581c42f8281af54fbdbd9f050099548060a
5240 F20101122_AAATDL milleson_m_Page_069thm.jpg
387190f7925a5f9325e0bff52ef947a3
b3624017a3d4b98b62b3b45456f072e4a27bf32d
1404 F20101122_AAASYG milleson_m_Page_095.txt
cac4208fa73311e47ec336c04684859e
484cb390809fb3b2993331122f246ff014e29ea0
5047 F20101122_AAATCX milleson_m_Page_061thm.jpg
32b44ddf88ff82dc8451c94d2999ca02
b7d910cc51188ccb9141d131d27d46c73576ff6f
2055 F20101122_AAASXR milleson_m_Page_078.txt
39ccf536ed7744a83f64376f346abbe4
92c8e137e7e056419fce27af3a3418a2dded757d
6421 F20101122_AAATEA milleson_m_Page_077thm.jpg
eda8e3d37c432c6c2457931c605b0bbb
790534e9838b713f08b906bf00cb243bdb182343
18289 F20101122_AAATDM milleson_m_Page_070.QC.jpg
6a72e8911cfe1a5014a69fb0ea106acb
5e23f8c5e89b5e33e9447ed0666b6d3ace7c8963
2301 F20101122_AAASYH milleson_m_Page_096.txt
d346015e979f96f83107d140f0efb73d
ff908666212ca79b48cc489b910f3dc4ab1a97d7
17278 F20101122_AAATCY milleson_m_Page_062.QC.jpg
4cd8d71175a6d09db17aa5dbd80f227f
6362d8a705b6dec8c3a8cf172e15adc00723b372
1807 F20101122_AAASXS milleson_m_Page_079.txt
af03e7447bc299f753aea0dfe318c6cb
0d8ca334475ef02b2c0b353c84986c1feb44e0fc
24097 F20101122_AAATEB milleson_m_Page_078.QC.jpg
4344583e49016cd1cc7ac452c7c4aeb9
81a0b40867e20f2adf9cf545a4b201b30bb6dbbb
5089 F20101122_AAATDN milleson_m_Page_070thm.jpg
c2c8e8eb8181612417d5c1c741baa67e
adecca290c36bed88bea0b492bb10e37817c9344
2566 F20101122_AAASYI milleson_m_Page_097.txt
f2c388479057828216dc8338374b0743
85c2e896e612210233ee78713886589c4f3d4f45
4958 F20101122_AAATCZ milleson_m_Page_062thm.jpg
6f8ddc22ee02744cb681bb26b8f6ff20
35226ba5dc55591c4049cfa2fff7d1d24e51b848
1897 F20101122_AAASXT milleson_m_Page_080.txt
0227f1469aea9c818aac246b5bc8d479
46c561430be488fd077030c7c7decd436ec84632
19316 F20101122_AAATEC milleson_m_Page_079.QC.jpg
ce59dbbc0dcf3af4e9cf584491be453d
0f59c68a6e5577438595b6455ece5a70573cf707
16078 F20101122_AAATDO milleson_m_Page_071.QC.jpg
3d706c71eedcb57ea21866c9b7cdc1f2
e661301fd6fed5f5a173c01ba26ab758d509df2d
2611 F20101122_AAASYJ milleson_m_Page_098.txt
b688512ad4b4e0cf72908d123b9cf04f
7bdd12ec99155bce390072fb0c73ca5a7d41e4c7
1929 F20101122_AAASXU milleson_m_Page_081.txt
671a774680d510f3b537615550e9defd
fe0c24aafb51876b738ff83b28ef92e48e0ed519
5346 F20101122_AAATED milleson_m_Page_079thm.jpg
393ff499b17441b716ace4256a4a3cb3
5586df8ed7e823d42a7393183a3fa25324847ead
4538 F20101122_AAATDP milleson_m_Page_071thm.jpg
8936bbab4b278609e5b39a6010b3bff8
b08457905659433b4d01d843243094103daace94
212 F20101122_AAASXV milleson_m_Page_083.txt
ea4f768d29837b5ae27f4967d4e5ce41
9486bdeb7bbcae64a2b0ed79f2812c84b712d365
22384 F20101122_AAATEE milleson_m_Page_080.QC.jpg
148d2115451471ac86ddee9bd1f3591a
d220060ba39d9cb971517cd2cd17b71f69192a37
21645 F20101122_AAATDQ milleson_m_Page_072.QC.jpg
b598656d3a04022cd2f16c79f93690b6
45c791a0ac28b1044d7e66280240091506addef4
2530 F20101122_AAASYK milleson_m_Page_099.txt
4c3b33f45ff567836418fd75c7e948cd
ac6c352b2af9ed0d863c7cf28500abeb444a2320
237 F20101122_AAASXW milleson_m_Page_084.txt
a3f34b84be64ba6f1f4255d601446592
18171d087970a75c7d065d5f4adbfbb93571ac8b
6277 F20101122_AAATEF milleson_m_Page_080thm.jpg
e985691c5b1c831628c5717157a5e70d
b36473fec4166c9e3ed93f3eaa80e923deae3368
6261 F20101122_AAATDR milleson_m_Page_072thm.jpg
ea638a018139edccd97e5ee0c7adfe5a
97655560cc27f8fc15802b47e0017d65edf016c1
F20101122_AAASYL milleson_m_Page_102.txt
b98f8f9b374616564dc8663ec4d9a761
252edfaa481ebff4006a2a499f8ae2edfbc16281
1557 F20101122_AAASXX milleson_m_Page_085.txt
638eeb70d46395f4d1c8d172175f6455
1715a9e350299278b73c71a9297428de4fd91c4f
22687 F20101122_AAATEG milleson_m_Page_081.QC.jpg
9c560f810d0837f6aa8840fdd78351bc
573c967d60ee43b29a4b39e8e813ff815135e3c9
8140 F20101122_AAASZA milleson_m_Page_006.QC.jpg
02fde3927931ce1c3ee326c81b603e32
90d92be5feff4fc8eeb27c757c89f01133349eab
20434 F20101122_AAATDS milleson_m_Page_073.QC.jpg
1e4bb5db52e5fffa59a0989058bbb548
47ef81980c158fc1c9f6ec6d415c23eef4be26e9
2510 F20101122_AAASYM milleson_m_Page_103.txt
842be0f784bd3e7b91e6ae8ccf69c380
27b31e2c00a06e8de5e2cfa14895e53a69aeea78
13 F20101122_AAASXY milleson_m_Page_086.txt
2727fa3b8e78a930ec6ca0011ed09a73
20e24349f0076fb709b988487bfa4fa8d1641894
6530 F20101122_AAATEH milleson_m_Page_081thm.jpg
2d72ff3736e48dbbd0117a735dae6b27
67193d43ee6d86c51acf3c7f69b69a9a925ba419
2672 F20101122_AAASZB milleson_m_Page_006thm.jpg
163220e2f15a617b8858fcb17ca07e60
d59b0de3a62d8b8f3d9b4afceecd4a046b3637e2
5398 F20101122_AAATDT milleson_m_Page_073thm.jpg
fcbfb6b16a004cb9202a9eccccf8bc0c
33385911ffa42e92ff1553494a7b2f98890da3f9
2456 F20101122_AAASYN milleson_m_Page_104.txt
a9b339c9ae066aae3beb5bebcbe27d44
0b5c9dfb5ed698a623a92049a426561632b4d1bd
865 F20101122_AAASXZ milleson_m_Page_087.txt
5c13b141f9a9f166c2892623b3531383
af9c7afd7203bde7bea11e3c1b2f4d5b31fcc077
18030 F20101122_AAATEI milleson_m_Page_082.QC.jpg
548df7d7caa22d9214bcebb870e2428c
396675752037f1aa5b720d694272ec12cf926939
8391 F20101122_AAASZC milleson_m_Page_007.QC.jpg
e14edbec5ca29ff83eb244d80a9d7d53
a750ff53f9e2d96091bb623aa3e89c5cd5aad1c4
1628 F20101122_AAASYO milleson_m_Page_105.txt
bfbd022d59ca355c25a603797b23502e
1982f274dd65a65cafd4883d164a2a3828e6d473
5177 F20101122_AAATEJ milleson_m_Page_082thm.jpg
aa1305c8ecc0e0afdb73c6b3eef184c0
9ee585bbd0453bfac883d07a6f496cd3bc4bc49f
2501 F20101122_AAASZD milleson_m_Page_007thm.jpg
0ff208ef9fa577988b11ca53464f7bd8
afaf915b51441e774985032583648f65c3662bab
24675 F20101122_AAATDU milleson_m_Page_074.QC.jpg
ba90b310f021fdb8f9b8f2ccd941f55e
232afb986b8c162d5c11646b1c38903adcb54932
1117 F20101122_AAASYP milleson_m_Page_106.txt
0b55ff7e18c8bd219712b64a5e885dc5
1b2ab2ae1e835dda8bfa480266a6937dc32ba0d0
13497 F20101122_AAATEK milleson_m_Page_083.QC.jpg
5990552e06ce3c8ac474e925ebac84af
227b4ea9fb9424cf4b6ddf1af018420cf413bf83
23434 F20101122_AAASZE milleson_m_Page_008.QC.jpg
51ce581627032fa8f7c0b18889464c0d
7bceab5ad28cbba3e737b08c9875914a4e90ba77
6783 F20101122_AAATDV milleson_m_Page_074thm.jpg
3bb898bcfbf55826e350f8c546ab7dc0
c50799213c2180df02ee529002267b2a5147187f
756788 F20101122_AAASYQ milleson_m.pdf
77c49f589761338e3f0a839016595bec
1c3aeae19756aa8ab308171fec2bbed7f048995e
4210 F20101122_AAATEL milleson_m_Page_083thm.jpg
f0272a3031991904ad71153c457a3f11
ac38a02d7206de89a510a7ffed41df9f521118be
6153 F20101122_AAASZF milleson_m_Page_008thm.jpg
b506cabddd8be7d5c5b45834b8fa8168
00ad85fc84478f651aedd96028ec9fe881b68fe8
21168 F20101122_AAATDW milleson_m_Page_075.QC.jpg
d1431bb80ff1ae13e7291f71987e2f3b
75861dddc7862d47e5e51f000fd7e2e2219e923d
2436 F20101122_AAASYR milleson_m_Page_001thm.jpg
9621a1cdfa0ecd843c4e8af7134e60a7
101374e7ef8f09ad01bb931c7525ef0a9be3cf88
5422 F20101122_AAATFA milleson_m_Page_093thm.jpg
9e863b0a03eb8e5863b1cfb6344edfe1
678312c93be1adf6cf2fdcc7fde69890fa6faa31
15504 F20101122_AAATEM milleson_m_Page_084.QC.jpg
34f8a66f65fed71ee0401608ecaed5b0
1d67bb28ee0c43d9552c4e26f4c34155492f6320
28160 F20101122_AAASZG milleson_m_Page_009.QC.jpg
68c83acaafd36cf2e05172eabea0f8e6
a04dfeb141583d635a847fc7b38de00a3df9e3d9
24194 F20101122_AAATDX milleson_m_Page_076.QC.jpg
76348fa27892a0ed833929a750c0579a
41664da851fb7f63af2e23171eb89074c16cd6e6
3339 F20101122_AAASYS milleson_m_Page_002.QC.jpg
f1c85b9bfd4aeabe1e57d28bc1157d65
75d09ad178cebea3acd240f05f29e1f59a8d72d6
24558 F20101122_AAATFB milleson_m_Page_094.QC.jpg
ca82fbc6c687b46f282cc2aa0f30cac6
ee5563a3e186ef07f74abe071ab1e2bf4d0b86b5
16854 F20101122_AAATEN milleson_m_Page_085.QC.jpg
8158620b72311026bfb3ef161de7e317
df98c88474ae5fc6fe1bff70af926f3e9b5e3330
7198 F20101122_AAASZH milleson_m_Page_009thm.jpg
85ddb04c590221fd5b6fe33a2ea77d88
b08bcf08c197f3841a439b4c84939125dfaa2a31
6689 F20101122_AAATDY milleson_m_Page_076thm.jpg
eee5c951c807f2741d7d95305813a233
65adc428d5ee1e69899237315bda31bedeb0076c
1389 F20101122_AAASYT milleson_m_Page_002thm.jpg
7b3cca1251ecb7be22fe069f967c3fc3
0b8aa5f2089289f0e7063cdc5c4fc039c08d2d60
6379 F20101122_AAATFC milleson_m_Page_094thm.jpg
232097fc576424f534403edfc3798368
70cb93d1a027ae51f85d6b4fdcf0c65a836d5534
2845 F20101122_AAATEO milleson_m_Page_086.QC.jpg
63879d52f8fd293bad40191f757c4d93
86a049c9ade46f7018975cfc8217d86517603e3f
13780 F20101122_AAASZI milleson_m_Page_010.QC.jpg
1d059d67fa5671d266c2f36594031531
255f5b72b90b7b6579052a199093d59aed74943d
22602 F20101122_AAATDZ milleson_m_Page_077.QC.jpg
257d4a537cb7a6cf9e96d2aa7ab44c34
ddea07c130eb805719afc974437ba4a0139a06b4
7494 F20101122_AAASYU milleson_m_Page_003.QC.jpg
bcc121da5fa42e4570670736396a8412
204919d383fecd44832e9a74f132404e6fa610ba
18612 F20101122_AAATFD milleson_m_Page_095.QC.jpg
92c2279f6f25a0a6984ab045fd81a1b9
b6a140ead5b22e880de50afae5bad467464bf21d
11591 F20101122_AAATEP milleson_m_Page_087.QC.jpg
532bb681e8bed620cecbacd54d4a69b8
00cc6f93f69540376c7d1397cfba9d0cb88ca95b
3722 F20101122_AAASZJ milleson_m_Page_010thm.jpg
d37bfb2229a99f4686e8cfef462b0aae
585b620560b535f10d77e00ea7700f590e99883e
2544 F20101122_AAASYV milleson_m_Page_003thm.jpg
b52faa31ba22ce88f2ec924093791fbd
b97b4d0a51108a8a71534a7912ddcfd3caf4c755
5110 F20101122_AAATFE milleson_m_Page_095thm.jpg
305db430d29126fbcd0b7d3ac912e3f3
e289437170dadf835280f45ea40ace36f0181548
3750 F20101122_AAATEQ milleson_m_Page_087thm.jpg
b89dd9e5144e0c7a3c5ffeed76891314
2bbdef7604e6d90aba05eeab2fd720b4acdd6089
18138 F20101122_AAASZK milleson_m_Page_011.QC.jpg
afe05b9e82c19b2c70645e113db1d4c5
f4cf35685d03f216481f136146090e9edbf78e89
18432 F20101122_AAASYW milleson_m_Page_004.QC.jpg
ee813b1be6fe60df540584a78f3cbe7d
0334f1c90a2be6d02a3c2830d9a47aeb57f87183
23290 F20101122_AAATFF milleson_m_Page_096.QC.jpg
da89c89ce17cb57b5edef745e7c04073
46235faae72b2ba0bf9a5a383644fa73349bf139
18984 F20101122_AAATER milleson_m_Page_088.QC.jpg
ab2d80df246ff7240dec28a9e52d0c87
f5323ae7721c0e9aca750d3f283601953edaff73
4930 F20101122_AAASYX milleson_m_Page_004thm.jpg
a517cb8b9568302f74395e01c4536320
db29dbdbc169ad8f1cffb828c54e1271dbde347a
6224 F20101122_AAATFG milleson_m_Page_096thm.jpg
104ad2ff3d39f5853498fafcdf0f867e
342e73f8b884836b0fdd348bfcebb3da66d14a73
5032 F20101122_AAATES milleson_m_Page_088thm.jpg
38edb51db088d8cb49ea42683c4cad5d
0638338ff0b7d606fa7b740cbe5a6b642bf596a5
5264 F20101122_AAASZL milleson_m_Page_011thm.jpg
f63080b28760331d51ccb4c8792af90e
b38f903c1d77b26404def5bf1e3a74c208515e23
F20101122_AAASYY milleson_m_Page_005.QC.jpg
7f89c4359b0f5ae4d91962c8fab0757a
a833605658ff794bc783a772c62b29a9b8818d73
27253 F20101122_AAATFH milleson_m_Page_097.QC.jpg
12c17a126107835a95fc67c4ed5044ca
fb6df485a604a94dcda3a0417b47d2f6bd721c12
18337 F20101122_AAATET milleson_m_Page_089.QC.jpg
102b8a52e98b6b8aaf29362d5a7c4737
cd006338d89dbc332a55ef12835e19c0f25d739e
6089 F20101122_AAASZM milleson_m_Page_012thm.jpg
391779bbbbb699191354e0844cbef78a
679a20d2eb6f49db19891db93019714162871c14
5765 F20101122_AAASYZ milleson_m_Page_005thm.jpg
01a94d5a21a446ff22dacf5bab122248
eff43c4d6d4d503c65d133b38a702c1d70ba70be
7235 F20101122_AAATFI milleson_m_Page_097thm.jpg
a2ed970b0556ec004d05c41d1f1d4bfb
bfe87d09bc23bd243b815484c96e39998c0bdb4b
5576 F20101122_AAATEU milleson_m_Page_089thm.jpg
0f50c5997a3063e2da5018d382f8abbc
dbcfa185b836338fa1f002b25e8e183a8a768fd0
20650 F20101122_AAASZN milleson_m_Page_013.QC.jpg
12becd31f05579ea2a8cab46e7ddd59b
57693d0c08d827742adc86253bab69a969a0c35c
25534 F20101122_AAATFJ milleson_m_Page_098.QC.jpg
eb53f010d81c4638188000033e2bccf8
f64316f5f016a434667fcf03f1a3e0e753332b3b
6135 F20101122_AAASZO milleson_m_Page_013thm.jpg
b0769536bffa6ea2b058d00def483600
a6667a764d23b8367810edd63df2efa0943e38d0
6797 F20101122_AAATFK milleson_m_Page_098thm.jpg
18414aa46b24d2d4dc83291a1a386b70
67dd345f63e4efce46ed715f0e45f064771fd315
5302 F20101122_AAATEV milleson_m_Page_090thm.jpg
b6bd11683dfe21c4e0b824c07e59217f
aea22f6388aff5cc96a95a19045504fb2bc4e967
24621 F20101122_AAASZP milleson_m_Page_014.QC.jpg
c2ff15be96e0fcfe84fbc53cbdc34509
234357d72077dde4b90412ea005ba6b01932cc50
6944 F20101122_AAATFL milleson_m_Page_099thm.jpg
f46641f44f2223be13450b12ea528b5e
efdd248c8ad310500fc38bbc51c6c7fd503b5807
5052 F20101122_AAATEW milleson_m_Page_091thm.jpg
0c02db3bdfd92df726042f5f7a919047
fad7988897e8d8862fbcece03b6aa45c7801d10a
F20101122_AAASZQ milleson_m_Page_014thm.jpg
7d45cb713cfdff66b66c10997c39bc2d
6136b70febfbf2a279eff96c979de3787ed6722c
24215 F20101122_AAATFM milleson_m_Page_100.QC.jpg
d97659646115ee123fd74b08836d25d5
f4954422b8eaa3aa3a7fecb89d9e9c6fde55f94d
18509 F20101122_AAATEX milleson_m_Page_092.QC.jpg
1b180d5353ef8b5a68dd7d27dac19073
428d85dcd15075aaee253bbabc4e1aa66ef9847b
F20101122_AAASZR milleson_m_Page_015.QC.jpg
d98d798ebe14287e2433b15e71bfeebb
c2b328e27638825e7458b7d014a5350963855b4e
6666 F20101122_AAATFN milleson_m_Page_100thm.jpg
94a8f688630228da43e5961fc062b4cb
f3168d899fe55cce1e8824e2253c4a3b869222c8
5167 F20101122_AAATEY milleson_m_Page_092thm.jpg
1cb83c8286ede513f85d79d92c43086f
05a87ec3b4a5e0eafd39da6c350def559e87471d
6604 F20101122_AAASZS milleson_m_Page_015thm.jpg
5d285bbaa86b460fbb1032b5967b3174
d7392a8d5d2670b69d81fdd899c21b3c5669fd75
24365 F20101122_AAATFO milleson_m_Page_101.QC.jpg
a31ff11333e5e4496e6368322e0576ea
9561e2ceaba288dc8054b19300addd3e8ff2096c
18869 F20101122_AAATEZ milleson_m_Page_093.QC.jpg
4f7063c691fb27652031c68dca664be3
92f7721d8033d59cf3110b677e7364d305b71403
23299 F20101122_AAASZT milleson_m_Page_016.QC.jpg
816725876d825b16faf66bf37980edb4
7673ca653e0e11457f313c93a4e10fe45273f80f
6766 F20101122_AAATFP milleson_m_Page_101thm.jpg
936c49ccc1b796b85924e599955cddcb
0cc68b94f5c425f4265e91716ca9a1222aa9e3a2
6623 F20101122_AAASZU milleson_m_Page_016thm.jpg
cee55db1ec6ad764c10d078714e48715
db9889ec6ba561ad7798849a1095f688afd0ea88
25308 F20101122_AAATFQ milleson_m_Page_102.QC.jpg
56624f0b1921a75ce7fff49bc62bfab6
fe61eaf0e9b7c458a9fad9bac1c031bb8a8d1979
6510 F20101122_AAASZV milleson_m_Page_017thm.jpg
a5336d5fa2a674a2636fc3f3ef442834
fc8a985d143056247aeb68aef98939e550050f13
6913 F20101122_AAATFR milleson_m_Page_102thm.jpg
d4412e3142d8e0926a3f7973e8fcb6d5
3705a64207db8c0d065d2001ae73bd48b87e3cb7
23379 F20101122_AAASZW milleson_m_Page_018.QC.jpg
4e30b98abeb2f96b5f8a44e789d85812
59efe760713648121ae5ee595ced42273b30e4e6
6818 F20101122_AAATFS milleson_m_Page_103thm.jpg
d1b5f590fe774c2dd869b4978d47b529
d3e1486fc2224fe283fd27119dd00b4b4c9ca7f3
6447 F20101122_AAASZX milleson_m_Page_018thm.jpg
49eb2dc45cf5a8b483ae56b8f2898450
361a0fdb5296c6df37d979911e25caf1b0fa9693
24641 F20101122_AAATFT milleson_m_Page_104.QC.jpg
08ac8a83f79e5ff96e795efc9d0f297b
15186b7d68482d9e4c9980b6cc1db21b7afd911b
4392 F20101122_AAASZY milleson_m_Page_019thm.jpg
0fa7ca8eb90554ce5c0902b92518c197
c87fdb92f632347cdc91ea1255b6531d8495cf88
18012 F20101122_AAASDF milleson_m_Page_085.pro
45582d533b2f072a14e53bd061b54aef
37aec43a1ff934fe83479322904ba1cf7d27b93a
4955 F20101122_AAATFU milleson_m_Page_105thm.jpg
343dab8934fa70a5c8abe7a3c19da320
5abb1387d5914ec15c76f8d4feddc84ac4a65ac1
24910 F20101122_AAASZZ milleson_m_Page_020.QC.jpg
b97a9fafa1397b26c00d0c50e6f324a5
3253ebaef7dbcc773f4617d52b41245bcf5b29c3
2029 F20101122_AAASDG milleson_m_Page_043.txt
8f3af0954140595ea1e6ebfb7cf0ae0e
f3118940c60ae2e34bb7c048476f827df6f7e7bf
14077 F20101122_AAATFV milleson_m_Page_106.QC.jpg
4b0e6dc4d80fe6d3f666166a42f50449
d6222f2eb13fd54fd2b42e274668123f0b8b45c6
81632 F20101122_AAASDH milleson_m_Page_008.jpg
2f0445d2ea3018cd84afef68602a2dbf
a5f9d219885a9cc71fd910d8718e71e812ff89e3
17329 F20101122_AAASDI milleson_m_Page_105.QC.jpg
3f3ea07ea1c0a2ea41775daa5f59afae
4f4e71e265abd5d4b61270b2870935f9b4f4a25e
4152 F20101122_AAATFW milleson_m_Page_106thm.jpg
2b1e1ebb00703d7cd1b694ea10ff790c
88924c8c35bb60b042fec3a5c1d2dd44e5723f60
19689 F20101122_AAASDJ milleson_m_Page_067.QC.jpg
813b143e5c32958277e59e3b79253f77
0b35b97f5bd35745bcdbec2cd16a4ff2a203f983
124150 F20101122_AAATFX UFE0012043_00001.mets FULL
3040f6de2c1d7d22af015f3bc0423b54
7108df264f657e4737be7fe16e96f50d57b87f61
1251 F20101122_AAASDK milleson_m_Page_086thm.jpg
27de5dcfaa7ca44e3c260e0fdc8a6a61
d4e1939b1dabeec94206074ad43e8346a0eb6d22
21661 F20101122_AAASDL milleson_m_Page_012.QC.jpg
63609105b192972604fa91a4742fc004
15c1d0e48272a9053d3c408fbd7867456ca8cbf5
39572 F20101122_AAASEA milleson_m_Page_067.pro
56225cb12094bdbdc03693f554dbc403
23f51c567615d04f2f72f7f170080c0c1550c79e
5191 F20101122_AAASDM milleson_m_Page_085thm.jpg
a71b39cca852b33050df7c3977cbff84
c16b37435a78e16e40408fa934401f1530efab46
F20101122_AAASEB milleson_m_Page_093.tif
194651099858a3e40e241b4f76913a83
43d32a2a748d91bde4198657c29f5fdc6bc6e4a4
1964 F20101122_AAASDN milleson_m_Page_017.txt
15f5dacc4cc3af0ccb8ad51d997e506d
33185ed5b7a5393b79d1f026b95b91b1772cafd2
1713 F20101122_AAASEC milleson_m_Page_060.txt
530fe5e6aacd0b0089af0b555a8754a0
62066837fa5217122b8c2bf0798647a82e493b2c
2353 F20101122_AAASDO milleson_m_Page_100.txt
8e58471c742e69a47e407343dc3ffbf7
7e52ba38fadc0bfc71c3babc52c84c7ee093bb99
F20101122_AAASED milleson_m_Page_046.tif
6ca5b865abd489b7cedc21e0e430a603
f6b658bcfe77550a13037a5e35a3a764c9df078b
73992 F20101122_AAASDP milleson_m_Page_078.jpg
371fcb88fca3ffca126ae19c3813bd26
54963114622f4dd12ed44fbf51d70efe4bcd6574
23779 F20101122_AAASEE milleson_m_Page_043.QC.jpg
c342f1d958f042537cf76daa151c0368
9817ee28503568435ca89289642bf1d4207f861a
F20101122_AAASDQ milleson_m_Page_003.tif
89dc9fa221a09dbaef3ec888d61bff37
97a34a038f214dd63e0912d7b4ebc57c5c501e97
24996 F20101122_AAASEF milleson_m_Page_099.QC.jpg
ec548ec67b6cd4de095a772086c105c0
88830212ca8c18d21ced6b878d7db58bb9fd2d2d
26619 F20101122_AAASDR milleson_m_Page_003.jp2
2872d6cc39825588615e54bc5e52dc53
18e8ad186799ca969f1cd44dd0bbbd9e2ee2bc73
704452 F20101122_AAASEG milleson_m_Page_046.jp2
4b9367d2f59260a674cbf7e7a213ae2d
50cc07ed78901aa03131a5adce98baa07570683a
22888 F20101122_AAASDS milleson_m_Page_001.jpg
dc47dc46edb94e7677fd27634878428c
f7bafa3a0316e83d168bab5df38d03e0273e9406
F20101122_AAASEH milleson_m_Page_013.tif
743ac8023eb1a64dfe966b9c0a10f386
6e8150df38a3761eb4ecad615d99e78c35e57f70
638088 F20101122_AAASDT milleson_m_Page_019.jp2
ccaa55c51412e711b23c721c143f1fd0
ed34f84f11bd0aa916fdc947c8a4eddd9a852b64
23314 F20101122_AAASEI milleson_m_Page_017.QC.jpg
fe6469f687eb2ba551405b18a2e68c6b
00c65aab209b4c6b212575ac991e24285edf4836
4063 F20101122_AAASDU milleson_m_Page_063thm.jpg
a27a6c46a2506486b82b93276827fce2
ac0b681e34083a340973b99d44154d28f1cad86d
F20101122_AAASEJ milleson_m_Page_043.tif
b4cbc109247639929f728807299542f6
c65e663dee30b7aa0fb91d6809e5b767c1950da2
42918 F20101122_AAASEK milleson_m_Page_075.pro
cb35354af4ae1437e885a19ff3576f01
1ef5c66eab6b58dc6318a91c1071621642de8bbf
109876 F20101122_AAASDV milleson_m_Page_009.jpg
0875a187117723baf201c38712368d0d
4307c3b2952b1b8a85315884e9c592cd0e0f5c0b
1784 F20101122_AAASEL milleson_m_Page_092.txt
c7471cd403c5e91260137f9923f7946f
17f6944a3013f5eec34fffa54fd1dcb7a27aacfe
6761 F20101122_AAASDW milleson_m_Page_020thm.jpg
b4dd1c554c1bd047623f2174e3607b74
24a164dcad7228a15425c720e9685af94649714d
1820 F20101122_AAASEM milleson_m_Page_073.txt
92a6a3c03350c2a84b846eba4fcffdc3
bf18b9469516c3dd9853b4c05ed69535cb9e98f2
72083 F20101122_AAASDX milleson_m_Page_043.jpg
00834a333831f3f7906434cabb45a9e3
6f5a6eee7bc51609a9f684674e7367b4b31851fe
114481 F20101122_AAASFA milleson_m_Page_076.jp2
38b7cc9531c6cc30f4b471d1d6ffff72
0aaff9b18a9dedb3fd04e224413d18bd4f45de49
1709 F20101122_AAASEN milleson_m_Page_011.txt
6c3c8fa3de237cd86ba37ce11bf43b1b
65cbc1f37441add2df9307dfbcb91ce34b162116
4676 F20101122_AAASDY milleson_m_Page_005.txt
c4d3882e424551897a2b2cef627854d7
7c549e0b81bacd316828ba73a6cc85a426c4d890
33137 F20101122_AAASFB milleson_m_Page_090.pro
4548c43ee9ac804672c3733bb378d383
4a7e660235244b00f24365b647c7cbd613207826
F20101122_AAASEO milleson_m_Page_080.tif
d54fded87dbb827b563ec90730936f49
6233876c2aaed9ef8ab90d27c3fb9b1707ac2b48
6877 F20101122_AAASDZ milleson_m_Page_104thm.jpg
f0a431277337984c88b8983cde565094
fe71cc444c5c38afb102f7b75527371a5324e370
3435 F20101122_AAASFC milleson_m_Page_086.jp2
815e09cbfebb6750a100d1aaffcc5397
897961b2f02a170ac7d3cbc4d46f0ac8b3170781
53570 F20101122_AAASEP milleson_m_Page_068.jpg
13fe6cea9423bd9b0978b5a11ac9f402
1334012ffdb03d5aa22c1a8d4a7aac9f4a35bee2
66498 F20101122_AAASFD milleson_m_Page_049.jpg
13f1ad314acf006432d978991116bb05
e1e206c690525f2b705f16e2afcdd453f8454cef
F20101122_AAASEQ milleson_m_Page_103.tif
99fbe6333c0010fb165f13042de7ca38
8e9e0582bfea3e01715954c83c0a62f548c09483
5791 F20101122_AAASFE milleson_m_Page_055thm.jpg
f7d96f592347eab86f3b66ecb7fd5c6f
32b4fbe7cf5316b095f124d7a494596d309b1382
17410 F20101122_AAASER milleson_m_Page_037.QC.jpg
759c09d7ad9ca5113b06f9ff7ea97787
6aafaa992051beec631863849c326ab19bc8d0f2
101276 F20101122_AAASFF milleson_m_Page_026.jp2
82116f76e0eccbd5a999dbf9fc8d66d6
c1068eaafdcb93c269d868298d3712633be3ddbc
6587 F20101122_AAASES milleson_m_Page_078thm.jpg
06859407402dd570f7cb8745890173fb
e063b488d85997e4c8ac86458f5e728707fa3a53
19923 F20101122_AAASFG milleson_m_Page_087.pro
d29f844daf9e7517a290708f7d0e1163
00514cb78871617cbccae9d72583f01403d30d78
F20101122_AAASET milleson_m_Page_002.tif
6df88b54305c1ad1c439a8b7e6de25ae
3ccc63d752b2fe472c88e216fbef87ec8e63a399
5725 F20101122_AAASFH milleson_m_Page_075thm.jpg
462459c490e8b98f8a5bed382ace099a
a9f7732ccdc2337b0f5e4c573b22eee6863a478e
7358 F20101122_AAASEU milleson_m_Page_001.QC.jpg
0b7af8fcf1a2b9efe00848f1f5bffed3
39ada0ee45774a445e3c587f8988d7171f7d4030
34282 F20101122_AAASFI milleson_m_Page_091.pro
29e5610797fb02efabe0d79b7a00e498
a9578a5f69551d49ef5297d6646d78d237cd90d7
15597 F20101122_AAASEV milleson_m_Page_019.QC.jpg
d9b0ca31485651b1e2d12a95a88e9f86
26d1fdcd41e756a4e0cc712bb4aa6eb966184ecb
48880 F20101122_AAASFJ milleson_m_Page_081.pro
86a8e658bc24b2786b06db39390afb69
9a2363509e8339efaf29c31a483d0537c20946dc
68418 F20101122_AAASFK milleson_m_Page_050.jpg
b15952896903a8c02aa514835e31fea6
c56fd342eb8e027bab3f72a714c4b5a1374f1322
6440 F20101122_AAASEW milleson_m_Page_041thm.jpg
56e2784b12e6d18fb6fee866b49c92c8
371bbfe047dd916e458c98b5c60b86bd21044568
6647 F20101122_AAASFL milleson_m_Page_025thm.jpg
df7b6a1b2382574b6f3039dedb4dfea5
fc88ddf7216a51b333843a2503e5cffefb8a1b83
F20101122_AAASEX milleson_m_Page_082.tif
8e48937b8025f2ca8272c9d3acf32119
b663494ad8a5e91afe09f1ee955a6f6fd83dbb88
160991 F20101122_AAASGA UFE0012043_00001.xml
0bf891670a4145c0392e53aa82750b9b
15f0d53e2cf21e532c162796eddbdc55c05fb9a9
116460 F20101122_AAASFM milleson_m_Page_014.jp2
8cd02924b933a1a2dd11d8df2b965bd5
be22fcdf1f0652f80806c9558f0eac1d3367530a
52570 F20101122_AAASEY milleson_m_Page_078.pro
f410f877f73d0b5c89e9d5b6e71b0feb
5a967899857ef502649ff22675a05b8d4a2896de
F20101122_AAASFN milleson_m_Page_010.jpg
caf9a018b00c2b1fe9ebe778df42f7bc
ba3177c2a0ca730c0e314294277f501b8ae05c80
24610 F20101122_AAASEZ milleson_m_Page_103.QC.jpg
d6aebb59f3fac198e2c1e4905544af78
69ce40d0d888bc14ebbe2a504537d53f99e826c9
F20101122_AAASFO milleson_m_Page_094.tif
3ba9f7a7ea18515123a00700960fee99
97ab7dc05c0f57f951898b5f89884477955cb583
10425 F20101122_AAASGD milleson_m_Page_002.jpg
373f117c97dd536d4b9222da443a7a1c
c4b09f25423d125a7f0cf65835fa1c4d21639824
4807 F20101122_AAASFP milleson_m_Page_084thm.jpg
9a8464c90c18e413e880099e44c23a41
e87282ba1818949adf1f173bf8727405872b634f
23038 F20101122_AAASGE milleson_m_Page_003.jpg
0e052be501f7657e8191904a7e6a8e83
0812f568426e3149770623d95b8f67a2422f2a22
2442 F20101122_AAASFQ milleson_m_Page_101.txt
95b0b886300fb3d7f3c51c8e7c0fb603
a0f431a647dc05ae08d452da56783c790cf466c4
72471 F20101122_AAASGF milleson_m_Page_004.jpg
07292da12b106aff247edcfa793b2e30
5e0275ae758a9f05664049b5e0ca4f8bcd4a126f
1491 F20101122_AAASFR milleson_m_Page_082.txt
834204a83f3f91375eda48ee8b08f802
d12e7aa10424a0bbd723be43a6a9027127e06850
97727 F20101122_AAASGG milleson_m_Page_005.jpg
8a776e3fad2baf0f6ec77fddb23429a6
0a53292fa6ea1558a496676b4b584c867999762d
34928 F20101122_AAASFS milleson_m_Page_055.pro
c7745d745f5686824c17e4f427f5aabb
f0eef888c0baa8c47759dfbeb23a159da5ecdb5f
29761 F20101122_AAASGH milleson_m_Page_006.jpg
fdce060ec27c4136c719e7e38fa11e04
74eed66ebe438b580b0144833c4b052276461df7
61798 F20101122_AAASFT milleson_m_Page_039.jpg
b85e37154d54493d16298f11ca133193
b4939483154df74ba14a116f2682629bb3128b5e
28498 F20101122_AAASGI milleson_m_Page_007.jpg
c19750324ef05b38dd0a13fb83d7c21a
04c657682432d23e4c06f4367b5a726075d279cd
17081 F20101122_AAASFU milleson_m_Page_091.QC.jpg
6379da7d1f582a5b2cfbeaa27b64493d
2286a8e7d4276bc092a5a3580c91aef1b58747b7
58299 F20101122_AAASGJ milleson_m_Page_011.jpg
c8a4481f73ffd5bd8005519b355d14e3
cdff7f9afab5a7cecc094aa70b63d9717081eaa0
16527 F20101122_AAASFV milleson_m_Page_033.QC.jpg
927042564c15a0d5004674f3cf1b2852
8bcac7d54200cda61ab5bd46a3b504a1d8a77f9b
66434 F20101122_AAASGK milleson_m_Page_012.jpg
3203826340527c9b41e7ec230428fba5
068266172aa72a41779712e22a39459505cb1bae
63398 F20101122_AAASFW milleson_m_Page_097.pro
2ba27474d1ca32b2e26c349a2bc867cf
808e57185b0d6b76469ef556f29461f84f4d316c
64384 F20101122_AAASGL milleson_m_Page_013.jpg
6f3657aefc3792b78686fb7a7cbfbc7b
b9cd87cd6adafc10724f04b73fece197eff42ca6
71884 F20101122_AAASHA milleson_m_Page_028.jpg
1b6f3829470a54183f2f53ba92fda22f
e027b4849036c9884f001082910d603133a41cc2
75104 F20101122_AAASGM milleson_m_Page_014.jpg
4b22b02d12a65ba4ea1ab958ea79cbbc
dea47c0b60b18baf9e3fecb7a74c694f688a4bfd
55700 F20101122_AAASFX milleson_m_Page_037.jpg
12fa74e2e23eafc5780a076f402694fe
7d58c90d3a7aeb0dfa6f210e56659b1886029c88
66084 F20101122_AAASHB milleson_m_Page_029.jpg
cae5456d91ef2cf10c1128d728f848f0
6a46022adba8feb85b78d6f7dadc23a1b57cdf68
72737 F20101122_AAASGN milleson_m_Page_015.jpg
aa657b204b834bccc2b2fe07c30f53cb
7511bca24b5fa76b217a038287355f71cb7ca527
17766 F20101122_AAASFY milleson_m_Page_090.QC.jpg
fc2f79769f206ee8d9065cc9e01cd162
7c215bf07a3ab654b12ef4c43360c32a35909e8a
45505 F20101122_AAASHC milleson_m_Page_030.jpg
2f03cb9ced4fe73a44e25ed26386a23a
7b11eaf2f890d3b4a5dcfa6ca4f60f94b32610d5
71649 F20101122_AAASGO milleson_m_Page_016.jpg
7be54f01ef5927ae19ea509dd0b0e362
a269dee48c8bbcdfaab87334d1a2040290f3c596
619932 F20101122_AAASFZ milleson_m_Page_007.jp2
bc6f1c174c3a2f8455b441e4e4cd1960
0cfc7803c304fd3239fc97e4c5d6c2244b2115ab
53113 F20101122_AAASHD milleson_m_Page_031.jpg
662602454996820bc9e14ce727a206b2
a0883bf76276a652f6c63aa60fc9fae006df2c7c
71303 F20101122_AAASGP milleson_m_Page_017.jpg
de7bbcfbde51b4395207d3da823de67c
999c1f077708da94b8bc4cb21f915d53c93f873d
45473 F20101122_AAASHE milleson_m_Page_032.jpg
0e6c380ab85b855aafad35c315196089
d0f96461569fa952fb8b43926828de70a07c0c23
72243 F20101122_AAASGQ milleson_m_Page_018.jpg
88b19b0442e105d02665564d195b4e3d
3c696e1fdebffc1bce5e64db28f950b3b94aac35
47563 F20101122_AAASHF milleson_m_Page_033.jpg
6f218dbcd0101f10b5d93f8d2368333b
d56cf628085d57fc047a136aac6124c10383acbb
53578 F20101122_AAASGR milleson_m_Page_019.jpg
bd8e09f54fcbd268ff092db794be1b45
ece65e48dfe1feb3e615f224b5a70cabddf1ac7d
57268 F20101122_AAASHG milleson_m_Page_034.jpg
6b4d8920b5e55d573abf1882dc083a81
e9ef5dd90300b3f9ccb2f47ecfc44fcda0e9ad6e
77024 F20101122_AAASGS milleson_m_Page_020.jpg
2295067fe42c88eeb7bdc0109a7db9c1
4a14ad6a119a6a5d6a4e4d554eff324b275a6ae2
54873 F20101122_AAASHH milleson_m_Page_035.jpg
982a4c632ae34cd37dad847b1dd0c250
bc0cf0653c5e12c96efd2a846b1ecce7c8577242
72697 F20101122_AAASGT milleson_m_Page_021.jpg
45b777dd45070dccc4db55181384b6d4
d425cf21915f342f5bc6ee538b1ab1dc3e406bbf
71838 F20101122_AAASHI milleson_m_Page_036.jpg
fd59a14c0fa1268a73a18daa40919dac
4ecb35908ffe8a9fa7be3c96afb626e45ed64ae9
72689 F20101122_AAASGU milleson_m_Page_022.jpg
6319914d63b125d77043e077e8c6f549
da25e93cf33c6a113edfdcafcf1c8829db569653
73295 F20101122_AAASHJ milleson_m_Page_038.jpg
1b1ca33e04c36df6b46917fd247d7e65
0fc54ffa07ac6def51b140df51a77e5795a0af9b
72278 F20101122_AAASGV milleson_m_Page_023.jpg
ab869dc4236091a6f7aa8baf548a917f
494d30d7e4bbdb54cec716c0bac5b7d7197d2892
66098 F20101122_AAASHK milleson_m_Page_040.jpg
f1ce4a50fa6944a7dec8df5c322e1adf
867a70838d39993eeb87ed777abd4b5808abb70f
66456 F20101122_AAASGW milleson_m_Page_024.jpg
5d93a270e4ef2a2da59a8463ba0b5ce4
6cb932f8ebce9a056bf18dc5669c48b594888654
72644 F20101122_AAASHL milleson_m_Page_041.jpg
05f9477d65440bece6dabae32d15a062
79da1e319f968a374a1530494fe631645689c98b
72865 F20101122_AAASGX milleson_m_Page_025.jpg
6d3447f5702147e96e22e20216f207fc
db80824cb913a8dfdd1d01600fbf94e74c991314
74369 F20101122_AAASHM milleson_m_Page_042.jpg
b1807b54f5fc4e95f344bb82d03cf3f9
dd69d469ad34698ef40b235d2f7a6a6561412510
47254 F20101122_AAASIA milleson_m_Page_059.jpg
8ada5a8adddaeea8ffcde723521ab957
98ce2d379c59fcf10eab33ea2b99619656f3eed6
73505 F20101122_AAASHN milleson_m_Page_044.jpg
9d63220f355901a03738f73faa49b529
0d2b8de97b6004e0c8aa301b86ef05cc33be080a
66396 F20101122_AAASGY milleson_m_Page_026.jpg
9ad8863354fcee99f858ddb6427f0145
5aad1375b62636ab4be9c51b976b4caa267236df
61538 F20101122_AAASIB milleson_m_Page_060.jpg
619382f775b8d3868b73f0c0dd65b461
efdd9ccfb379905e68720181d47bb8d9d6a50493
69569 F20101122_AAASHO milleson_m_Page_045.jpg
c39415dacbac1cda2fe20b99edb5b95d
70caf40ec8810e3c7c45576a4377901f3c41356f
72381 F20101122_AAASGZ milleson_m_Page_027.jpg
ff8bf144b2818829815225e7de9fc2cc
2ed3555b632e5b9913e79c36fc2005f6b7adfa1e
56943 F20101122_AAASIC milleson_m_Page_061.jpg
6815324d522bfb6b6d9205067c15bf97
b6069fb6197ae7e2d3520368b7af1dff7ee5396c
50220 F20101122_AAASHP milleson_m_Page_046.jpg
a14f43b1eef15d16a32b49ab33a61aa3
5304cac8a6e407b143a691255d3251789da2df21
54550 F20101122_AAASID milleson_m_Page_062.jpg
17d87fba26e7cc7a823c804ee92e2168
b4874582595c2ae838234bdd2f3d21d83e0d9223
67337 F20101122_AAASHQ milleson_m_Page_047.jpg
24efe6009185540d0406df289771edc5
b20e9054c30d842c1089e3e28f3744e60cee19c8
48055 F20101122_AAASIE milleson_m_Page_063.jpg
d87bf71582daff101f0faaff60f192e2
834079a05adfcefb10c782536b5377e732e9a75a
70859 F20101122_AAASHR milleson_m_Page_048.jpg
0f3b73dae522718647e5abb524a17afa
baec5ec9eb89b3ff532f7fa0e31060e23f108160
46070 F20101122_AAASIF milleson_m_Page_064.jpg
b7df5870e6752b3cacff123e902d3bf6
d2e04b06eb0fab975b34d1cdd4cbd683285ae536
73662 F20101122_AAASHS milleson_m_Page_051.jpg
40d9d8bcd6d488a49c5183cf5be92ea6
b111a6b6b6affbf9e190a3bf145c3aedeca9e938
58069 F20101122_AAASIG milleson_m_Page_065.jpg
5d3ccf28ca7f144faff79a956e89d86c
f0ccd023908fc9f32729ed541c97aec7b4a09ef8
71767 F20101122_AAASHT milleson_m_Page_052.jpg
1ef02cb0e5480c343cdc6146f9a4626c
5e68f5e260c693a85ab0b496ba67e21acb4b6420
59805 F20101122_AAASIH milleson_m_Page_066.jpg
492400f031775dced869ec046ff0b10f
7ffe38ad838a5201cf4223d6025c70e0e95bbee9
62470 F20101122_AAASII milleson_m_Page_067.jpg
758fdf257d5e1a82fb7aad60f342ca2e
e7bd60cd4cb5d1fe1b3e74b02d6d9cdbefd2f3c8
67981 F20101122_AAASHU milleson_m_Page_053.jpg
032e62a433d86f99d398f1910c613515
4c71fca2b9f2f155a8b84351baac6384776a899a
58294 F20101122_AAASIJ milleson_m_Page_069.jpg
1cc67b4cc982195a4269d7ceef49dcc7
aade49e970053fbc547c7e971d9f128d63f1516b
66481 F20101122_AAASHV milleson_m_Page_054.jpg
ba48728cdd0fe752b9aa85dd5b453115
304c441320fc83dca4b55eceb6ce1ef22b7b6825
59227 F20101122_AAASIK milleson_m_Page_070.jpg
d4f167f114a4fad74baf57c93dd04bc2
cea53692796c74a2543aae8869449ab58db465dd
63676 F20101122_AAASHW milleson_m_Page_055.jpg
d41a30321aef7295e62505517f6a1e63
08021fb68815e7e7999d716d771828b5e39e0d33
53377 F20101122_AAASIL milleson_m_Page_071.jpg
462bb6c2eb6deb5c3e252dfeb09778c8
e9040f563697352967f845695c642d380fb7338a
61437 F20101122_AAASHX milleson_m_Page_056.jpg
392a1a52e09d8a2ea3d3dbbd88f8f00d
a430676ee48e85ee2dc998107cba503867c7de05
36943 F20101122_AAASJA milleson_m_Page_087.jpg
9f1aaa71bd58e22fd86de545ca993ca8
5b44da15be3ba6c748ab2e438b248581378fc62f
65613 F20101122_AAASIM milleson_m_Page_072.jpg
0e78cc2a5f60c7142146651cdcf15413
b525da2d9852603b38af8203f3fb421a0d97bc4c
45852 F20101122_AAASHY milleson_m_Page_057.jpg
7e9157f4e1b2a7c5c5bfda19cca3791b
8ae2008f71ab3d13b6c5ca79e00895b52d15825a
63531 F20101122_AAASJB milleson_m_Page_088.jpg
f1ae10fe718dfc39f108dbcdc024c894
074f0c140e026dd34a4ac1d8af6aada5e4606ba1
65782 F20101122_AAASIN milleson_m_Page_073.jpg
76a177db8eaac51ad6947b7f845cd959
e621ebd83d2f06250d5b97e9baff230ccb038a24
58845 F20101122_AAASJC milleson_m_Page_089.jpg
03cbbba1945c90aa67ce0cfdb3635774
afb9f864f3774aaf9c9225cf5837a5cfeedf01d3
74956 F20101122_AAASIO milleson_m_Page_074.jpg
9648a2860658d3f5c5ef1d5df63e88f8
2313bbef6f216b47878efe3733dcbc75835b1f9b
F20101122_AAASHZ milleson_m_Page_058.jpg
a415a7682e299d16bdfde58ddb939ef5
5688cae56982ed9c5ee6edf38068cd20dc864e6f
57892 F20101122_AAASJD milleson_m_Page_090.jpg
a72fe8023be77bef2808eb49ebef4356
304174cba84f996473ae92b1efcaba6b33c29f24
68410 F20101122_AAASIP milleson_m_Page_075.jpg
b3e6d7b945e05a2a9dc1ce0ca652bd5b
e2bcc44fef0e8bb5e8bf90401106ca3016b5eacb
55626 F20101122_AAASJE milleson_m_Page_091.jpg
c90a225521d58241e06025def905c72b
9def62a6b1574eb16c0e1b27b297e642fbf176ec
74345 F20101122_AAASIQ milleson_m_Page_076.jpg
c4749e367616e4d5f3c3b383885b30da
e2603988886d9d2bd4e9fb23c3d0b96e92c7fec6
66153 F20101122_AAASJF milleson_m_Page_092.jpg
256c771333cd061222891d9fbbd38b46
3299fe15b3cbbc37ace882888f1857c89453d1be
72443 F20101122_AAASIR milleson_m_Page_077.jpg
58d8309d7ce24dc719dca6801eb98854
677e9fb885e87bb810fa39b75ed4764ed054758b
59489 F20101122_AAASJG milleson_m_Page_093.jpg
2aa6edaab8d7e1cd175be08307f65956
8179ac09aafdfe1feb3d75094901dc8791d79acc
60459 F20101122_AAASIS milleson_m_Page_079.jpg
2e3f08fd793e09de4fd97a805e1acea3
a9616bcc6fb0d04122f2dce6d0ef4c7f3c085c16
85583 F20101122_AAASJH milleson_m_Page_094.jpg
1b06bfb3a32fd73152735f37ef2372bd
69e508f7bc7a15dc3914b93a9286c5adcc024f35
68783 F20101122_AAASIT milleson_m_Page_080.jpg
60ecd5fe6742683d8007ef018a766560
d0bbaf22de55aa2164aef5588378e9354c4f2a30
61011 F20101122_AAASJI milleson_m_Page_095.jpg
6744960bafd408dcf3b6694e7823b825
bc2edc166783f6d4e3f97d9c6800489ddd57a540
68839 F20101122_AAASIU milleson_m_Page_081.jpg
7e301a17cbf5363d6265dde1936b31c0
965a685c15cf6f1bf8b3f36a184e2b6accf9407c
81804 F20101122_AAASJJ milleson_m_Page_096.jpg
c0a4f7e1a3d2cc807267893d6e2d1ff2
78f9b8628b323bea8b344a3d55fd887dc9feae2d
54577 F20101122_AAASIV milleson_m_Page_082.jpg
9953f32fa0e8f1b6ccb736f426b753b0
09f7d9b9040930feb98c6422d11abdb82928cf3d
97478 F20101122_AAASJK milleson_m_Page_097.jpg
cb3c498d68adb7127fcc37703756a728
ef8b0f19c953e30e22bec9aa3188bb2ccea8fd4e
41046 F20101122_AAASIW milleson_m_Page_083.jpg
48ef928ae0029f8c5951f1da7ce404fe
2fff2af0ac653d72dbef84a1118c1e7b2815c26e
90226 F20101122_AAASJL milleson_m_Page_098.jpg
4d52266c3380d43aabdb83a683d65163
7323a83511b93eb16bedcbf8ae404a602f2c1b17
49010 F20101122_AAASIX milleson_m_Page_084.jpg
496a7095cd4d598434164c4f2aef977f
8f06d222030649e58e91c7051381767167f7c927
88406 F20101122_AAASJM milleson_m_Page_099.jpg
475662a882fe69120cdc3935af1a303e
dee9ed1a225a434fe1e5ed9078f1e85eb1d955ac
53159 F20101122_AAASIY milleson_m_Page_085.jpg
bdb43f1b3b9ab099b9549ccc166993a1
50582ff91fb6ef2bb456d8a814741c093fe8d80b
1051985 F20101122_AAASKA milleson_m_Page_009.jp2
2aa93f306461c18c743c3c64786c9f26
4fd066cadfcfbec13e8604944f93e2b794284591
84253 F20101122_AAASJN milleson_m_Page_100.jpg
ba16d9a01f9e8cc864be7a051c20c5ce
4d504a2e9902617739987469e7c76eb6e0c7715d
8960 F20101122_AAASIZ milleson_m_Page_086.jpg
f126f8c76c336b1b3addb5cc3b7e4812
93b7725dd788556f4a3853dd16be6a1b71290cc1
1051939 F20101122_AAASKB milleson_m_Page_010.jp2
f130fb68f1470a9546beea76a288769b
13df4ece4cdc8dfac68d7c2a114075ec0521f960
83634 F20101122_AAASJO milleson_m_Page_101.jpg
9ad7ab20e6617cb9a35daa5fe9dca6f1
9f67067d3e71f5618a75c4616cd1a1e518c5ff65
85109 F20101122_AAASKC milleson_m_Page_011.jp2
d8a0ffff1d7bde7689eaf1196d02433d
fa7aaf02c6fe5cc4e38e0a93ceca172a2192d495
85538 F20101122_AAASJP milleson_m_Page_102.jpg
5916f2ba271c88080e8add71af762036
82497c260be763777ffaac86e68ce9abf85233e9
102187 F20101122_AAASKD milleson_m_Page_012.jp2
641f1ed60ccc23f82f334a1594be801b
96d802b3286990d857dc3adb52e04778b0d9bd77
80610 F20101122_AAASJQ milleson_m_Page_103.jpg
f5481d279f0729c51a021494b6d290a5
a02dcce8abb1798cc637408cb88307cc6aaedeab
96839 F20101122_AAASKE milleson_m_Page_013.jp2
5c869ee9d51e28adc3f1d0ffbe9389a1
907422587c2477b3fc1dc762d1fe5a5d680342d1
84935 F20101122_AAASJR milleson_m_Page_104.jpg
cdc88250a00304d8754a1567879f5c72
27d103f7c2262d088296d2036c7210c196a5ab40
111219 F20101122_AAASKF milleson_m_Page_015.jp2
9959b923bd18b6e40eb502496b21b7f7
74607884511c3dd4dec2e9f3c3e102d4e024357e
62131 F20101122_AAASJS milleson_m_Page_105.jpg
bd3abb8a93e499b6bea64646844e18d4
25ab9a99e1c84866e10c00fd41a78f7a608d649c
109843 F20101122_AAASKG milleson_m_Page_016.jp2
33463b4b007df4e80ece550922f9f648
b97dd2909f6fd232c87403ce2ec8acd95adc389b
43477 F20101122_AAASJT milleson_m_Page_106.jpg
380279537434f5610849ff482afffaad
0601225c5be86badfe30de7463e86baed87f7149
108791 F20101122_AAASKH milleson_m_Page_017.jp2
d5826fca331979c9db6667a9e9a8cde0
f3e7a80bb60c8e8753d1b92e60baaf5efc9a02c6
23547 F20101122_AAASJU milleson_m_Page_001.jp2
f0412ce815de1d0f64bd8d64652d8288
93ada2e9d038505d265c01affa443da42a3e01f9
109879 F20101122_AAASKI milleson_m_Page_018.jp2
8e52984e88e3eae004c30df3b01181a0
bb6583568246386ccbe4c55031ddff4aa18b6fcb
5869 F20101122_AAASJV milleson_m_Page_002.jp2
4c8429503560e3fe795b2dbd5211af1d
3b000a7b8531031099afbacca2e14ffd300a6d26
118666 F20101122_AAASKJ milleson_m_Page_020.jp2
269f5f43163f013d5b83cb7ef01c2546
d08c5c713d7d971be97ec00afa931af5e3a69e09
1051986 F20101122_AAASJW milleson_m_Page_004.jp2
63575d52d72b3f49b4c1ac3191ea6753
7594c9971fc64043dc7b9f6c454b1b605c26d73a
112313 F20101122_AAASKK milleson_m_Page_021.jp2
f8f3f3e6b61c57e29529000013677011
e965fa2efd73f33553c869388db991eb772a25b6
1051979 F20101122_AAASJX milleson_m_Page_005.jp2
beee4c629e50fa325c9b7cb6945972b3
048d02e2b621e8bf98daa2c50413f1d7f494ab8e
111994 F20101122_AAASKL milleson_m_Page_022.jp2
dcce8d68828fd2ae53bbaf3e6a7fe634
2031574bcf70d9fae50a7ef7e134134d81b78ac9
658756 F20101122_AAASJY milleson_m_Page_006.jp2
d15af3a52cde826fde805e172ba158e6
17216cfca675633c262a7637f3aafba57fb8c4de
113109 F20101122_AAASLA milleson_m_Page_038.jp2
0bc39d35176ff30838f51f03afda77ef
6e24118399cce38c3639fd6c9b75a6d27d85e082
110025 F20101122_AAASKM milleson_m_Page_023.jp2
eec536e97c451d88e506b3a51dba4076
91cee1e54243e41de1aebeb9f27d77bf89a37247
1051967 F20101122_AAASJZ milleson_m_Page_008.jp2
685b0b0de481b38bce6293188192d27d
f857aa0e0ee5546c1ab9893033406b08680c80e6
92218 F20101122_AAASLB milleson_m_Page_039.jp2
6d6d7114e2abf8ede02a9652cd76c275
21574f306c0f1c1ccb48b2694e19b3f4f1c446bd
101512 F20101122_AAASKN milleson_m_Page_024.jp2
33d9ac77740ed97cce216bd3804996e9
f8ab24c4277e86d4343713b4b0fbc78805dc8a4c
96579 F20101122_AAASLC milleson_m_Page_040.jp2
c0556b010e60b730a7c237dbf056b810
d7f7f4eee7b79f9096ae1466c525405b5b9b569b
111628 F20101122_AAASKO milleson_m_Page_025.jp2
dcc76e1ee1a1c628b1dee80c77f18963
5fb50a607c8bddfa9b111ef3ebb3adf107ce072d
113543 F20101122_AAASLD milleson_m_Page_041.jp2
7495a16e2edfad1c0be6fbb02ed4f98c
4f0eb37f008210450a2892e5ebe1cd3e6342e541
108750 F20101122_AAASKP milleson_m_Page_027.jp2
139f20caba95dac5fd4afa17a7da27ff
3a0e2c8767ef16159ebd7255c4af0e3bb346f197
113194 F20101122_AAASLE milleson_m_Page_042.jp2
3f85bfb6fff6e62f8c095c90078731ac
d79ab1da15c2bc1c5b74380542cffb1597d92e67
110131 F20101122_AAASKQ milleson_m_Page_028.jp2
b89b4eb9f68c41ad5c1120bdb7dade49
cc23549904fd525d5d2aa77b66bda7c6cbc16c11
110963 F20101122_AAASLF milleson_m_Page_043.jp2
fb470e81b872d330b9df3448f501d540
5fbf83685451f8d31bfecc5c24673ed7b21ffebc
99094 F20101122_AAASKR milleson_m_Page_029.jp2
017af30708b291f737b7a5cb5a7ad5ff
b3d6551eb71605ff3db3438f118733440b805eb2
111578 F20101122_AAASLG milleson_m_Page_044.jp2
385c8a804e88a6c858eb134ee259e7b7
42802be04548a1df2a9ea5f610a7b40c472994c3
593306 F20101122_AAASKS milleson_m_Page_030.jp2
d706bee6b511ff86bf42f07103023cbd
0571db58a1e87e8018b5e4255befa14723430be5
106992 F20101122_AAASLH milleson_m_Page_045.jp2
1f9022c5b0882aa74133a9e9b842762a
243b20147d3e1a99a99778b32bc24f554e604233
701779 F20101122_AAASKT milleson_m_Page_031.jp2
d099c6cda92586c6e1caf917d88434bd
8311f1aef7f85c3a230d619f621ea670b069e2a8
100567 F20101122_AAASLI milleson_m_Page_047.jp2
8e325320f7c793673d4371ac019007a1
82d92fc8e9d8874fb04db68816287e6827187137
608888 F20101122_AAASKU milleson_m_Page_032.jp2
85c7549a3df3813c6daf713b485e7fcc
cc2aef617869fb8f9f8cea286c3dfb332878312e
109324 F20101122_AAASLJ milleson_m_Page_048.jp2
6cdf27160d8d1dfcac89341cfab23a6c
9390e2ddc26f68182a3b8adebe578d7b594df82c
632164 F20101122_AAASKV milleson_m_Page_033.jp2
c12cf59621b74a02885dd2f7a2d2a395
d64b1a68001fac069b435e61ae463d8d11e30e60
101685 F20101122_AAASLK milleson_m_Page_049.jp2
4b4c0646ac86de339e73671eacb16c33
096618a19c72d862d90ec3da3bbe2284eb7ae7b9
735519 F20101122_AAASKW milleson_m_Page_034.jp2
0cbf4945f0443bc62a008347bbea074e
151c813bfd4962eecf03fea7b090afed096a9580
104289 F20101122_AAASLL milleson_m_Page_050.jp2
8e7680ecdc7c5c27066d79752cbe694b
ee01dc07b9c753aac32f757d595c37b0f15afb48
768305 F20101122_AAASKX milleson_m_Page_035.jp2
56b974b9c00db310f22bee027ec97670
c732651d30d308f8623a06a9f8e9e4f767f0e815
799170 F20101122_AAASMA milleson_m_Page_065.jp2
9e40fb9a1d6a79f3ffaf663fc97625d5
c4bdc21c59c13fbaa989db3fe82793936975e090
111509 F20101122_AAASLM milleson_m_Page_051.jp2
cd316c3ff5dac3014ea0944435a5d551
12b8e74e7e67949bbf168ec0e4c07f811534046b
111306 F20101122_AAASKY milleson_m_Page_036.jp2
6da0445528a7114e4aba5d46d267b464
51289ce8ac795124275e8121fbcfd2b35dedb757
830574 F20101122_AAASMB milleson_m_Page_066.jp2
2f6a96475ac0a76923fa6b0a57d993bc
570cec80db69d17df1f21e4e83af6865f60b45df
108976 F20101122_AAASLN milleson_m_Page_052.jp2
fd9765195e8720e6e2de45cbc9808da7
2c33d1483e97fa11dab6cccafd1be981b8f6af6a
765197 F20101122_AAASKZ milleson_m_Page_037.jp2
d54aab5e4339140d84430adc72eae923
ea98022c913aca467cc2c0f7f4cc38dde868b5f8
874619 F20101122_AAASMC milleson_m_Page_067.jp2
756764881dde79bae977b65a9b065cef
fa54d35e4889fe7ae6a282b72c731714ad6a616d
103030 F20101122_AAASLO milleson_m_Page_053.jp2
b8b0ab61e5adbca908c8e040123c6227
fac1d754a2eb97297c81b487f5421b22b819523a
718446 F20101122_AAASMD milleson_m_Page_068.jp2
2b184be6f14b6f0dd06af2bc64e4e6ee
bd3fa3c34280dc316c958cb24fa33dd8a8b2e809
99138 F20101122_AAASLP milleson_m_Page_054.jp2
6374bf0b41f8619176d2f7ff59a179b2
d055113642b928e7d24edd3e61a185246924dd2c
800063 F20101122_AAASME milleson_m_Page_069.jp2
70207d9e2ffcd568905c9b941bb43c84
41366c22e1a19657cb21448e8d4ffcea40b5e991
843974 F20101122_AAASLQ milleson_m_Page_055.jp2
3f25661b946214c255665d926a9c2023
28bda4ea0f82cf4ccb8102eb0853a5c199218a2e
796048 F20101122_AAASMF milleson_m_Page_070.jp2
39b2e25ab75039532d7b11cf416882a6
bc4bc8facadaf9fd45babc397477bb7b9279012e
90045 F20101122_AAASLR milleson_m_Page_056.jp2
08ed626d5ed5e9a68dac3a116c3ccd76
9aa22e276ac752fa6cdd1ae8526ac3b550c8b1d4
699309 F20101122_AAASMG milleson_m_Page_071.jp2
31117d866e94b6df02f82ef59934ea28
6e186abe0e2374a90c73e97a63f67223b8e403ba
605654 F20101122_AAASLS milleson_m_Page_057.jp2
232452b8051dbdf5820db150feb4b2d4
1bd5dba33f5e43b5764f58f2fcaa2c1288862f45
94918 F20101122_AAASMH milleson_m_Page_072.jp2
b90a49881ce82469e1cfc7088226cc72
dc01e5a4d1ed55f139d15d997ae25e147701a86b
78381 F20101122_AAASLT milleson_m_Page_058.jp2
c883c5b0b2885cef0a5a2aa6509da682
63ad4a0e2b04150f665659f038a901a3f636816c
946342 F20101122_AAASMI milleson_m_Page_073.jp2
6cd1c619cfbeb167630c466925ed50d2
5858c306ee5927e0d99a8edb198436f74774076e
653643 F20101122_AAASLU milleson_m_Page_059.jp2
210f6b0a88b336bb7b556544ad9bd6b2
d37e2c954063c447068110b20e7a9d1d910cf0cc
114969 F20101122_AAASMJ milleson_m_Page_074.jp2
26a1a4c37a49d1f48efd3a9159c6f438
4790e90b7d04d618932120084fc2d4abd3fe6b1e
93730 F20101122_AAASLV milleson_m_Page_060.jp2
ad94b98aae3ad9703de4095a8c7407b0
2d470399a81494efd95cab52c33dc8ad6d1fe981
954974 F20101122_AAASMK milleson_m_Page_075.jp2
d569ccbf239f021cb6960ac85e9b19f9
7f897931408031015dcfacaf7d897c290ea36f7c
798657 F20101122_AAASLW milleson_m_Page_061.jp2
a9b9700c42d8905581880e128488f32c
da86468f928da35cba20e199101b48c7172e2cbc
957079 F20101122_AAASML milleson_m_Page_077.jp2
65bfe0b0924664d2b1103125f2dbc75a
26043a8bf33facbb3b0c247e4f357ccaa10be362
761721 F20101122_AAASLX milleson_m_Page_062.jp2
a1f001aa28c68a340461e9007fae2d24
729d6f2f9d86ad45adff301e1ff6125e150d7df1
113319 F20101122_AAASMM milleson_m_Page_078.jp2
33de31156c2562235920c5b11edc5430
80ac0cc6e6a14c0277423e399250157d8757651b
624979 F20101122_AAASLY milleson_m_Page_063.jp2
c171e74f12a3ebbdb036d14cbd8641e4
143f055750c1a23e0e20dc7825c8d07148259aca
83099 F20101122_AAASNA milleson_m_Page_093.jp2
313e6368483aeec6c9b196ed28ab03e8
b18b777bae2c987f43258135f1dd39365da1fe72
883187 F20101122_AAASMN milleson_m_Page_079.jp2
8651cf0b12ee1a67e69ab3b96133f7ce
42c0922f64c92cab80dbdaf09702b9651c33eb21
602024 F20101122_AAASLZ milleson_m_Page_064.jp2
b956c0c09c62ef1da20b49a41ef82188
c855bfcdbde94875f7993f9ae01afdd44346f17b
116387 F20101122_AAASNB milleson_m_Page_094.jp2
eefb6217aee650e75a0aa120696e8736
6ae3d5727e2c6a109d4bf3e794d89506f1f7941b
104850 F20101122_AAASMO milleson_m_Page_080.jp2
82488cf940db38a9d95882d75326afd6
f1fbfcc303834208b7271f5f0fda332aae4a294a
86919 F20101122_AAASNC milleson_m_Page_095.jp2
7866675b1032b6052b339355a5975da2
2ec032c490264c4183c036ef3ad7deb8fd44b439
106037 F20101122_AAASMP milleson_m_Page_081.jp2
86df554cfa622c08553ca2314f80eb9a
4e36b5e71c39a3204d72af29b562f2ef97a2363c
116463 F20101122_AAASND milleson_m_Page_096.jp2
9f5bd9a2ca25ea5a21df74217913b47d
5148b7e62982b1f925214d90c0dd36d372c6ed33
82978 F20101122_AAASMQ milleson_m_Page_082.jp2
3ec48cf606b50806472c0e5777d2d124
eb4528c16a04ef053cff2658e497477685ff28b5
1051938 F20101122_AAASNE milleson_m_Page_097.jp2
915d456d278f0eaf55afc54e7b43181e
56701c08883d28bdf65da00b243337ec71aa80e5
132654 F20101122_AAASNF milleson_m_Page_098.jp2
040c33d83ce9f20e716217f231e6cdc2
b528b38c06206780231da6b0e7ab90bebc052be3
864284 F20101122_AAASMR milleson_m_Page_083.jp2
26c6c3d9fe877d0e3d8a45bb436c19a3
51eab8db9817f714d3f4d826450c37facd17b497
128480 F20101122_AAASNG milleson_m_Page_099.jp2
cff287a64087e8df5851c05e25085e27
cd0350898f62fb119bb14e9fe2001271f86ce9bb
F20101122_AAASMS milleson_m_Page_084.jp2
bacafa51d4bafcb2318b9f657057bb97
ab26acfdf312f30ca3fe1482a119f74404b0cf53
117855 F20101122_AAASNH milleson_m_Page_100.jp2
dab30bc9f4c2b00b23ce467cb15936fb
1273a9e59a3f6c3c2c0b69952fcd96c9fc552be2
F20101122_AAASMT milleson_m_Page_085.jp2
b5f2273c21646d483ee74135e8d18173
164dae2731cdda89284f57e656de4ddf2884be7b
125521 F20101122_AAASNI milleson_m_Page_101.jp2
c2c0b8676a7f2e4d6c7a3c576f52ce49
d9a29ef13e26e883cddcbc90fb33537a3d7b8c92
49383 F20101122_AAASMU milleson_m_Page_087.jp2
7a8889ccc504d3104122a638ec3a9915
d483153726b83550230b45f840bfab5a80dd7ad0
132205 F20101122_AAASNJ milleson_m_Page_102.jp2
d151a3ecbf109480aa55829098842c5e
57ba6575064891bd9a9cf19f46bfd682004b36e6
90680 F20101122_AAASMV milleson_m_Page_088.jp2
e7ce84a461767559dc180598265b2a43
a3b00d0a3d9c75ed5abd210cebb555154cb1151c
126870 F20101122_AAASNK milleson_m_Page_103.jp2
c6658932ae030efc57af7cc74e0e03b7
2512da5089abe504e4e72cdcdd99acb97674bc5a
86186 F20101122_AAASMW milleson_m_Page_089.jp2
18cf591d24c2fcd3bc89613764128607
adf63d8e61f0223f6fff9a6b08cc54f138da47e4
126908 F20101122_AAASNL milleson_m_Page_104.jp2
0dd81dd4674dc21005886c62e51f22b9
637c8914e0eef8d29657deb131c08fa6fd20e24f
76993 F20101122_AAASMX milleson_m_Page_090.jp2
823c8d6f9a5097583f9bd0548209a1e7
11cdbd86e26ecabdf5fc29aac1f108281946c7fd
F20101122_AAASOA milleson_m_Page_016.tif
890ce042a77cafc8f0aaca4344682f9b
ec23fbdb15a95922525334bcd13f4e2eb6b4b3e1
82042 F20101122_AAASNM milleson_m_Page_105.jp2
b909eefc28975d3faff4299c497dd4a4
a549b2128ff86a8152d6e6bafa186fc28cf2a5d0
77119 F20101122_AAASMY milleson_m_Page_091.jp2
f16b1ee27946a637b3f31a404741e6b0
7c608d29dc6ed248147a154649066dbff1421081
F20101122_AAASOB milleson_m_Page_017.tif
d0ba1d9a3c6170d871a9379b1dfe3fc0
b54d876304c0db530421880478fb74063e9e74ed
60260 F20101122_AAASNN milleson_m_Page_106.jp2
cd473a516feebe672e3cb1b714508d19
fb3c184993f660d687107b8b85baa9c7160228a2
93095 F20101122_AAASMZ milleson_m_Page_092.jp2
a37deb116887bd37722b3c9ce7f0f602
1075f4ff9ee36004a14b2afa48fff0d26f5e47ea
F20101122_AAASOC milleson_m_Page_018.tif
e56c8fc041e606e85016c42ee97c850d
f21c21c61c709ee172dde964c87883b563ceab09
F20101122_AAASNO milleson_m_Page_001.tif
73b53826ea8d95370538d4bdf2e01dc6
c062ddae4a1a132821b78e0db8c519bf9ab091f8
F20101122_AAASOD milleson_m_Page_019.tif
c37732d8816e1bb941c14e3444337ff1
9fdb833eafbf502b0c8f50863f98dd9140363a29
F20101122_AAASNP milleson_m_Page_004.tif
955e77a02b7c32bfe1e4ead4cd8f2869
ccd542f6c0a79dcf9f165c0665906a1a596c24d0
F20101122_AAASOE milleson_m_Page_020.tif
cf85f72d4bc7ca104a544a4ba46abc64
99b97079ea5e61a6c310a5f62994d26a8ec24b75
F20101122_AAASNQ milleson_m_Page_005.tif
0325a9b76308a09422edf68e7ec88f06
191c212c21146d6af22b1f91cb209df0b6249ee6
F20101122_AAASOF milleson_m_Page_021.tif
4fb21f88329d1e815bae7faacc2beaae
f76960a7901a2ac24d5582b411538c9356427415
F20101122_AAASNR milleson_m_Page_006.tif
02753ec9e4dbf1bd578dcf6b2b185bdb
0f3558669518b38a1491da70b06d0ca17a3a2f2a
F20101122_AAASOG milleson_m_Page_022.tif
d48112a3dd5e2f56314951ede320b237
c1983821d45278bd128532a95f70b1d50bdc3c85
F20101122_AAASNS milleson_m_Page_007.tif
17e450650d56778e5fff91b5825490f1
ef619c1b247ca8fe16ca0ab19db8e688445e56dc
F20101122_AAASOH milleson_m_Page_023.tif
617ca80550c308dd409e6fff43feb69e
937e8d594bb21be8973f9853ed7acf1063857cf9
F20101122_AAASNT milleson_m_Page_008.tif
efb72de8513ce14705fe1592e6d5aa1e
c0f4a5b14cf1784306a51d86bc53ed0e4e2265ae
F20101122_AAASOI milleson_m_Page_024.tif
0abdb75aeb77d59bbac8be6acaa2fdf7
66f1abcf012816ca1003ee0c0d8dc08f7e9f9550
F20101122_AAASNU milleson_m_Page_009.tif
c15e72b729ac820677b86daf1beac3c9
57780605deae7a64fc0f150b39d17a03f8311fde
F20101122_AAASOJ milleson_m_Page_025.tif
10d29807f6e3a8c6a897d10f5f9c7ea7
c5955b2772a5c8b2e6ddc1307ac056fe6f159f1b
F20101122_AAASNV milleson_m_Page_010.tif
8935972714ff6a54ed756f30f1e46508
6abc0d448e57e7d91f563b622421264e4f036e1f
F20101122_AAASOK milleson_m_Page_026.tif
c02adfabe079fa3aafd7bc196fd49a9f
3cc6431c23ac15961a6d23c36cb36f0b8aa80c44
F20101122_AAASNW milleson_m_Page_011.tif
c095c6f8905fa958d2f6fdc291dd485d
65d8ac2d26319929ff632a9f1d7c076397c92cc2
F20101122_AAASOL milleson_m_Page_027.tif
0e43fcfb1196efa94860ef16bf82dd28
d711a4906e3ae10110397cff073abf8d5f989418
F20101122_AAASNX milleson_m_Page_012.tif
f769b07807123183299f090f880dfea1
78969c3867cf8dc2f5ea85194797c83bb89d0873
F20101122_AAASPA milleson_m_Page_042.tif
85a8eb75ef543720d8129576d18c7d95
acffe1baad2185ba1026a881de6449d0d2716336
F20101122_AAASOM milleson_m_Page_028.tif
a6f81369a5f3073e82ecc7dad8a0e5a4
bd633a80c4b581743820376224366475f737d88f
F20101122_AAASNY milleson_m_Page_014.tif
1e865d540a7d16e9b36af454f3d4c27d
19e941587ba298c202c692e0c8f3fccbad836b45
F20101122_AAASON milleson_m_Page_029.tif
3d57d1359046462f7e7a2b3ec5191173
55eb902e7622db97e8e4a275aa032378dfa41656
F20101122_AAASNZ milleson_m_Page_015.tif
665dbf108b82bae8831525193d2f3eb8
5316aebceb4fceb6830db5ff28a225504b7ebb68
F20101122_AAASPB milleson_m_Page_044.tif
92a3854d589334b7f5e06a9ade10ada3
48ac6c89a0cea5a0ccb33e5dd1fa6ea85ea98c77



PAGE 1

THE ROLE OF BIRDS AND MICROSITES IN THE REGENERATION OF SOUTHTEMPERATE RAINFOREST By MICHAEL P. MILLESON A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Michael P. Milleson

PAGE 3

ACKNOWLEDGMENTS I thank my parents first and foremost for helping to get where I am today. I thank the Fundacion Senda Darwin, Biological Station, Senda Darwin. I also thank Traci M. Darnell for all of her help and support, and Mary F. Willson and Juan J. Armesto for their guidance. Finally I thank my awesome advisor, my wonderful girlfriend, my wacky dog, and each and every one of my friends. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iii LIST OF TABLES ............................................................................................................vii LIST OF FIGURES .........................................................................................................viii ABSTRACT .......................................................................................................................xi CHAPTER 1 AVIAN SEED DISPERSER ACTIVITY AND AVAILABILITY OF GERMINATION SUBSTRATE IN BACCHARIS-DOMINATED OLD-FIELDS IN SOUTHERN CHILE...............................................................................................1 Introduction...................................................................................................................1 Chilean South-temperate Rainforest......................................................................2 Disturbance, Arrested Succession, and Consequences..........................................3 Alternative Hypotheses for Arrested Succession: Seed Dispersal vs. Germination Limitation.....................................................................................4 Seed dispersal limitation................................................................................5 Seed germination and seedling establishment limitation...............................8 Research Design...........................................................................................................9 Frugivore activity hypothesis..............................................................................11 Germination site hypothesis................................................................................11 Methods......................................................................................................................12 Study Site.............................................................................................................12 Study Species.......................................................................................................12 Frugivore Activity Hypothesis............................................................................13 Bird censuses................................................................................................13 Focal samples...............................................................................................14 Seed traps.....................................................................................................15 Germination Site Hypothesis...............................................................................15 Seedling transects.........................................................................................15 Substrate availability transects.....................................................................16 Germination and survival experiment..........................................................16 Results.........................................................................................................................17 Frugivore Activity Hypothesis............................................................................17 Bird Censuses...............................................................................................17 iv

PAGE 5

Focal samples...............................................................................................18 Seed traps.....................................................................................................19 Germination Site Hypothesis...............................................................................20 Seedling transects.........................................................................................20 Substrate availability transects.....................................................................20 Germination..................................................................................................21 Discussion...................................................................................................................21 Frugivore Activity...............................................................................................22 Seedling Establishment.......................................................................................24 Recruitment Foci.................................................................................................26 Conclusions.........................................................................................................27 2 RAINFOREST RESTORATION SCENARIOS FOR BACCHARISDOMINATED OLD-FIELDS IN SOUTHERN CHILE: A SIMPLE ECOSYSTEM MODEL AS A DECISION MAKING TOOL...................................28 Introduction.................................................................................................................28 Chilean South-temperate Rainforest: Natural Disturbance Regime and Arrested Succession.........................................................................................28 Modeling Restoration Scenarios..........................................................................31 Model Description......................................................................................................33 Overview.............................................................................................................33 Old-field Characteristics......................................................................................35 Seed rain..............................................................................................................37 Seed Germination and Seedling Survival............................................................40 Creation of New Foci..........................................................................................41 Tree Growth and Survival...................................................................................43 Sensitivity Analyses............................................................................................45 Management Scenarios........................................................................................47 Results.........................................................................................................................48 Initial Conditions.................................................................................................48 Coarse woody debris....................................................................................48 Baccharis......................................................................................................49 Sphagnum and grass.....................................................................................49 Tree cover.....................................................................................................50 Length of edge..............................................................................................50 Bird density..................................................................................................53 Sensitivity Analyses............................................................................................53 Germination and tree growth rate.................................................................53 Survival rate of seedlings on dead wood......................................................54 Rate of cover loss to forest...........................................................................55 Rarity of new recruitment-foci formation....................................................56 Management scenarios.................................................................................57 Discussion...................................................................................................................57 Relative Importance of Seed Dispersal and Germination Limitation.................57 Restoration Methods / Management Scenarios...................................................61 Assessment of model scenarios....................................................................61 v

PAGE 6

Recommendations for efficient reclamation of Baccharis fields.................62 Assessment of the Model.....................................................................................68 Complexity...................................................................................................68 Relevant scales.............................................................................................68 Model improvement.....................................................................................69 APPENDIX A ICONOGRAPHIC REPRESENTATION OF MODEL.............................................71 B MODEL EQUATIONS..............................................................................................75 LIST OF REFERENCES...................................................................................................84 BIOGRAPHICAL SKETCH.............................................................................................94 vi

PAGE 7

LIST OF TABLES Table page 1 Comparison of the number of seedlings found beneath non-tree cover types to the number found beneath trees to estimate the odds that a seedling would be established there.......................................................................................................44 2 Parameters used for sensitivity analyses. Values represent mean of 6 fields..........46 3 Field conditions, input, and time required for 50% regeneration under 3 different management scenarios and three different starting conditions................................60 vii

PAGE 8

LIST OF FIGURES Figure page 1 Conceptual model of options for use of an arrested successional site.......................7 2 Average number of frugivores counted during 10 minute point counts at 12 sites in Chilo, Chile........................................................................................................18 3 Mean number of frugivore visits during 30-minute focal samples to clusters of trees and adjacent single trees in degraded old-fields in Chilo, Chile (N = 24).....19 4 Mean number of frugivore visits during 30-minute focal samples to fruiting and non-fruiting trees in degraded old-fields in Chilo, Chile (N = 14).........................20 5 Density of seedlings found growing on each substrate type in 9 degraded old-fields in Chilo, Chile...............................................................................................21 6 Percent cover of various substrate types across six degraded old-fields in Chilo, Chile. Error bars represent +/-1 SE.........................................................................22 7 Number of seeds germinating on each substrate type in 18 trials, in degraded old-fields in Chilo, Chile........................................................................................23 8 Seedling density divided by substrate availability for various substrates in Baccharis-dominated oldfields throughout Chilo, Chile........................................25 9 A simple representation of old-field regeneration as conceptualized by the model........................................................................................................................34 10 The value of S, or seed rain per meter of forest in the old-field, changes with increasing tree cover in the old-field........................................................................43 11 Correlation between age of Drymis winterii (as determined by coring) and diameter at breast height (DBH)..............................................................................45 12 Correlation between the age of Drymis winterii (as determined by coring) and tree height.................................................................................................................47 13 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial amount of coarse woody debris (CWD) is increased...........................................................................49 viii

PAGE 9

14 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial amount of coarse woody debris (CWD) is increased with zero initial trees in the old-field................50 15 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial coverage of Baccharis is increased..............................................................................................51 16 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial coverage of Sphagnum is increased.............................................................................................51 17 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial coverage of grasses and ferns is increased...................................................................................52 18 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial number of trees is increased...................................................................................................................52 19 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the length of adjacent edge is increased...................................................................................................................53 20 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial number of frugivorous birds is increased..................................................................................54 21 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the germination rate is increased...................................................................................................................55 22 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of tree growth is increased or decreased..............................................................................................56 23 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the mortality of seedlings on dead wood is increased.............................................................................................57 24 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of seedling mortality increases...................................................................................................................58 25 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of out-competition of Baccharis is increased..............................................................................................59 ix

PAGE 10

26 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of out-competition of Sphagnum is increased.............................................................................................59 27 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of out-competition of grass and fern is increased........................................................................................61 28 Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the value of r (additional rarity of new recruitment foci establishment) is increased.......................................63 29 Time predicted for regeneration of forest (to 5, 25, and 50% forest cover) with low densities of seed dispersers, compared with time required for regeneration with very few micro-sites suitable for germination, compared to default values of each (see text)......................................................................................................65 30 Model prediction of the effect of the number of saplings planted annually on the rate of forest regeneration (to 25, 50, 75, 100% forest cover).................................67 x

PAGE 11

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science THE ROLE OF BIRDS AND MICROSITES IN THE REGENERATION OF SOUTH-TEMPERATE RAINFOREST By Michael P. Milleson December, 2005 Chair: Kathryn E. Sieving Major Department: Wildlife Ecology and Conservation Determining the mechanisms of arrested succession at restoration sites can influence understanding and management of landscape scale patterns and processes. On Isla Grande de Chilo, a large continental island in southern Chile, conversion of south-temperate rainforest to pasture is occurring at a high rate. In some cleared sites agricultural activity cannot be implemented due to invasion of persistent shrub fields comprised of Baccharis spp., a scrubby bush in the Asteraceae. Baccharis-dominated fields serve no economic purpose and are of limited use to wildlife. If they could be restored to native forest then landowners and endangered endemic wildlife species would accrue greater benefit. The goal of this study was to identify the relative importance of two likely limitations on natural regeneration of native forest in Baccharis-invaded sites on Isla Grande de Chilo, Chile. I tested two alternative hypotheses for arrested succession in my system: that shrub fields persist because of lack of seed dispersal, and lack of appropriate substrates for seed xi

PAGE 12

germination and seedling establishment. In chapter 1 I present findings of field studies (experiments and comparative observations) showing that both ample seed dispersal and provision of suitable germination sites must occur to increase the rate of forest regeneration in Baccharis-dominated shrub fields. Sites with few or scattered trees received significantly less avian seed-disperser visitation than did sites with more than approximately 30 trees/ha or with trees in clumps of several individuals. Additionally, substrate types with the greatest seedling density were the rarest of the substrates available. In the second chapter, I develop a dynamic systems model that incorporates processes actuating both hypothesized mechanisms of arrested succession to address realistic scenarios for restoration given constraints and goals relevant to local landowners. I use field data and findings reported in the literature to parameterize the model. Based on the model, the following suggestions are made. First, areas selected for forest regeneration should be adjacent to mid-successional or old growth forest stands in order to ensure sufficient avian seed dispersal. Second, the focus should be on providing sufficient germination substrates and making the field attractive to avian dispersers. The model also showed that given both avian seed dispersers and suitable germination sites, the most important deterrents to forest succession are the competitive effects of Sphagnum and Baccharis cover in the fields. Finally, the model demonstrates how these deterrents to regeneration can be overcome simply by planting a few trees each year in old fields that have minimal germination sites a realistic recommendation given landowner constraints. xii

PAGE 13

CHAPTER 1 AVIAN SEED DISPERSER ACTIVITY AND AVAILABILITY OF GERMINATION SUBSTRATE IN BACCHARIS-DOMINATED OLD-FIELDS IN SOUTHERN CHILE Introduction A common goal of ecological restoration is to quicken the natural pace of secondary succession (Hobbs & Norton 1996). However, in cases of arrested succession, when a damaged or degraded ecosystem does not return to the original state (Brown & Lugo 1994), the causes of the arrest must first be identified and removed, if possible, before secondary succession can proceed (Parker 1997). Three general types of factors can cause arrested succession. Sites can be colonized by species that inhibit the growth or spread of species more characteristic of the desired ecological state (Connell & Slatyer 1977); abiotic factors pushed outside the local species ranges of tolerance by the disturbance can prevent establishment by representative colonizers (Milchunas & Lauenroth 1995); or, finally, disturbance can bring about conditions (biotic or abiotic) that favor the influx of an entirely different set of species to the site (Suding & Goldberg 2001). Determining the mechanisms of arrested succession at small scales can influence understanding and management of larger landscape scale patterns and processes (Bell et al. 1997). Once succession becomes arrested, site characteristics may change considerably, pushing the site over a threshold toward the basin of attraction of an alternate state (Laycock 1991; Lewontin 1969). When undesirable alternate stable states arise in landscapes, restoration ecological approaches can be used to identify and release 1

PAGE 14

2 the mechanisms that generate and maintain them. The goal of this research was to identify mechanisms that may be inhibiting forest succession in previously cleared fields in the Valdivian temperate rainforest region of southern Chile. The persistent shrub fields that may develop following clearing of Chilean temperate rainforest can occupy significant areas (30% or more in regions where forest clearing is advanced) and provide little to no ecological or economic productivity (Gude, 2000; personal observation). Therefore, these persistent shrub fields are manifest as an undesirable alternate ecosystem state at the landscape scale. In this study I examine alternative causes of arrested forest succession at a community-scale, in shrub fields at forest edges, with the goal of identifying factors maintaining persistent shrub lands that, via restoration work, could be released or altered to allow forest succession to proceed. Chilean South-temperate Rainforest The South temperate rainforest of the Valdivian region of Chile (3548 S) receives between 1,000 and 6,000 mm of rain per year, and is characterized by emergent evergreen broad-leaved trees (e.g., Nothofagus oblique, N. alpina, and N. dombeyi) and conifers (e.g., Podocarpus nubigina). Typical canopy and understory species include Drimys winterii, Weinmannia trichosperma, and several trees in the family Myrtaceae (Willson et al. 1994) Endemism is very high (Stattersfield et al. 1998) ranging from 45% in vertebrates to 90% in seed plants (Armesto et al. 1996; Villagran & Hinojosa 1997) The typical disturbance regime is characterized by periodic catastrophic disturbances such as earthquakes, volcanic activity, and fire. Windthrow and treefall gap creation are common occurrences (Veblen 1979) This disturbance regime prevents shade tolerant tree species such as Laurelia phillipiana and Saxegothea conspicua from out-competing the shade intolerant Nothofagus (Bustamante & Armesto 1995) Due to

PAGE 15

3 human settlement and agricultural land uses, much of the remaining Valdivian rainforest exists as fragmented patches in a matrix of pastoral, agricultural, and industrial forestry land uses, where increased gap creation, fire, and windthrow frequencies along forest edge have intensified disturbances in remaining forests (Willson & Armesto 1996) Forest clearing and associated human activities in remaining forest have resulted in global endangerment of endemic flora (Armesto et al. 1998) and fauna of the region (Stattersfield et al. 1998). Disturbance, Arrested Succession, and Consequences On Isla Grande de Chilo, a large continental island serviced by a system of ferries, forest conversion has been slower than on adjacent mainland (Rozzi et al. 2000) due to its greater economic isolation. Here, the rural life style is characterized by pasture creation (via tree cutting followed by fire) for milk and meat cows and sheep, by non-mechanized row crop production (oxen teams are often used to till the soil), and by fuel wood acquisition in the most accessible forest patches (Armesto et al. 1998). Large scale forestry (via clear-cutting) also occurs in the islands southern sectors, and plantations of pine and eucalyptus are increasing throughout Chilos rural landscapes (Armesto et al. 2001b). In rural communities, economic productivity for the people is partly determined by availability of pasture for livestock and wood fiber for cooking and building materials. This can be limited by the development of persistent shrub fields (dominated by Baccharis magellanica) following forest clearing that cannot be used for livestock or row crop production. Moreover, native forest succession does not occur readily in Baccharis-dominated sites. While native forest frequently reinvades logged sites that are not further disturbed by fire, and in some agricultural old fields left fallow, Baccharis fields frequently

PAGE 16

4 develop after forest clearing. It appears that Baccharis takes over especially where the water table may be higher than elsewhere, and significant soil inundation prevents establishment of both native trees (Bewley & Black 1982) and cultivars. The shrubs can be burned back but, without prohibitively expensive ditching and draining, fire alone does not often improve site utility for either agriculture or forest regeneration. Baccharis overstory may prevent establishment by forest tree species (Cespedes et al. 2002; Putz & Canham 1992), and is underused by the local avifauna (Gude, 2000). Therefore this shrub land formation is of low economic and ecological value, and it appears to be highly persistent. In the landscape of NE Chilo, Baccharis fields comprise around 30% of the land cover and are virtually unused by native wildlife species (T. M. Darnell, K. E. Sieving, unpublished data). Since farmers that clear forest for pasture or wood products and get Baccharis development in the cleared area usually move to clear a different site, if available, I view these fields as a restoration priority. Regeneration of forest on arrested successional sites would provide wildlife habitat and at least minimal economic benefit (forest products) for people, and this might protect forest in other sites from additional clearing (Fig. 1). In this study I focused on understanding factors limiting natural forest regeneration in sites dominated by Baccharis shrubs. Alternative Hypotheses for Arrested Succession: Seed Dispersal vs. Germination Limitation Forest regeneration can be limited by several factors, including competition, lack of nutrients, irregular disturbance regime, or allelopathy (Brewer 2002; Connell & Slatyer 1977; Kirkman et al. 2004; Mallik 2003; Wilson & Shure 1993; Cespedes et al. 2002). While all of these factors likely play a role in creating and maintaining a state of arrested

PAGE 17

5 succession, this study focuses on dispersal and germination due to their importance in this system (Armesto & Rozzi 1989; Papic & Armesto 2000). Seed dispersal limitation Insufficient seed dispersal can limit opportunities for establishment and growth of diverse plant species and, thereby, reduce vegetative structural heterogeneity. Given that more than 70% of all trees, shrubs, and vines in the Valdivian rainforests are bird dispersed (Armesto & Rozzi 1989), access to Baccharis fields by frugivorous birds and their activities in them are likely to define many parameters of regeneration. In a study by Armesto et al. (2001a), only 10% of the fleshy fruits collected in seed traps placed in rainforest fragments reached the margins of the forest patch, suggesting that even fewer would reach beyond forest edges and into shrub fields. Moreover, since the principal seed dispersing birds are forest species (Willson et al. 1994), the absence of suitable habitat for them in cleared fields could contribute to arrested succession. Seed dispersal is important in this site if seeds are able to establish in open areas or if they are deposited on suitable microsites for germination (Howe & Mirti 2004). Three possible factors have been identified that might make a site, such as an anthropogenic shrub field, unsuitable for use by frugivorous birds: a lack of perches, a lack of structural diversity, and a lack of food. Several studies have found that seed rain is positively correlated with vegetation that offers natural perching sites (Debussche & Isenmann 1994; Ferguson & Drake 1999; Harvey 2000; Kollman & Pirl 1995) and with the availability of manmade perches (but see Holl 1999; McClanahan & Wolfe 1993; McDonnell & Stiles 1983). In this study, I assessed the importance of natural (tree) perch availability in Baccharis fields on avian frugivore activity.

PAGE 18

6 A lack of vegetative structural complexity may also result in decreased visitation to a site by seed dispersers. Holl (1998) and McDonnell and Stiles (1983) found that perches that are more complex received more seed rain. Cardoso da Silva, et al. (1996) also found a positive correlation between structural complexity and bird use of a site. Structural complexity, and its positive effect on bird use, is also increased when trees are found in close proximity to one another (Toh et al. 1999). Complex vegetative structure may appeal to birds for reasons such as increased cover and more diverse microhabitats. In sum, enhancing complexity in target degraded sites may be an important factor in restoration where bird dispersal is a central constraint on inputs of seeds. To address this aspect of seed dispersal limitation into shrub fields, I assessed the effect of simple natural perches (lone trees) versus more complex perching and cover for frugivorous birds, represented by clumps of trees. A third factor limiting seed dispersal may be lack of food; fruiting vegetation in a restoration site can attract birds to make more visits during which they are more likely to defecate seeds. Although Holl (1998) found that using fruit as bait did not increase bird visitation, Cardoso da Silva, et al. (1996) found an increase in the use of an abandoned field when naturally occurring fruit resources were higher. Moreover, Slocum and Horvitz (2000) found greater seed dispersal beneath fleshy fruit producing trees in Costa Rica. Wunderle (1997) also suggests that, in general, the presence of fruit plays an important role in attracting seed dispersers. Thus, a potential consideration in restoration efforts is the availability of fruit to frugivores in the target sites. In this study, I addressed this possibility by assessing the relative influence of fruiting and non-fruiting trees in shrub fields on frugivore visitation and activity.

PAGE 19

7 Pasture Creation: Ditching, Draining, and Burning Release Arrested Succession Ornithocory Increased tree seed germination Agriculture Native Forest Regeneration Purview of this s tudy Provides fuelwood/forest products Decreased amount of landscape in arrested state Reduced need for clearing new tracts of forest Reduced impact on the landscape Options for Restoration of Baccharis Fields Figure 1. Conceptual model of options for use of an arrested successional site. The left path helps alleviate continued fragmentation pressure on temperate rainforests in S. Chile, but is expensive. The right path enhances regeneration of forest and wildlife habitat in areas that would otherwise have no wildlife or economic values. Both have the same end result from a landscape level perspective; however, fully identifying how to enhance regeneration is the purview of this study.

PAGE 20

8 Seed germination and seedling establishment limitation Once successfully dispersed to a site, tree establishment and growth can be limited at various life stages. Germination of dispersed seeds can be prevented by physical characteristics of the microsite (insolation, moisture, litter depth, pH, etc.; Houle 1992; Peterson & Pickett 1990; Streng et al. 1989), and by seed predation (Hulme 2002). Other critical life-stages include the seedling and sapling stages, at which point plants are especially vulnerable to herbivory (Hanley 1998) and fungal attack (Rey & Alcantara 2000). In Baccharis fields, inundation, desiccation, and nutrient limitation may all affect seed germination. Year round surface water inhibits germination of most tree species (Bewley & Black 1982), and is likely to do so under the hydric conditions that characterize Baccharis fields (Papic & Armesto 2000). Moreover, as Baccharis becomes established in cleared sites development of a surface layer of Sphagnum is commonly observed (Ruthsatz & Villagran 1991) and could reduce germination of seeds in at least two ways. When Sphagnum accumulates in a disturbed site, it creates acidic and nutrient poor conditions that are known to exclude colonization by forest species (Van Breeman 1995, J. Armesto, K. Clark, pers. comm.). Seeds may also desiccate readily when deposited on top of Sphagnum mats, because while this lichen requires standing water to grow, the top layers can be well above available moisture (Van Breeman 1995). It has been suggested that all of these extreme conditions limiting seed germination in hydric forests and old-fields can be alleviated by the availability of dead wood on the ground (Papic & Armesto 2000; Takahashi et al. 2000). Woody debris in Baccharis fields fluctuates less in water content relative to soils (Papic & Armesto 2000), and therefore, could ameliorate both low and high water stresses on seeds falling in bare ground or Sphagnum-laden sites. Moreover, large pieces of woody debris from tree

PAGE 21

9 trunks (commonly called nurse logs) can collect organic debris, preventing nutrient limitations (Harmon et al. 1986). After germinating on dead wood, seedlings subsequently become established by growing down into the soil. With sufficient maturity attained while supported by nurse logs, tree seedlings/saplings of wet forest can then tolerate, and even alter conditions of the soil. In other wet temperate forest systems, conifer germination is largely dependent on the availability of nurse logs (Hofgaard 1993; Simard et al. 1998), and in northern hardwood forests, yellow birch and red spruce densities are 24 and 5 times greater, respectively, on nurse logs than on the forest floor (McGee & Birmingham 1997). In premontane Costa Rican pastures, four of the most common woody species were found significantly more often on logs than on surrounding microsites (Peterson & Haines 2000). In a temperate Chilean forest system, seedlings of eight tree species common to the forests of Chilo, were found growing primarily on dead wood on the forest floor (Christie & Armesto 2003; Lusk 1995). Additionally, Papic and Armesto (2000) showed that survivorship of one-year-old seedlings of the five most dominant tree species found in the region is higher on woody debris in logged fields. Thus, it is likely that post-clearing fire applied by land owners in my study system reduces the availability of deadwood and that this is a limitation on germination success of seeds arriving into these fields. Research Design I considered two potentially interacting hypotheses to better understand the processes inhibiting succession in south temperate old-fields and the potential for manipulating them during restoration to forest. The first hypothesis, that a lack of avian frugivore activity is limiting forest regeneration, is based on the possibility that a lack of perches, structural diversity providing cover, and/or a lack of food is making the

PAGE 22

10 Baccharis fields unsuitable for use by frugivorous birds. Rather than use artificial perches, I conducted three comparative observational studies to examine the influence of naturally occurring variation in tree density and fruiting activity on bird activity at the scale of 0.5 ha old-fields. To test the second hypothesis, that the limited availability of suitable germination microsites is limiting forest regeneration, I conducted one comparative-observational study examining the relationship between the presence of seedlings on various substrates and the availability of these sites, and one experimental study examining germination rates on different substrates. This study was conducted along forest/old-field boundaries at sites occurring over a 400-km2 area. At the scale of my study plots, the phenomena of interest are localized bird movements, microhabitat choices (e.g., perches and feeding sites), and small-scale changes in substrate availability, which influence individual tree growth and distribution. At larger scales, other forces such as economics, large-scale disturbances, and patch context are operating to determine overall extent of forest versus other land uses in the regional landscape. But the purpose of this study was to examine small-scale processes potentially under the control of individual landowners seeking recommendations for managing their small parcels comprising the total area (sensu Hostetler 1999). With-in the context of a given field, the study was conducted close to the forest edge because successional processes fostering forest intrusion into hydric fields (e.g., frugivore activity, seed fall, and deadwood accumulation) occur from the forest edge outward (Armesto et al. 2001a; Armesto & Rozzi 1989). Data generated here were used to identify limitations on seedling establishment and to parameterize a systems model for comparison of different forest restoration scenarios (Chapter 2).

PAGE 23

11 Frugivore activity hypothesis Assuming that increasing frugivore activity correlates strongly with seed movement (Westcott & Graham 2000), I studied frugivore visitation rates to old-fields with varying numbers of trees (Debussche & Isenmann 1994; Ferguson & Drake 1999; Harvey 2000; Kollman & Pirl 1995), clusters versus single trees (Cardoso da Silva et al. 1996; Holl 1998; McDonnell & Stiles 1983; Toh et al. 1999), and fruiting versus non-fruiting trees (Cardoso da Silva et al. 1996). I predicted that bird activity would be influenced by type (fruiting vs. non-fruiting), occurrence, and distribution (solitary vs. clumped) of trees in fields and proximity to forest. Specifically, I predicted that avian frugivore abundance would be greatest in old-fields with greatest numbers of remnant trees, that avian frugivores would be most often associated with clusters of trees rather than single trees, that avian frugivores would be more often associated with fruiting trees rather than non-fruiting trees, and that avian seed deposition would be greatest beneath trees. Germination site hypothesis To test whether germination sites for forest trees are limiting in old fields, I surveyed fields for tree seedlings and identified 6 relevant micro-site types; beneath trees, coarse woody debris (CWD), CWD beneath a tree, beneath Baccharis, Sphagnum, and bare ground. I then surveyed fields for the availability of micro-sites that were relevant to seedling establishment, and tested to see if sites promoting seedling establishment were limited. I predicted that I would most often find seedlings growing on dead wood beneath trees, and least often on Sphagnum moss. I also tested actual establishment rates by planting seeds on three different substrates (decaying wood, bare dirt, and Sphagnum) and comparing their germination and survival success. I expected to see greater rates of

PAGE 24

12 establishment on dead wood when compared to Sphagnum moss and bare ground (Papic & Armesto 2000). The major assumption is that sites that are suitable for germination are also suitable for further survival, which is not always true (Gunnarsson & Rydin 1998). This assumption was not tested, however Papic and Armesto (2000) found that seedling survival is higher on coarse woody debris than on bare ground. Methods Study Site The study was conducted during the months of January and February 2001-2002, at and near Estacin Biolgica Senda Darwin, a field station located on Isla Grande de Chilo (9,600 km2 ) roughly 10 km from the coast of Chile (4155S, 7335W). The main woody tree species colonizing disturbed habitat are the avian dispersed D. winterii, and E. cordifolia, and the wind dispersed N. nitida (Veblen 1985). The study fields are located in the northeastern part of the island, near the towns of Manao and Linao. Mean annual rainfall is 1906 mm (peaks in Austral winter; June-September) and the mean annual temperature is 11 C (Armesto & Figueroa 1987). Study Species The primary seed dispersers in this system are Elaenia albiceps (white crested elaenia or fio fio) and Turdus falklandi (austral thrush or zorzal; Willson et al. 1994). The white crested elaenia occupies forest interior, edges, and clearings in Nothofagus forest. Its breeding season is from November to February (Fjelds & Krabbe 1990). The austral thrush makes use of a variety of habitats ranging from Nothofagus understory to gardens, parks, or brushy country. Its breeding season begins in October and ends in February (Fjelds & Krabbe 1990). Nothofagus nitida was the numerically dominant tree

PAGE 25

13 species in the post-disturbance shrub fields, followed by Drimys winterii, Amomyrtus meli, Eucryphia cordifolia, and Podocarpus nubigena. Frugivore Activity Hypothesis Bird censuses I conducted bird censuses during the breeding season (January and February 2001-2002) between the hours of 06:30 and 09:30 to examine the effect of trees in fields on the number of avian visits to fields. A total of 12 0.5 ha rectangular old field sites were selected, including 4 sites in each of the following categories based on the density of emergent forest trees (> 10cm dbh); low (with zero to 18 trees/ha), medium (3048 trees/ha), and high (more than 58 trees/ha). Site selection was constrained by proximity to the field station and thus was non-random. However, sites were at least 200m apart, and in most cases greater than 1000m apart, limiting the chances of non-independence. For each sample, I delimited a 100m by 50m section of old-field adjacent to a forest edge that contained Baccharis magellanica. Nine sites were sampled 3 times each in 2001, and an additional three sites were added in 2002. In order to avoid confusing temporal effects on frugivore abundance with site effects, all sites were censused once before any site was censused a second time (with two exceptions, due to travel restrictions). Censuses were only conducted on non-rainy mornings. For each sample, I recorded each frugivore seen moving from forest into the 0.5 ha section during a 10 minute period. Ten-minute point counts probably allowed double counting to occur. However, I was not concerned with movement from forest to field per individual bird, but rather total number of field visits per unit time. Whether by one or by several birds, each visit has an equal probability of resulting in a defecated seed. Since all of my sites were located near forest patches large enough to support many individuals of my study species, I assumed that

PAGE 26

14 linear densities of these birds territories along forest edges were comparable among sites. Thus, my census plots sampled visits by an equal number of individuals. The effect of field type (high, medium, or low number of trees) on number of avian frugivore visits was subjected to a Kruskal-Wallace one way analysis of variance. Focal samples I used a sub-set of the sites described above to conduct 30-censuses of two specific trees or tree clusters between the hours of 07:00 and 10:00. Sites were selected based on the availability of trees that fit the following design. To compare clusters of trees to single trees, I selected a cluster of trees, usually mixed species, and a single tree, equidistant from the forest edge and within 20m of one another. I defined a cluster as a group of two or more trees where each tree was within 0.5m of the foliage edge of its nearest neighbor. Mean cluster size was 10.18m (+/1 S.E. = 0.8430m) circumference at the outer edge of the crown. The mean crown circumference of a single tree was 4.00m (+/0.4624m). After selecting the trees, I placed myself in an inconspicuous location that provided an unobstructed view of both trees and clusters and counted the number of avian frugivore visits to either the cluster of trees or the single tree during a 30-minute period. 24 single-cluster pairs were censused. I also conducted focal samples comparing fruiting tree species (Drimys winterii) to non-fruiting tree species, using the same methods described above. Drimys winterii were not presently bearing fruit in four of the 14 pairs sampled. The effect of cluster type (cluster or single) and tree type (fruiting or non-fruiting) on the number of avian frugivore visits was analyzed using Mann-Whitney U tests.

PAGE 27

15 Seed traps To test for a difference in seed rain with distance from the forest edge, I placed 120 seed traps in clusters of three throughout four fields located at and near Senda Darwin. Seed traps were placed in clusters of three to increase the area sampled at each location. Seed traps were modeled after those used by Amesto et al. (2001a). The traps were constructed from a metal ring 30cm in diameter, supported by three metal stakes approximately 50cm above the ground. Seed catching area for each trap was approximately 0.07 m2. Plastic netting (mesh size = 2mm) was attached to each ring to collect the seeds. I placed 20 of the trap clusters within 25 meters of the forest edge, and 20 from 25 50 m from the edge. The design was slightly unbalanced, because there were insufficient trees within 25 meters of the edge. To determine whether avian seed dispersal was higher beneath trees, I placed 19 of the traps directly beneath a tree, and the other 21 at randomly chosen, non-tree locations. Traps were placed at the end of January 2001, and were checked at the end of February 2001, the beginning of January 2002, and the end of February 2002. Only seeds that were of a different species than the tree above the trap were counted, unless there was good evidence that the seed had been dropped by a bird (i.e., fecal material evident). I tested for the effect of distance from edge on seed rain and the effect of location (tree or non-tree) on seed rain using Mann-Whitney U tests. Germination Site Hypothesis Seedling transects In 2001 I set up five transects in each of nine fields to determine where seedlings were actually growing. Fields were chosen at random from the 12 fields that I sampled for frugivore activity. I placed 100-m transects parallel to the edge at 10, 20, 30, 40, and 50 meters from the edge, in order to control for distance from edge. Along each transect,

PAGE 28

16 I searched at three randomly selected points for each of the following substrates: bare ground, dead wood, bare ground beneath a tree, dead wood beneath a tree, beneath Baccharis magellanica, and on Sphagnum moss for a total of 18 randomly chosen points on each transect. At each point, I searched a 5m radius for the presence of the substrate, and counted the presence and number of seedlings growing on the nearest one-meter square area of the substrate. All trees less than one centimeter in diameter were considered seedlings. Seedlings were not identified to species, but were differentiated from non-woody species. Substrate availability transects In 2002, I randomly chose six of the 12 fields to analyze substrate/cover type availability. In each field, I established five 50-m transects perpendicular to the forest edge, in order to incorporate any variability along this gradient. Using a measuring tape I paced the transects, visually estimating the percentage cover of each substrate type for each meter of distance, arriving at a percent cover for each substrate type over the total 50 meters. A Chi-squared test was then used to compare the number of seedlings found at each substrate or cover type to the percent occurrence of each type. Germination and survival experiment In February 2001, I planted 1080 seeds of two species; Drimys winterii and Amomyrtus meli, on three substrates; dead wood, bare ground, and Sphagnum moss to compare rates of germination and survival on different substrates. Drimys and Amomyrtus were chosen based on availability and their prevalence in successional fields. At each of 18 locations, in two large Baccharis fields at Senda Darwin, I placed 10 seeds of each species on 0.1m2 area of each substrate. Only locations where all three substrates occurred within 2m of each other were chosen, to control for localized site effects. Due

PAGE 29

17 to this constraint sites were chosen based on availability, and thus were non-random. At each location, I placed seeds in three microhabitats, each of which constituted a treatment: (1) dead wood, (2) bare ground, (3) and Sphagnum moss. Prior to planting, any existing seedlings were removed, and the substrate surface was lightly scraped to remove any existing seeds. Seeds were lightly pressed into exposed soil or Sphagnum moss and dropped onto crevasses or pressed into soft portions of CWD. In an actual defecation event, seeds would be accompanied by fecal material, and would have been subjected to intestinal acids, which may enhance germination (Traveset et al. 2001). Each location was marked with small metal stakes and flagging tape. In January of 2002,, I recorded the percentage of seeds that had germinated and survived. The effect of substrate type on germination and survival was evaluated using a Kruskal_Wallace test. Results Frugivore Activity Hypothesis Bird Censuses The mean number of frugivores counted during 10-minute point counts for fields with low, medium, and high numbers of trees was highest for the sites with the most trees (> 29) and lowest for the sites with less than nine trees (Fig. 2). In total, 75 frugivore visits were recorded. Elaenia albiceps accounted for 64% of the observations, and Turdus falklandi made up the 36%. Mean number of frugivore visits at high sites was ~4 times higher than and medium sites, where mean frugivore visits was ~5 times higher than low sites. Sites with low and medium numbers of trees were both found to be significantly different from high tree sites, but not from each other ( 2 (2)= 7.78, P =0.02 N=12; Fig. 3).

PAGE 30

18 highmedlowNumber of trees in field 0246Mean number of frugivores Figure 2. Average number of frugivores counted during 10 minute point counts at 12 sites in Chilo, Chile. Sites are grouped as high (>29 trees), medium (15-24 trees), and low (0-9 trees) numbers of trees (N = 12). Error bars represent +/-1 SE. Focal samples During focal tree sampling, clusters of trees received significantly more bird visits on average than single trees (Z(46)= -3.38, P = 0.001, N=24; Fig. 3). When circumference was controlled for, clusters still received significantly more visits (Z(44)= -2.396, P=0.02, N=46). Clusters received 6.6 times more visits on average than single trees. However, there was no significant difference in the number of visits to fruiting and non-fruiting trees, though the sample size was lower (Z(26)= -0.48, P =0.73, N=14; Fig. 4) Fruiting tree species received approximately twice as many visits on average as non-fruiting tree species.

PAGE 31

19 ClusterSingle Tree configuration 0123Mean number of frugivore visits Figure 3. Mean number of frugivore visits during 30-minute focal samples to clusters of trees and adjacent single trees in degraded old-fields in Chilo, Chile (N = 24). Error bars represent +/1 S.E. Seed traps Four hundred and ninety four seeds from fruiting tree species were collected in the seed traps over the course of one year. Of these, 482 were found beneath trees (out of 16 trap clusters) and 12 were found in random non-tree locations (out of 2 non-tree trap clusters). After discounting seeds that I was unable to verify as being bird dispersed (i.e. seeds of the same species as the tree above the trap and not embedded in fecal material), there were 110 seeds dispersed to locations beneath trees and 12 dispersed to random non-tree locations. Seed rain of verifiably dispersed seeds within 25 meters of the edge did not vary significantly from seed rain between 25 and 50 meters from the edge (Z(38)=-0.03, P=0.98, N=20). Dispersed seeds were significantly more common beneath trees than in random non-tree locations (Z(38)=-3.33, P = 0.004, N=20).

PAGE 32

20 Germination Site Hypothesis Seedling transects I found the highest density of seedlings beneath trees, followed by dead wood beneath trees, Baccharis, bare ground, dead wood, and Sphagnum (Fig.5). FruitingNon-fruiting Tree category 0.20.40.60.81.0Mean number of frugivore visits Figure 4. Mean number of frugivore visits during 30-minute focal samples to fruiting and non-fruiting trees in degraded old-fields in Chilo, Chile (N = 14). Error bars represent +/1 S.E. Substrate availability transects Baccharis was the most common substrate type, covering 34.57% of the sampled area followed by ferns, Sphagnum and bare soil. Trees, trees with dead wood, and dead wood were the least common substrate types (Fig. 6). Seedlings were found in significantly higher numbers than would be expected beneath trees, and on dead wood beneath trees, and in significantly lower numbers than expected on Sphagnum moss (X 2 (5)=192, P < 0.001, N=157).

PAGE 33

21 00.40.81.21.62Tree + DeadWoodTree Dead WoodBare GroundBaccharusSphagnumSubstrate/Cover TypeSeedlings/ square meter Figure 5. Density of seedlings found growing on each substrate type in 9 degraded old-fields in Chilo, Chile. Germination Of 1080 planted seeds (N=360 on each substrate type), only 7 Drimys seeds germinated and survived for one year. Five of these were on dead wood, two were on bare ground, and no seeds placed on Sphagnum moss germinated and survived (Fig. 7). Discussion I conclude that seed dispersal limitation is the most important factor limiting regeneration of Baccharis fields for the following reasons. Fields with no or very few trees received almost no frugivore visits. Since very few seeds were found in traps placed in open areas, some sort of perching structure appears to be necessary to jumpstart regeneration. Availability of suitable germination substrate, while also important, seems to be secondary to seed dispersal, since some seedlings were found in

PAGE 34

22 all substrate types. The latter finding suggests that although some substrates might be better for germination, increased overall dispersal into any sites could increase germination rates. While the history of the study fields is not fully known, most were logged at least 30-50 years ago (I. Daz, Pers. Comm.). Presence of seedlings and trees in some of my study plots suggests that succession to forest is not actually arrested close to forest edges, but slowed enough to be experiencing limitations on regeneration representative of truly arrested sites. Tree DWTreeDead WoodBare GroundBaccharisSphagnumFernSubstrate / Cover Type 0.000.100.200.300.40Mean proportion of cover Figure 6. Percent cover of various substrate types across six degraded old-fields in Chilo, Chile. Error bars represent +/-1 SE. Frugivore Activity Following my predictions, fields with more trees received more visitations from frugivorous birds than fields with fewer trees. However, lack of a significant difference between low and medium tree density suggests the existence of a threshold value for the number of trees required to significantly increase the number of visitations, alternatively,

PAGE 35

23 my sample size may have been too low to detect a difference. Based on my study design, the threshold could be around 15 trees per hectare. Therefore, at minimum, my results imply that if forest is logged and the land owner wants forest to regrow, then leaving at least 15 trees/ha will positively influence the number of frugivorous birds frequenting the area (and enhance the rate of regrowth; see Chapter 2). Due to the limited area of my study plots, however, point counts on a larger set of fields varying more systematically in the number of remnant trees should be conducted to obtain a more accurate determination of numbers of trees needed to attract seed dispersers. 25071080020040060080010001200Bare Ground Dead WoodSphagnumTotalTotal PlantedSubstrate typeNumber of seeds Figure 7. Number of seeds germinating on each substrate type in 18 trials, in degraded old-fields in Chilo, Chile. Focal sampling revealed that trees growing in large clusters attract more frugivores than trees growing singly. This may be due to the fact that a cluster of trees provides greater cover from predators and is more likely to contain multiple food resources than a single tree (Debussche & Isenmann 1994; Ferguson & Drake 1999; Holl 1999;

PAGE 36

24 McClanahan & Wolfe 1993; Slocum & Horvitz 2000). It is also possible that clusters of trees simply receive more visitors due to their greater volume and likelihood of being in the path of birds moving across open areas. But in any case, it may be more effective, when leaving remnant trees during forest clearing, to leave them in small clumps. Although, this may seem counterintuitive for pasture creation, trees can be useful in providing shade and windbreaks for cattle. Additionally, if pasture creation is successful, remaining trees can be used for fuel wood when it has become apparent that Baccharis and Sphagnum are not becoming established. Data from the seed trap study suggest that frugivores overwhelmingly deposit seeds beneath trees. Although density dependent mortality can be greater beneath trees (Howe & Smallwood 1982; Janzen 1970; but see Hubbell 1979; 1980), in this system it may be compensated by conditions under trees that are significantly better for germination and survival (i.e., elevated and drier soils, reduced daily variations in temperature and humidity, and a lack of competition from shrubs and grasses; Nepstad et al. 1996). This combination of seed attraction and favorable conditions for germination and growth characterizes what has been called recruitment foci, or points in old fields from which regenerating forest grows outward (McDonnell & Stiles 1983; Slocum & Horvitz 2000; Toh et al. 1999). Seedling Establishment The location of seedlings in fields further suggests that some of the seeds deposited beneath trees survive to the seedling stage. Seedlings were significantly less prevalent on other substrate types. After trees and trees with dead wood, the next most common location for seedlings was beneath Baccharis. Because Baccharis is so common relative to colonizing or remnant trees, on a per plant basis, Baccharis shrubs receive much less

PAGE 37

25 seed input than trees in old-fields. However, Baccharis is occasionally used as a perching site by the smaller bird species, and some seedling establishment might occur in the absence of trees, albeit at an extremely slow rate. When seedling density is divided by substrate availability, it becomes apparent that the rarest sites represent the highest seedling establishment (Fig. 8). 020406080100120TreeTree + DeadwoodDead woodBare soilBaccharisSphagnumSubstrate TypeSeedling density/substrate availability Figure 8. Seedling density divided by substrate availability for various substrates in Baccharis-dominated oldfields throughout Chilo, Chile. Testing germination success on different substrates yielded little data. Low germination and survival rates in my study (0.97 %) may accurately reflect an extremely low germination rate under natural conditions. In a study of germination of the same species under green house conditions, however, a much greater percentage of seeds germinated (~75 %; Figueroa et al. 1996). Actual germination rates may have been higher, however, because I was unable to check for germination between field seasons. It is quite possible that germination rates were much higher, and that subsequent survival

PAGE 38

26 was low. Moreover, it is not yet known for this system how gut passage affects the seed germination process (Traveset & Willson 1997). Further testing of seed germination on a variety of substrate types after frugivore gut processing would be beneficial. Recruitment Foci I found more seedlings growing beneath trees than any other location. This is somewhat expected a priori (Nepstad et al. 1996; Slocum & Horvitz 2000), and my work also suggests that this is due to some combination of increased seed input and better conditions for survival of seeds and seedlings beneath trees (Figs 2, 5). From the seed trap portion of the study, we know that seeds are being dispersed, and that the area beneath trees is the main recipient of these dispersed seeds. What we dont know is whether this area is also better for seed germination. However, the presence of a tree suggests that the area was historically a favorable microsite, particularly because most trees used for this study probably came in after the fields were created; the oldest tree that I cored was 54 years old. Typically, dispersal away from a parent tree is thought to increase seed survival by decreasing intraspecific competition (Hubbell 1979), as well as by providing escape from predators and pathogens (Connell 1971; Janzen 1970). However, in the case of Baccharis dominated old-fields, survival away from the parent tree is likely much lower, unless the seed is dispersed beneath the canopy of another tree. For future work it will be important to examine how distance from a tree influences germination success in sites where succession is arrested, as the rate of survival should reach a maximum closer to the tree, then drop off drastically, reflecting the unfavorable conditions of the site (Houle 1995). Ascertaining how close to a tree a seed needs to fall to maximize its germination success would facilitate restoration planning, by allowing estimation of optimal tree planting density.

PAGE 39

27 Conclusions In order to speed-up or initiate forest regeneration in Baccharis dominated shrub fields, lack of seed dispersal and a lack of suitable germination sites must both be overcome. My study suggests most strongly that seed dispersal is the main limiting factor. Moreover, I can conclude that the simplest means of overcoming arrested succession will be to plant tree saplings in fields where forest regeneration is desired. And while my work did not find overwhelming evidence that the presence of dead wood (nurse logs) is necessary to overcome germination limitation, evidence that this would be important is mounting from this and other rainforest systems (Harmon et al. 1986; Papic & Armesto 2000; Takahashi et al. 2000). Therefore, based on the field studies and literature review presented here, I put forward the following conservative recommendations for reducing Baccharis coverage where it exists, or fostering forest development rather than shrub development where forest is freshly cleared. (1) Plant, or preferably leave behind, a small number of forest trees after clearing; at least some in clumps. (2) Leave a significant amount of coarse woody debris throughout the area of interest to provide safe germination sites. In the next chapter I explore recommendations in greater detail using an empirical systems model of forest regeneration under different starting conditions in order to allow landowners greater certainty in applying these simple recommendations to their own fields which may vary in starting conditions that could affect the outcome.

PAGE 40

CHAPTER 2 RAINFOREST RESTORATION SCENARIOS FOR BACCHARISDOMINATED OLD-FIELDS IN SOUTHERN CHILE: A SIMPLE ECOSYSTEM MODEL AS A DECISION MAKING TOOL Introduction Determining the underlying mechanisms of arrested succession at local restoration sites can influence understanding and management of larger landscape scale patterns and processes (Bell et al. 1997). For example, arrested succession is one way that stable and widespread community types (or alternate stable states; (Laycock 1991; Lewontin 1969) can arise that, in turn, define the mosaic of ecological and socioeconomic characteristics of human landscapes (Naveh 1994). Therefore, when undesirable alternate stable states arise in landscapes, restoration ecological approaches can be used to identify and release mechanisms that generate and maintain them. In this study, I identified potential mechanisms of arrested succession in mesic old-fields resulting from anthropogenic clearing of south-temperate rainforest in Chile. The community type produced that is undesirable, from both socioeconomic and ecological perspectives, is a persistent shrub community dominated by Baccharis spp. Chilean South-temperate Rainforest: Natural Disturbance Regime and Arrested Succession The South temperate rainforest in the Valdivian region of Chile (3548 S) receives between 1,000 and 6,000 mm of rain per year, and is characterized by emergent evergreen broad-leaved trees (e.g., Nothofagus oblique, N. alpina, and N. dombeyi) and evergreen conifers (e.g., Podocarpus nubigina). Canopy species include Drimys winterii, 28

PAGE 41

29 Weinmannia trichosperma, and several trees in the family Myrtaceae (Willson et al. 1994). Vertebrate dispersed species make up approximately 70% of the flora (Armesto & Rozzi 1989) and include Drimys winterii, Amomyrtus luma, A. meli, Eucryphia cordifolia, Weinmannia trichosperma, Podocarpus nubigena, Laurelia philippiana, and N. nitida. Two bird species, Eleania albiceps, and Turdus falklandii disperse the majority of seeds. Endemism is very high in this biome (Stattersfield et al. 1998), ranging from 45 % in vertebrates to 90% in some groups of seed plants (Armesto et al. 1996; Villagran & Hinojosa 1997). The natural forest disturbance regime is characterized by large-scale periodic catastrophes including earthquakes, volcanic activity, and fire that occur relatively rarely in a given site. Wind throw and tree fall gap creation occur much more frequently at any given location and at smaller scales (Veblen 1979). The natural disturbance regime prevents shade tolerant tree species such as Laurelia phillipiana and Saxegothea conspicua from out competing the shade intolerant Nothofagus (Bustamante & Armesto 1995) and helps maintain a typically diverse tree canopy composition across scales and throughout the Valdivian region. Human disturbances include widespread clearing of forest followed by burning and then conversion to agriculture (especially in the lowlands) and to industrial forestry plantations of exotic pine and eucalyptus. Typical farm and plantation plots are much larger, and more highly altered, than natural clearings. Productivity of cleared lands can be limited by the development of persistent shrubs (dominated by Baccharis magellanica) following forest clearing that prevent livestock or row crop production and natural forest succession. Apparently the Baccharis shrub canopy develops where the water table is, or becomes, elevated following forest

PAGE 42

30 clearing. While soil inundation alone can prevent establishment of native trees (Bewley & Black 1982), Baccharis overstory may also prevent establishment by desirable species (Cespedes et al. 2002); Putz and Canham 1992). Landowners can burn back the shrubs but without prohibitively expensive ditching and draining, fire alone does not reliably improve site utility for either agriculture or forest regeneration (Pers. Obs.). Baccharis species are common disturbance-related site invaders in many regions where they are native, as in this study (Stylinski and Allen 1999; Sarmiento et al. 2003). A natural ecotonal vegetation type called Magellanic moorland, occurring at higher elevations in Chile, has similar plant community properties to the Baccharis shrublands that represent arrested succession in the lowlands (Ruthsatz and Villagren 1991). But monospecific Baccharis shrublands are not naturally widely distributed in the study area and are a direct result of clearing at a larger scale than the natural disturbance regime. On Isla Grande de Chilo, up to 30% or more of local landscapes currently persist in this arrested state, and extensive areas have been dominated by uniform Baccharis stands for more than 50 years (T. M. Darnell, unpublished data). Baccharis shrublands are resource poor and not productive of native wildlife (Sieving unpubl. data) or agricultural produce. Farmers that clear forest and then see the development of Baccharis shrub in the cleared area will move on to clear another site, if available, to try again. Therefore, from the perspective of maximizing biotic productivity in this landscape, restoration of Baccharis fields to either pasture or native forest would be beneficial. Regeneration of forest would provide wildlife habitat and forest products for people, and viable pasture creation would stave off further clearing of native forest (Fig. 1). Given that forest clearing and human activity in southern Chile have resulted in

PAGE 43

31 global endangerment of endemic flora (Armesto et al. 1998) and fauna of the region (Stattersfield et al. 1998), native forest restoration is a high priority. In this chapter I explore two likely causes of arrested forest succession in through systems modeling of restoration scenarios I developed based on field data (from Chapter 1). In Chapter 1 I tested two alternative hypotheses for arrested succession in my system: that shrub fields persist because of lack of seed dispersal into them, and that they persist because associated conditions prevent seed germination and seedling establishment. I found evidence that both mechanisms are operating to suppress succession in Baccharis fields on Chilo Island, in addition to supporting justification based on the work of others. Therefore, in this chapter I develop a dynamic systems model that incorporates processes actuating both mechanisms of arrested succession in order to address realistic scenarios for restoration given constraints and goals relevant to local landowners. Modeling Restoration Scenarios Several restoration scenarios are possible for the patchwork of cleared old-fields that exist in the temperate rainforests of southern Chile. One possibility involves the creation of drainage ditches followed by controlled burning, to dry out the shrub fields and create pasture. While this does not replace forest, it does reduce the amount of additional forest that needs to be cleared. A second potential approach is to plant rows of trees with high rates of evapotranspiration (Ferro et al. 2003); trees can function as pumps in this scenario, reducing the hydric conditions and allowing easier conversion to pasture or forest. However, both of these approaches are costly, and depend on monetary resources of individual landowners; many farmers in the region do not have the resources to employ these techniques (Personal Observation). A third general approach,

PAGE 44

32 and the focus of this modeling exercise, is the restoration of native forest via restoration of secondary forest succession in shrub-dominated fields. Forest regeneration on arrested sites could lead to some economic gain for landowners (from forest products or provision of livestock shelter and understory browse during winter) and may, thereby, lessen pressures to clear more forest. In Chapter 1 I presented evidence that seed dispersal can be limited if birds are not attracted to shrub fields. Attractiveness is determined by the number of trees, whether they are fruiting, and their spatial distribution. Planting trees can help attract seed-dispersing birds into shrub fields, and potentially provide better microsites for germination and seedling establishment. It is also clear that because ground-level substrates where seeds might fall can be either too wet (bare ground) or too dry or acidic (Sphagnum), coarse woody debris (CWD) coarse woody debris is the best germination substrate because it will catch and germinate seeds that would otherwise be inundated in water or desiccated in the Sphagnum layer (Lusk 1995; Papic & Armesto 2000). Therefore, suitability of germination and seedling establishment sites are determined by relative availability of CWD versus adverse substrate conditions. Both planting trees and adding CWD cost money and labor. Therefore, the goal of the model is to assess the gain (in rate of forest regeneration) against the estimated relative costs of inputs. This chapter presents a dynamic systems model of seed disperser response to field characteristics, seed germination processes, and tree establishment and growth in the old-fields adjacent to south-temperate rainforest on Isla Grande de Chilo, Chile (4155S, 7335W). The purpose of this model is threefold. The first is to expand our understanding of the relative importance of seed dispersal by birds and the limitations of

PAGE 45

33 seed germination as key components of forest succession. The second is to explore methods for restoring rainforest by simulating removal of limitations to the process of succession (i.e. lack of trees, insufficient germination sites). The third objective is to assess the model as a predictive tool that can be used by managers and landowners who want to restore native forest, given certain constraints (i.e., time, money, labor). I use field data (Chapter 1) and information from other studies, to parameterize the model, which necessarily simplifies certain aspects of the regenerative process. For example, interactions between tree species, soil properties, and nutrient levels are ignored. Sensitivity analyses were applied in cases where model parameters are estimated or derived from the literature to analyze their importance, to provide a range of outcomes when a parameter is not well known, and to generate useful hypotheses for further testing. Model Description Overview The model is comprised of five subsystems: old-field characteristics (size, cover and substrate types, length of edge), seed rain, seed germination and seedling survival, creation of new recruitment foci, and tree growth and survival (Fig. 9). The underlying premise of the model is that succession acts as a positive feedback loop during early stages such that trees present in the old-field act as an attractant for avian seed dispersers. As seeds are dispersed to suitable germination sites, the number of trees increases as does area covered by tree canopy and, in turn, this attracts more seed dispersers and alters the habitat to create more germination sites (Alcantara et al. 2000; Slocum 2001). At later stages, as tree density increases, homeostatic properties slow forest growth (e.g.,Tappeiner et al. 1997; Wills et al. 1997). Rather than attempt to explicitly model

PAGE 46

34 every aspect of the ecosystem, select attributes are modeled incorporating available data from field studies. For example, instead of modeling all processes that could influence tree growth rates; I incorporated rates of tree growth recorded in nearby systems (see below). Stella 7.0 (Wallis et al. 2001), an icon based modeling environment, was used to create this model. Following is a brief description of key model equations. The model is presented in its entirety in Appendix 1 (iconographic form) and Appendix 2 (equation form). tree growth and survival seed germination and seedling survival seed rain Figure 9. A simple representation of old-field regeneration as conceptualized by the model. Dark arrows represent the flow of material (through space or time i.e. seeds or tree growth), dashed arrows represent the influence of one component on another. creation of new recruitment foci old-field characteristics

PAGE 47

35 Old-field Characteristics The parameters I defined to reflect old-field characteristics include: area of ground cover of 5 different types; area of tree canopy cover; and the rate of spread of tree cover in the face of competition with other ground cover types. Cover types defined here (and observed in the field; Chapter 1) are trees, Sphagnum moss, course woody debris (CWD), Baccharis, grasses and ferns, and bare ground. The category grasses and ferns refers to vegetation that is greater than 0.25 m in height and that is sufficiently dense to limit the amount of light striking the soil. The category bare ground refers to either exposed soil or patches of sparse grasses and soil. Each of these cover-types either promotes or inhibits tree growth. Tree growth in this instance refers to the increase in total tree canopy area throughout the old-field. Other trees, CWD, and bare ground act as promoters of tree growth (i.e., area of canopy cover can increase in and around areas with these cover types). On the other hand, tree seedlings have difficulty getting established in areas with Sphagnum, Baccharis, and grasses and ferns. Therefore, growth of forest area is inhibited where these substrate types occur (Chapter 1). To represent proportion of canopy coverage, the canopy cover of trees in the old-field is calculated relative to the size of the old-field. This is shown in Eq. 1 Ca = Fa /Oa (1) where Fa is the area of ground covered by tree or tree canopy (also referred to as a focus), and Oa is the area of the old-field. When determining the actual area available for tree growth, the various substrate types are subtracted from the total size of the old-field, as in Eq. 2 Ga = Oa (SPcov + Bcov + GFcov) (2)

PAGE 48

36 where Ga is the area available for trees to grow, SPcov is the area of old-field covered by Sphagnum moss, Bcov is the area of old-field covered by Baccharis, and GFcov is the area of old-field covered by grasses and ferns. This means that trees are in competition for resources with grasses, shrubs, and mosses. In other words, when shrubs and mosses dominate an old-field, the area available for new tree growth is much lower than in an old-field without competing cover types because there is less available space and resources for the trees to grow. Little data are available concerning the rate of spread of Sphagnum moss. However, it has been shown that Sphagnum out-competes small plants by creating acidic, anoxic, and nutrient poor conditions (Van Breeman 1995). This model incorporates Sphagnum as a static feature that is slowly out-competed by trees, which, once rooted with a canopy over the top of the shrub layer, can grow uninhibited. Although the actual mechanisms underlying tree-shrub and tree-Sphagnum competition are poorly understood at present, it is assumed that as forest trees grow, microclimatic changes occur, making the habitat more suitable for trees and less suitable for Baccharis and Sphagnum. Figueroa and Lusk (2001), for example, found that Baccharis shrub canopy is highly shade intolerant. Additionally, Ohlson et al. (2001) describe the impedance of Sphagnum growth by Pinus sylvestris in a boreal bog ecosystem. Eq. 3 calculates the rate at which growing trees can out-compete Sphagnum moss. SPcov(t) = SPcov(t dt) (IF P > 0.99 THEN Fc *(1/180) ELSE 0) dt (3) where SPcov is the area covered by Sphagnum at time t, P is the proportion of the old-field covered by any substrate type other than bare ground (including forest), Fc is the circumference of the recruitment foci, and dt is equal to one month. The value in

PAGE 49

37 parentheses, 1/180, is an estimate, subjected to sensitivity analyses, as little data regarding Sphagnum-tree competition is available. Therefore, as the edge of the recruitment focus meets a patch of Sphagnum moss, it takes 15 years to shade out each square meter of Sphagnum moss that is adjacent to the regenerating patch of trees. The time required for forest to out-compete opposing cover types is unknown, and requires further research. Therefore, very conservative values were chosen. Eq. 4 and 5 show how Baccharis and grasses and ferns are similarly handled. Bcov(t) = Bcov(t dt) (IF P > 0.99 THEN Fc *(1/288) ELSE 0) dt (4) GFcov(t) = GFcov(t dt) (IF P > 0.99 THEN Fc *(1/48) ELSE 0) dt (5) It takes 24 years to shade out 1m2 of Baccharis, and four years to shade out 1m2 of grass or ferns. Other old-field parameters include the length of forested edge bordering the field, and the number and size of existing trees in the old-field, and are chosen to represent the field of interest. Seed rain Based on data collected (Chapter 1), it is assumed that avian-dispersed seed rain occurs with equal frequency within 50m of the forest edge, given appropriate conditions (this was the scale of the field studies in Chapter 1). Hence, the dynamics of this model are relevant to forest growth within 50m of the forest edge at larger distances into shrub fields, certain rates may differ. Two measurements were used to calculate the potential for seed dispersal into a field; the number of avian frugivores inhabiting forest adjacent to the old-field, and the seed rain beneath trees in old-fields. Bird censuses were conducted in forest edge habitat adjacent to Baccharis dominated old fields to estimate the number of frugivores in

PAGE 50

38 adjacent forests. Between 07:00 and 10:00 during January and February 2002, I walked two 200m line transects into each of three different forest patches and counted frugivores seen or heard within 40m of the transect. Due to the dense nature of the forest, distance was estimated, not measured. Forest patches were chosen based on proximity to pre-existing study sites and transect direction was chosen randomly. The average number of frugivores counted was 12.5 per 80m wide transect. Elaenia albiceps will travel hundreds of meters from the forest and into the adjacent old-field, if there is sufficient structure in the field (Willson et al. 1994). This means that for each meter of edge there are 0.16 birds with access to the old-field. This method is prone to under-counting (Bibby et al. 2000), but to avoid over predicting rates of forest regeneration, the outcome was not modified. This is high relative to linear densities of most breeding birds, but the primary seed disperser, Elaenia albiceps, occurs at very high densities in this system (Rozzi et al. 1996; Willson et al. 1994). To simplify the model, Turdus falklani was not included. To estimate seed rain, 60 seed traps of roughly 0.07 m2 in surface area were placed beneath trees in 2 old fields (see Chapter 1 for details). Over a one-year period, 92 verifiably dispersed seeds were found in 60 traps. Hence, 4.2m2 of old-field received 92 seeds in one year, generating a rate of 21.9 seeds/ m2*year, or 1.83 seeds/ m2*month beneath trees. Eq. (6) shows how the potential seed rain (PSR) was calculated. PSR = (Sr *Oa)*(L/100)*(Db/.015625) (6) where Sr is the number of seeds dispersed per unit of attractive habitat in the old field, L is the length of the edge, and Db is the density of birds in the forest edge with access to

PAGE 51

39 the old-field. As the number of trees in the old-field increases, the ability of the birds to disperse seeds equally to all areas decreases. However, this levels off to some degree because as the old-field becomes converted to forest it begins to provide suitable habitat, increasing the seed disperser population (Figure 10 shows how the value of S changes with the number of trees present in the old-field). The second part of this equation, (L/100)*(N/16), adjusts the amount of seed rain for the length of edge, based on the bird survey data suggesting that there are 16 frugivorous birds per 100 meters of edge. Because the value of S is derived from real-world sampling efforts where the edge and density were 100m and ~0.16 birds/m of edge, respectively, these values allow the model to adjust S to varying conditions (i.e. when there is 100m of edge N = 16, if there are more or less than 100m of edge the model adjusts N). Actual seed rain is based on the proportion of old-field covered by trees. The portion of the potential seed rain that is dispersed is equal to the ratio of area beneath trees to total field size (0.5 ha). However, most of this seed rain does not contribute to the expansion of recruitment foci. In many systems, seeds that fall under existing canopy tend to die off or form a seedling bank where they replace adults that die (Antos et al. 2000; Stewart et al. 1991; Szwagrzyk et al. 2001). From the trees perspective, avian frugivores both bring seeds from other species and individuals as well as disperse seeds for the tree. In any case, this results in a seed shadow that typically results in a leptokurtic distribution conforming to a negative exponential curve (Willson & Traveset 2000). Long distance dispersal events are dealt with below via the creation of new foci. For the purposes of the model, it is the short distance dispersal that we are mainly interested in, as seeds dispersed too far from the tree are likely to be dispersed in poor

PAGE 52

40 quality habitat unsuitable for germination, and would not influence regeneration in focal sites at this scale. Based on this traditional model of seed shadows, I expect seed rain to be highest beneath the tree canopy (where it contributes to replacement of existing trees) and, due to the lack of alternative perching sites, to drop off quickly as distance from the canopy increases. Additionally, I expect conditions in most of the old-field to be unsuitable for germination such that with the exception of rare conditions (discussed below) survival is only possible in the immediate vicinity of existing tree canopy, resulting in a process of nucleation (Debussche & Isenmann 1994), where the only seed rain that contributes to tree growth occurs in the outer edge of the tree canopy. Therefore, only the seeds that fall in the outer 0.5 meters of a tree canopy actually contribute to the growth of recruitment foci in the model. This is represented by Eq. (7). Fg = Af ((Af /)-0.5)2*) / Af (7) where Fg is the part of the focus where seed rain contributes to growth in area and Af is the area of the focus (m2). For simplicity, recruitment foci are assumed to be circular in shape. Eq. (8) describes the actual seed rain (ASR). ASR = PSR*(Af/Oa)* Fg (8) Seed Germination and Seedling Survival Germination rates were determined by planting seeds of two species in various substrates throughout an old field (see Chapter 1). Because the rate of germination and survival was 0.97% and may not have adequately mimicked natural conditions, I compared it to the germination rates obtained by Figueroa and Lusk (2001) for Drimys winterii; these were observed to be ~75% annually in gap conditions, and as low as 16% in low light conditions. However, in this study, seeds were protected from predation. In a study on seed predation in the same area, as many as 65% of seeds were removed by

PAGE 53

41 predators (Daz et al. 1999). Taking a conservative approach, I chose 6% (65% of 16%) as the germination rate for the model. To account for the large difference, I ran a sensitivity analysis to determine the impact of changes in the germination rate (see results section). At rates below 1%, changes in the germination rate have a large impact on rates of forest growth; above 1% the difference is much less. Eq. (9) shows how germination is calculated. G = Sf*Gr (9) where G is germination, Sf is the number of seeds in the recruitment focus, and Gr is the germination rate. The proportion of seedlings that become established as saplings is based on a mortality of 1% per week (for Drimys winterii) under controlled conditions in high light and 4% per week in low light (Lusk & Del Pozo 2002). I took the conservative approach and chose 4% per week to arrive at a number for the model. This value may be low because it was determined under controlled conditions, but I explored this also with sensitivity analyses. Due to the prevalence of standing water in arrested old-fields, the presence of dead wood, acting as nurse-logs, increases the chance of seedling survival. Papic and Armesto (2000) found that survival was 12 times greater on dead wood that on surrounding terrain. This is reflected in the model by setting the death rate for seedlings on dead wood 12 times lower than that of other seedlings. Creation of New Foci In addition to the growth of any existing recruitment foci in the old-field, there is some chance that new foci will be created as time passes. Since birds perch on Baccharis or other small shrubs in the field (Personal Observation) or occasionally defecate while flying, there is a small possibility that a seedling will become established at a new

PAGE 54

42 location in the old-field each month. Determining the value of this likelihood requires further research, and estimating it is quite difficult. However, I made the assumption that there is a 100% chance of a seedling becoming established below a tree in any given year (Slocum 2000), and that the proportion of the various cover types influences this probability of occurrence. Adjusting for area, I then compared the number of seedlings that I found beneath non-tree cover types (Chapter 1) to the number found beneath trees to calculate the odds that a seedling would be established there annually (Table 1). However, the presence of a seedling does not necessarily guarantee a surviving sapling. To further represent the likelihood of sapling establishment, I included the transition probabilities (e.g., Rey & Alcantara 2000) determined by Papic and Armesto (2000) for survival of seedlings. However, because the process of survival from seed to established sapling in the harsh conditions of a cleared old-field is poorly understood, I erred on the side of caution and assumed that only 10% of the sites were suitable for new foci establishment to occur. Eq. (11) shows how the random chance of new foci appearing is handled. IF Af = 0 (on dead wood) THEN MONTECARLO 1 ((0.10/12)*DW)*0.23)*r) or (11) (on bare ground) THEN MONTECARLO ((0.06/12)*BG)*0.02)*r) ELSE 0 where DW is the percent cover of dead wood in the old-field but not part of an existing focus, BG is the percent cover of bare ground in the old-field, and r is the additional 1 Monte Carlo is a function that returns either a 1 or a 0 each time step. The number in parentheses determines the frequency that a 1 is generated.

PAGE 55

43 rarity of new focus formation (r=0.1 in this case). Values are divided by 12 to convert the annual value to the monthly time step used in the model. (To adequately represent multiple foci with the Stella software, it is necessary to array many of the variables in the model. The if-then-else structure of the equation ensures that new foci do not appear where one already exists.) This equation generates a random chance of a seed falling, germinating and surviving, which increases with the presence of dead wood and bare ground. 00.20.40.60.811.21.41.61.8202004006008001000120014001600Area of tree cover in old-field (square meters)Seed rain per square-meter of forested old-field (# seeds) Figure 10. The value of S, or seed rain per meter of forest in the old-field, changes with increasing tree cover in the old-field. This is because the existing seed disperser population becomes less able to distribute seeds as more attractive habitat opens up. Eventually, seed disperser population should increase, allowing the curve to level out. Tree Growth and Survival Ages of 16 trees in old fields were determined by coring in order to estimate the tree growth rate. Tree cores were extracted as close to the ground as possible, then

PAGE 56

44 sanded with progressively finer sand paper. After repeated sanding, growth rings were visible and were counted under a binocular microscope (Stokes et al. 1968). Cores were not cross-dated, so missing and false rings may be a potential source of error (Lusk 1999). Table 1. Comparison of the number of seedlings found beneath non-tree cover types to the number found beneath trees to estimate the odds that a seedling would be established there. Cover type # of seedlings found (x) Proportion of cover type to total area surveyed (p) Number of seedlings in field if entire 5000m2 field were same as cover type ([1/p]*x)=n Seedlings per meter2 (n/5000) Compared to tree Tree 89.00 0.04 2225 0.44 1.00 Dead Wood 7.00 0.03 233 0.05 0.10 Bare Ground 17.00 0.12 141 0.03 0.06 Baccharis 27.00 0.35 77 0.01 0.03 Sphagnum 8.00 0.15 53 0.02 0.04 Tree growth rate is based on the positive correlation found between age and DBH (Fig. 11) and age and height (Fig. 12). The standard tree for this model is a 10cm DBH tree with a canopy covering 1.68 m2 of old-field. This was determined by measuring the canopy cover of 15 adult trees and averaging the area covered. Trees in this model are broken down into 5 size classes; saplings with a DBH < 5cm, and trees with a DBH of 5cm, 10cm, 20cm, and 30cm or more. Between the size classes, canopy cover area is assumed proportional to DBH. The growth rate for D. winteri to each size class is based on commercial forestry data (Navarro et al. 1997).

PAGE 57

45 481216DBH (cm) 1020304050Age of tree (years) R-Square = 0.67 Figure 11. Correlation between age of Drymis winterii (as determined by coring) and diameter at breast height (DBH). Navarro et al. (1997) reported an annual mortality rate for D. winterii of 2.17%, which equals 0.181% mortality per month. The number of trees in each size class is governed by Eq. 12 SCx(t) = SCx(t dt) + (Gin Gout Dx) dt (12) where SCx is the size class at time t, Gin is the input of trees growing to that size class, Gout is the trees growing out of the size class, and Dx is the death of trees. Sensitivity Analyses I conducted sensitivity analyses in order to examine the impacts of changes in key components of the model on model predictions (e.g., Halpern et al. 2005). A base set of conditions representing an average field were used for all sensitivity runs (Table 2). These values were determined by line transect sampling. The proportion of

PAGE 58

46 cover/substrate type was visually estimated at each meter along five 50 meter transects in each of six fields. The proportions of cover and substrate types were then averaged. Table 2. Parameters used for sensitivity analyses. Values represent mean of 6 fields. Parameter Initial value (m2) used in sensitivity analyses Cover of Baccharis 1728.5 m2 Cover of Sphagnum 762.5 m2 Cover of Bare soil 603.5 m2 Cover of Tree (30 trees of size class 2) Cover of Deadwood 312 m2 Cover of Ferns and grasses 1513 m2 Field size 5000 m2 Edge length 100 m Bird density 0.16 birds with access/m of edge Germination rate 0.60% of seeds per month Each parameter in question was then run repeatedly through an incremental series of values centered on the mean value reported in Table 1(e.g., Hallgren & Pitman 2000). Sensitivity analyses that were run are listed in Table 2, and include characteristics of the field and model parameters such as germination rate and rates of competition between forest trees and other cover types Given that field data were collected in old-fields with less than 50% tree cover, it is doubtful the model, parameterized with such data, will sufficiently reflect the population and community dynamics of largely forested sites. Controls on, and relationships among,

PAGE 59

47 bird density, inter-specific competition, and microclimatic conditions are likely to change over time, especially as forest grows and changes the quality of the habitat. Therefore, I do not discuss the ramifications of model predictions regarding regeneration greater than 50%, and I present 75% and 100% forest regeneration model results as points of comparison only. Due to the stochastic nature of the model I report average times until regeneration. Finally, because of the sheer volume of computing time required to run multiple analyses, minimum, median and maximum values were averaged based on 6 runs each. 46810Height of tree (m) 1020304050Age of Tree (years) R-Square = 0.57 Figure 12. Correlation between the age of Drymis winterii (as determined by coring) and tree height. Management Scenarios I applied the model to three management scenarios; a high cost, medium cost, and a do-nothing scenario. The high cost scenario assumes that the land owner has sufficient

PAGE 60

48 labor and money available to add dead wood to the field, plant saplings, and perhaps even burn back some of the Baccharis and Sphagnum. In this scenario, non-forest cover types were reduced by 5% each, CWD cover was increased by 5%, and 25 saplings are planted each year. The medium cost model assumes that the farmer has an active interest in restoring the old-field, but has little resources to do so. This farmer may plant a few trees when the time is right, but will not be able to burn back the shrubs and lichen or drag abundant logs or cut pieces of CWD into the field. In this scenario CWD cover is increased by 1% and 10 saplings are planted annually. The do-nothing, or low cost, scenario assumes that the landowner lacks either resources or interest, and leaves the land as it is. Each management scenario was run with 3 starting conditions poor, medium, and high quality fields. These three field types represent points along the continuum from very low quality fields (no trees, a lot of Baccharis and Sphagnum cover, no CWD), to very high quality fields (many trees remaining, no Baccharis or Sphagnum cover, good CWD coverage, Table 3). In all, 9 starting conditions (3 cost by 3 starting condition scenarios) were run 5 times each. Results Initial Conditions Coarse woody debris On average, increasing the amount of dead wood has a positive impact on the rate of forest regeneration (Fig. 13). However, as more CWD is added, diminishing returns occur, particularly after 1-2% cover of CWD. In fields with zero initial trees, the impact of CWD is more pronounced (Fig. 14).

PAGE 61

49 0501001502002500.00%1.00%2.00%3.00%4.00%5.00%6.00%7.00%Initial CWD in Oldfield (percent cover)Years 25% 50% 75% 100% Figure 13. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial amount of coarse woody debris (CWD) is increased. Baccharis As the initial coverage of Baccharis increases, the time required for forest regeneration increases (Fig. 15). In fields with less than 30% Baccharis cover, little effect is seen, but this is because I am measuring time until 50% regeneration, and at these low values there is little obstacle to regeneration of 50% of the old-field. Above 30% Baccharis cover, the impact is drastic, but is dependent on the values chosen for the rate at which forest shades out Baccharis (see below). Sphagnum and grass Changes in the initial coverage of Sphagnum and of ferns and grasses show a similar pattern to that exhibited by changes in the coverage of Baccharis (Fig. 16 and Fig. 17). However, the percent cover required to slow forest growth corresponds with the rate

PAGE 62

50 at which forest is able to out-compete either cover type. Thus, Sphagnum has less impact than Baccharis, and grass less than Sphagnum. 0501001502002500.00%1.00%2.00%3.00%4.00%5.00%6.00%CWD in Oldfield (percent cover)Years 25% 50% 75% 100% Figure 14. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial amount of coarse woody debris (CWD) is increased with zero initial trees in the old-field. Tree cover Initial tree cover has some effect on rates of subsequent forest regeneration (Fig 18). The time required for 25% and 50% forest regeneration is somewhat reduced by adding more trees. Length of edge Increasing the length of the border with adjacent forest proportionately decreases the time required for forest regeneration because of the increased availability of seed dispersers and seeds. For example, the difference between 125 and 225 meters of edge is far less than the difference between 25 and 125 meters of edge (Fig. 20).

PAGE 63

51 0501001502002503000.00%10.00%20.00%30.00%40.00%50.00%60.00%Initial Baccharis cover (percent cover)Years 25% 50% 75% 100% Figure 15. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial coverage of Baccharis is increased. 0501001502002503000.00%10.00%20.00%30.00%40.00%50.00%60.00%Initial Sphagnum cover (percent cover)Years 25% 50% 75% 100% Figure 16. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial coverage of Sphagnum is increased.

PAGE 64

52 0501001502002503000.00%10.00%20.00%30.00%40.00%50.00%60.00%Initial Grass and Fern (percent cover)Years 25% 50% 75% 100% Figure 17. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial coverage of grasses and ferns is increased. 0501001502002503000100200300400500600Initial Number of trees (20cm>DBH >5cm) in old-fieldYears 25% 50% 75% 100% Figure 18. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial number of trees is increased.

PAGE 65

53 050100150200250300050100150200250Length of EdgeYears 25% 50% 75% 100% Figure 19. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the length of adjacent edge is increased. Bird density As expected, forest regeneration rate is positively correlated with bird density (Fig. 20). The range in bird density values correspond with roughly 150 years in forest regeneration times. Sensitivity Analyses Germination and tree growth rate The variation among rates lower than 20% is much greater than among rates between 20 and 70% (Fig. 21). The value for the germination rate was difficult to obtain, and additional research is needed before the model should be used to make predictions. Field observations (Chapter 1) revealed a low germination rate of 0.98%. Conversely, controlled experiments (Figueroa & Lusk 2001; Figueroa 2003) found a germination rate for the same tree species of 75%. A reasonable compromise would return a germination rate of 35.49%, but as the model demonstrates, this is too high because at such high

PAGE 66

54 germination rates, forest regenerates much faster than historical observations suggest. I conservatively chose a 6% rate of germination for the model. When subjected to sensitivity analysis, little change was seen for germination rates above 1% (Fig. 21). There is a ~125 year variation in time until 50% regeneration between 0% and 1% (Fig. 21). This extreme variation demonstrates that getting more accurate estimates of this value is critical to the successful prediction of forest regeneration. As seen in Figure 22, regeneration times vary proportionately to tree growth rates. 0501001502002503003500.050.10.150.20.250.3Initial Density of frugivorous birds in adjacent forestYears 25% 50% 75% 100% Figure 20. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the initial number of frugivorous birds is increased. Survival rate of seedlings on dead wood According to Papic and Armesto (2000), survival of seedlings is 12 times higher on CWD than on adjacent ground. However, variation in this variable had no impact on the outcome of the model (Fig. 23). Overall seedling mortality had a minor impact on time

PAGE 67

55 until forest regeneration (Fig. 24). The higher the seedling mortality the longer the time required until forest regeneration. 05010015020025030000.10.20.30.40.50.60.70.8Germination Rate (annual proporton of seeds)Years 25% 50% 75% 100% Figure 21. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the germination rate is increased. Rate of cover loss to forest I parameterized the model to allow for the possibility that forest trees modify the microclimate such that it becomes unsuitable for Baccharis, Sphagnum, or grasses. However, at high rates of competition with other substrate types, the forest does not grow fast enough to take advantage of the improved habitat conditions (Figs. 25-27). On the other hand, if the time required for forest tree species to out-compete cover types such as Sphagnum and Baccharis is very high (i.e. very low rates of cover loss), it has profound effects on the model, therefore, this relationship could dictate whether or not old-fields are in a state of arrested succession. For example, if Baccharis is lost any faster than 0.008m2 per month, there is no corresponding increase in forest cover because forest

PAGE 68

56 does not grow any faster to take advantage of the increased Baccharis loss (Fig. 25). At a value of 0.008 m2 for Baccharis reduction, it takes roughly 12 years for a tree to shade out 1 m2 of adjacent Baccharis. Although these rates of loss have no effect on time until 50% regeneration in the sensitivity analyses (i.e. average levels), if a particular field has a very large area covered with Baccharis or Sphagnum, changes in competition rates can be very important. 0501001502002503000.010.0150.020.0250.030.0350.04Rate of tree canopy growth (square meters / year)Years 25% 50% 75% 100% Figure 22. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of tree growth is increased or decreased. Rarity of new recruitment-foci formation For the model runs I chose an r (rarity of new focus formation) value of 0.1. As this value was chosen purely to keep the model estimations conservative, it is important

PAGE 69

57 to subject it to sensitivity analyses. Figure 28 demonstrates that this value directly influences the rate of forest regeneration, especially at lower values. 05010015020025030000.020.040.060.080.10.120.140.160.18Seedling mortality on CWD (proportion of seedlings / month)Years 25% 50% 75% 100% Figure 23. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the mortality of seedlings on dead wood is increased. Default mortality of seedlings is 0.16% seedlings / month. Management scenarios High quality fields with high cost input (additions of both CWD and saplings) show the fastest regeneration times, followed by medium quality fields with high-cost input. All field types show significant benefits from increasing the cost of input, however low quality fields require high-cost inputs to achieve any significant regeneration at all (Table 3). Discussion Relative Importance of Seed Dispersal and Germination Limitation When either seed dispersal or suitable germination microsites were set to very low values (limiting), the outcomes suggest that germination site availability is potentially

PAGE 70

58 more limiting of forest regeneration. At very low rates of seed dispersal forest regenerates significantly faster with adequate availability of suitable germination sites than in situations with very few germination sites. My results suggest that during the initial stages of reclaiming established Baccharis fields, germination sites may be the more important limiting factor (Fig. 29). My results agree generally with similar studies in that both dispersal and germination are limiting factors (Costa Rican abandoned pasture Holl et al, (2000); Puerto Rico Zimmerman et al. (2000). Unlike these studies, however, my modeling effort allowed me to explore the relative importance of one factor versus the other over a range of realistic values of starting conditions. 0501001502002503000.070.120.170.22Seedling mortality rate (proportion of seeds that die / month)Years 25% 50% 75% 100% Figure 24. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of seedling mortality increases.

PAGE 71

59 05010015020025030000.0050.010.0150.020.0250.030.035Square meters of baccharis cover lost per month per meter of new forest edgeYears 25% 50% 75% 100% Figure 25. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of out-competition of Baccharis is increased. 05010015020025000.0020.0040.0060.0080.010.012Square meters of sphagnum lost per meter of new forest edge/monthYears 25% 50% 75% 100% Figure 26. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of out-competition of Sphagnum is increased.

PAGE 72

60 Table 3. Field conditions, input, and time required for 50% regeneration under 3 different management scenarios and three different starting conditions. (HC= High Cost, MC= Medium Cost, LC = Low Cost, HQ = High Quality, MQ = Medium Quality, LQ = Low Quality). High Cost Medium Cost Low Cost HQ MQ LQ HQ MQ LQ HQ MQ LQ Sphagnum (% cover) 0% 5% 14.5% 1% 10% 19.5% 1% 10% 19.5% Baccharis (% cover) 0% 10% 25% 1% 15% 30% 1% 15% 30% Grass (% cover) 1% 10% 0% 1% 10% 0% 1% 10% 0% Dead Wood (% cover) 10% 6% 5% 6% 2% 1% 5% 1% 0% Saplings Planted/yr 25 25 25 10 10 10 0 0 0 Initial trees 100 50 0 100 50 0 100 50 0 Average time until 50% forest cover (years) 42 43 65 48 53 no regen 103 140 No regen I found avian frugivores to be critical for sustaining regeneration of old-fields throughout the progression of forest accretion to recruitment foci (Fig. 20). In this way, the model suggests that the most rapid regeneration will not be achieved without constant, ample seed inputs from birds visiting the sites. As fragmentation, forest loss, and degradation of remnant forest patches intensifies, inadequate seed rain could become limiting as frugivore density and activity declines. Currently, however, the high abundance of Elaenia albiceps (et al. 2005) and the species apparent tolerance for disturbance, ensure they are likely to be available at sites I studied (and most disturbed sites in Chilo) to provide seed input, as long as there are recruitment foci available to

PAGE 73

61 attract them into fields. In the case of reliable dispersal, the importance of both germination rate and availability of germination substrates (Figs. 15-17) including coarse woody debris (Fig. 13), may be relatively higher in determining variations in observed regeneration rate. But if seed dispersal ceases, then reclamation of Baccharis fields is unlikely to occur without intense and costly manipulations (see below). 05010015020025000.010.020.030.040.050.06Square meters of grass and fern lost per meter of new forest edge/monthYear 25% 50% 75% 100% Figure 27. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the rate of out-competition of grass and fern is increased. Restoration Methods / Management Scenarios Assessment of model scenarios Recommendations for restoring old-fields will vary somewhat according to the management scenario involved. The three field types modeled represent points along the continuum from very low quality fields (no trees, a lot of Baccharis and Sphagnum cover, no CWD), to very high quality fields (many trees remaining, no Baccharis or Sphagnum cover, good CWD coverage). Very high and very low quality fields require a different approach to ensure regeneration (Table 3). High quality fields (i.e. fields not fully

PAGE 74

62 cleared or burned, or already undergoing succession) can be essentially left alone, and regeneration will still occur, although according to the model it will still require roughly 100 years. In other words, high quality fields would represent fields in which succession is not arrested. On the other hand, in low-quality fields, no regeneration occurs without significant input, including a high level of sapling planting (Table 3). The model suggests that the initial starting conditions have a large impact on the outcome of restoration efforts. Whereas increasing the input from a medium cost scenario to a high cost scenario has little impact on regeneration times for high and medium quality fields (5 or 10 year improvements), 50% regeneration does not occur within 250 years in poor quality fields without high cost inputs. Therefore, if a farmer has no resources, and wants to accelerate regeneration, he needs to avoid burning when clearing for pasture and leave fruiting tree species scattered throughout the field. Whereas, if a farmer has sufficient resources, quick regeneration can be obtained in a low quality field if labor and money are expended to plant trees and drag wood into the field (as long as it is close to forest with dense bird populations). Recommendations for efficient reclamation of Baccharis fields Based on this modeling effort and exploration of scenarios derived from discussions with landowners regarding their experiences, an effective restoration strategy would need to concentrate in the following areas in the order that I consider them here. (1) Site location. Areas selected for forest regeneration should be adjacent to mid-successional or old growth forest stands in order to ensure sufficient visitation and seed dispersal by Elania albiceps (Daz et al. 2005). Seed dispersal into old-fields clearly helps sustain any attempts at restoration, whether a landowner is relying principally on manipulations (additions of CWD and plantings) or natural processes (Figs. 19 and 20).

PAGE 75

63 A landowner without resources for restoration, but with a variety of potential sites and flexibility in choosing which sites can be allowed to go back to forest is not without recourse. A high quality field (with trees and CWD) near a proper forest edge can be set aside for forest regeneration and has a high likelihood of becoming a forest without any manipulation, though even a small number of plantings could help speed the process considerably (Table 3). 05010015020025000.20.40.60.811.2Proportion of new foci establishedTime till regeneration 25% 50% 75% 100% Figure 28. Model prediction of the number of years required to achieve 25, 50, 75 and 100% regeneration of degraded old-field habitat as the value of r (additional rarity of new recruitment foci establishment) is increased. The best time to make the decision regarding placement of a potential regeneration site is before complete forest clearing occurs. If, for example, only partial clearing and no burning was done, followed by an observation period of a year; an assessment of a sites best use could be accomplished while the site was still of high quality (in terms of regeneration potential). If the physical conditions of the cleared site proved to foster Baccharis rather than drier land use options (e.g., pasture) then the site should be left to regenerate to forest while the landowner assessed other sites for non-forest land uses.

PAGE 76

64 This kind of assessment would save a great deal of wasted labor; as landowners we work with relate that certain Baccharis-dominated sites (apparently, the most mesic ones) can be cleared and burned repeatedly without successful establishment of pasture or cropland. After repeated burning and complete clearing, these sites rapidly reach states of stable Baccharis-Sphagnum cover. My work suggests this occurs because such sites are devoid of suitable germination substrates. Thereafter, these sites remain in an unproductive stable state of no, or only marginal, use to humans and wildlife (Darnell and Sieving, in press, Daz et al. 2005). Therefore, intensive clearing and burning of such sites, without timely assessment of future use, leads directly to more forest clearing and the spread of Baccharis. To my knowledge, there is no way of determining, prior to beginning forest clearing, what post-clearing site characteristics will be with respect to Baccharis formation, although some research is beginning to address this idea in other systems (Chanasyk et al. 2003). Therefore, I put forward the recommendation to partially clear and then wait to see what the site tendencies appear to be. (2) Germination sites. Given a proper location close to forest to insure adequate dispersal, then the next focus should be on providing sufficient germination substrates. Depending on the number of trees, amount of CWD, and shrub cover of the old-field, this could entail the addition of CWD and/or the removal of Baccharis and Sphagnum. The positive impacts of CWD are significant (Papic & Armesto 2000, suggested by model). However, the degree of benefit obtained from adding dead wood to a site depends on initial condition of the field. Fields show a slight benefit from an increase in the initial amount of dead wood, likely because any impact that dead wood might have as sites for new foci is swamped by positive germination conditions provided by existing recruitment

PAGE 77

65 foci. However, in fields with no trees initially, the impact of CWD as nurse logs is profound (Fig. 14). For example, with 0 trees, the time required for regeneration of 50% forest cover decreased from 200 years with no dead wood, to roughly 100 years when 2% of the field is covered with CWD. The importance of CWD in ecosystem function is no longer underestimated (Carmona et al. 2002), especially in facilitation of recruitment of new trees (McGee & Birmingham 1997; Slocum 2000). 050100150200250Mean time to5%Mean time to25%Mean time to50% Years Default values Low bird density (.02birds in adjacent forestper meter of edge) Low substrateavailabilty (0 CWD,500 m bare ground) Figure 29. Time predicted for regeneration of forest (to 5, 25, and 50% forest cover) with low densities of seed dispersers, compared with time required for regeneration with very few micro-sites suitable for germination, compared to default values of each (see text). Error bars represent 1 SD. (3) Shrub and Sphagnum management. Even though, in some cases, shrubs can act as facilitators to forest regeneration (Duncan & Chapman 2003; Holl 2002; Li & Wilson 1998; Zahawl & Auspurger 1999), they often have the opposite effect. Shrubs inhibit seedling growth in Appalachian canopy gaps (Beckage et al. 2000). Similarly, Denslow et al. (1991) found evidence for inhibition of tree seedling establishment by broad-leaved understory plants in the tropics. Finally, Hill et al. (1995) found that shrub canopies

PAGE 78

66 along utility rights of way in New York were very resistant to invasion. Therefore, in the context of restoring Baccharis fields given the presence of both avian seed dispersers and suitable microsites for germination, the single most important deterrent to forest succession is the presence of Sphagnum and Baccharis. This is because seeds simply cannot germinate and survive in direct competition with these two cover types (Table 3, Figs. 15 and 16). Instead of burning away the shrub and moss layer which can only be done during the two driest summer months and may destroy CWD and seedlings perhaps the most parsimonious approach is to reduce competition with these cover types at the edges of recruitment foci by physical means (cutting shrubs and trampling at the perimeter of recruitment foci). Providing a modest sized open buffer of suitable germination substrate (e.g., bare ground, or modest CWD) around trees and tree clusters would release one of the main inhibitors of the rate of regeneration suggested by the model; competitive effects at the edges of recruitment foci (Figs. 15, 16, and 17), and may be less labor-intensive than manipulations at the scale of whole fields. (4) Planting trees. Finally, I would suggest actively planting new trees in old-fields. The model demonstrates that planting just a few trees each year can have a large impact on forest regeneration (Fig. 30). Reay and Norton (1999) found that without plantings, restoration in a New Zealand temperate forest proceeded at a much slower pace. According to the model, an average field (as defined above; see Methods: Sensitivity Analyses) will require 157 years to reach 50% tree canopy cover. Planting 10 trees per ha per year will reduce the time to 83 years, as predicted by my model. Further increasing the rate of tree planting does little to decrease the regeneration rate (Fig. 30), likely due to limitations on the growth rate of trees, and competition from Baccharis and

PAGE 79

67 Sphagnum. However, as discussed in Chapter 1, there may be a threshold value for the number of trees required to lure avian seed dispersers into old-fields. Further research is required before this can be incorporated into the model. Although this model overlooks potential differences between species, planting trees of a variety of forest species and paying attention to the facilitative effects of various species may be very important (Jansen 1997). 0501001502002500102030405060Number of saplings planted per yearYears 25% 50% 75% 100% Figure 30. Model prediction of the effect of the number of saplings planted annually on the rate of forest regeneration (to 25, 50, 75, 100% forest cover). Finally, another type of planting, not addressed by my work, is of the native bamboo (Chusquea valdiviensis), which is very attractive to native understory birds (Reid et al. 2004). Large patches of bamboo occur naturally associated with all seral stages of south-temperate rainforest (Donoso 1996). It would be worthwhile investigating the influence of bamboo on tree establishment in these fields due to its potential to provide habitat immediately. However, the benefits of bamboo as a disperser attractant may be offset by its negative impact on seedling establishment (Donoso & Nyland 2005).

PAGE 80

68 Assessment of the Model Complexity While it is far from a complete description of the south temperate rainforest ecosystem, this model acts as a good starting point for future research and hypothesis generation. By necessity the model is far simpler than the ecosystem that it represents. All tree-related data were obtained regarding only one species (Drimys winterii). This species is not a poor choice, however, because it is one of the most common bird dispersed colonizers of old-fields (Armesto et al. 2001a) and occurs in all seral stages of forest in this region (Lusk & Del Pozo 2002). Models of greater complexity are in use for similar studies. For example, LANDIS a spatially explicit landscape model (Mladenoff et al. 1996) could incorporate a variety of tree and understory species with unique demographic characteristics into regeneration scenarios. Additionally, a more accurate (and much more complex) model might factor in such variables as carbon, other nutrient and water cycles and energy flows (Kirschbaum 1999). However, by maintaining simplicity, I was able to focus on parameters relevant to my original research question concerning the relative importance of seedling establishment and seed dispersal. Moreover, there is some merit in keeping the model simple and utilizing statistical methods (e.g., Monte Carlo method) to model the stochastic properties inherent in complex systems (Young et al. 1996). Relevant scales Because most key ecological processes of forest regeneration at edges take place within 50 meters of the forest, the effects of larger spatial scales were ignored, even though seed dispersers, for example, operate at larger scales. While I was constrained in my field approaches (Chapter 1) to work on processes relevant to field-forest ecotones, I

PAGE 81

69 do not think my conclusions concerning restoration of old-fields in general are inaccurate because forest regeneration in old fields tends to occur through accretion at edges of recruitment foci and remnant patches (Debussche & Isenmann 1994). In order to incorporate larger-scale site characteristics, like attenuation in seed dispersal with distance from forest, further study of avian frugivore movements would have been required. I deemed this unnecessary because in the landscape where I worked, most Baccharis fields were not more than 250-300 m wide. Frugivorous birds readily and quickly crossed these open areas between patches of forest, so it is unlikely that dispersal was dramatically different throughout the fields. In order to select sites that were equivalent in proximity to forest patches and other land use types, the plot size I used to collect data was easy to standardize from a design perspective. Perhaps the greatest caution in using the model is relevant to the combined effects of the long time scales involved and the potential that (over such long times) factors originating beyond the spatial scale of the site could be significantly involved in determining final outcomes (Parker 1997). Model improvement One major assumption of the model is that Baccharis, Sphagnum and grass are static features of the old-field (i.e. they do not increase in area). In reality they are likely to increase in cover area when not in direct competition with trees, but including this dynamic was not feasible. However, this simplification could overlook important dynamics of the system (Duncan & Chapman 2003; Holl 2002), and is deserving of further work. Finally, I suggest that the influence of standing water be incorporated into modeling efforts, and that greater understanding of the hydrology of these sites is

PAGE 82

70 necessary. Currently, standing water is assumed to be a factor leading to the increased survival rate of seedlings on CWD. However, an analysis of historic weather trends, combined with study of germination rates of immersed seeds may better illuminate the role that the hydrology of the system plays in forest regeneration. Moreover, it also seems important to establish the validity of a central assumption we make, based in part on observations, that the water table is affected by forest clearing. We assume that one of the reasons landowners clear so many areas that become wet and shrub-dominated is that a high water table is not evident prior to clearing, but that the water table often rises following clearing (e.g., Sun et al. 2000). This could occur in sites dominated by tree species with high transpiration rates. Other sites, with lower rates of transpiration, may be more easily assessed as to post-clearing hydrology. Further knowledge is needed regarding site characteristics that can predict the outcome of forest clearing (e.g., Chanasyk et al. 2003). If arrested succession can be avoided by informed choices before forest clearing, this would reduce the overall impact on an already fragmented ecosystem. For fields already in an arrested state, it is hoped that this research and further research will shed some light on returning old fields to forests that provide benefits to both humans and wildlife.

PAGE 83

APPENDIX A ICONOGRAPHIC REPRESENTATION OF MODEL OLD-FIELD CHARACTERISTICS L Oa PSR max length of edge seed pool Fa Ga undispersed seeds Ca P DW Ga 2 Foci Circumference SPcov SPloss Oa Bloss GFloss BG Ca Bcov GFcov P SP loss rate B loss rate GF loss rate 71

PAGE 84

72 SEED RAIN PSR 2 ~ Sr seed pool 2 ASR Sf seed death Af + total attractive trees undispersed seeds 2 trees in foci Db Fg SEED GERMINATION AND SEEDLING SURVIVAL Sf G seedlings seed death transition growth DCWDseedling seedling bank on nurselogs transition dw CWD death rate BG CWD Growth Dseedling dw proportion Gr seedling bank BG non dead wood proportion total dw in foci seedling mortality

PAGE 85

73 CREATION OF NEW RECRUITMENT FOCI growth saplings G1 seedling bank on nurselogs random chance of new foci appearing CWD Growth sapling death DW random chance new foci dead wood seedling bank r TREE GROWTH AND SURVIVAL SC1 sapling SC4 SC2 SC3 G1 G2 G3 G4 D1 death rate D4 decay D2 dead wood in foci D3 dead trees 4 growth modifier decayed dead trees decay of wood BG dead trees 2 dead trees 3 decay 2 dead wood in foci 2 decayed 2 decay of wood 2 decay 3 dead wood in foci 3 decayed 3 decay of wood 3 decay 4 dead wood in foci 4 decayed 4 decay of wood 4 total dw in foci

PAGE 86

74

PAGE 87

APPENDIX B MODEL EQUATIONS OLD-FIELD CHARACTERISTICS L(t) = L(t dt) INIT L = 0 Oa(t) = Oa(t dt) INIT Oa = 0 Ca = Fa/Oa Db = .15625 Fa = IF (total_attractive_trees+.00000000000001)*1.68 >Ga THEN Ga ELSE (total_attractive_trees+.00000000000001)*1.68 max_length_of_edge = (SQRT(Oa))*4 Bcov(t) = Bcov(t dt) + (Bloss) dt INIT Bcov = 0 OUTFLOWS: Bloss = IF P > 0.999 THEN ARRAYSUM(Fc[*])*(B_loss_rate) ELSE 0 GFcov(t) = GFcov(t dt) + (GFloss) dt INIT GFcov = 0 OUTFLOWS: GFloss = IF P > 0.999 THEN ARRAYSUM(Fc[*])*GF_loss_rate ELSE 0 SPcov(t) = SPcov(t dt) + (SPloss) dt INIT SPcov = 0 OUTFLOWS: SPloss = IF P > 0.999 THEN ARRAYSUM(Fc[*])*(SP_loss_rate) ELSE 0 BG = Oa-(Bcov+SPcov+GFcov+ (Ca*Oa)) B_loss_rate = 1/288 75

PAGE 88

76 Ga = IF (Oa-(Bcov+GFcov+SPcov)) <= 0 THEN .000000001 ELSE (Oa-(Bcov+GFcov+SPcov)) GF_loss_rate = 1/48 SP_loss_rate = 1/180 SEED RAIN seed_pool(t) = seed_pool(t dt) + (PSR ASR[1] ASR[2] ASR[3] ASR[4] ASR[5] ASR[6] ASR[7] ASR[8] ASR[9] ASR[10] ASR[11] ASR[12] ASR[13] ASR[14] ASR[15] ASR[16] ASR[17] ASR[18] ASR[19] ASR[20] ASR[21] ASR[22] ASR[23] ASR[24] ASR[25] ASR[26] ASR[27] ASR[28] ASR[29] ASR[30] ASR[AllFoci] undispersed_seeds) dt INIT seed_pool = 0 INFLOWS: PSR = (Sr*Oa)*(L/100)*(Db/.15625) OUTFLOWS: ASR[AllFoci] = (seed_pool*((Af[AllFoci]/Oa)))*Fg[AllFoci] undispersed_seeds = seed_pool-(ARRAYSUM(ASR[*])) Sf[AllFoci](t) = Sf[AllFoci](t dt) + (ASR[AllFoci] G[AllFoci] seed_death[AllFoci]) dt INIT Sf[AllFoci] = 0 INFLOWS: ASR[AllFoci] = (seed_pool*((Af[AllFoci]/Oa)))*Fg[AllFoci] OUTFLOWS: seed_death[AllFoci] = (Sf[AllFoci]-G[AllFoci]) Af[AllFoci] = trees_in_foci[AllFoci]*1.68 Fg[AllFoci] = (Af[AllFoci](((SQRT(Af[AllFoci]/PI))-0.5)^2)*PI)/(Af[AllFoci]+.0000000000000001) Sr = GRAPH(total_attractive_trees) (0.00, 1.82), (36.6, 1.82), (73.2, 1.82), (110, 1.79), (146, 1.76), (183, 1.73), (220, 1.70), (256, 1.67), (293, 1.64), (329, 1.61), (366, 1.58), (402, 1.55), (439, 1.52), (476, 1.49), (512, 1.46), (549, 1.43), (585, 1.40), (622, 1.37), (659, 1.34), (695, 1.31), (732, 1.28), (768, 1.25), (805, 1.22), (841, 1.19), (878, 1.16), (915, 1.13), (951, 1.10), (988, 1.07), (1024, 1.04), (1061, 1.01), (1098, 0.98), (1134, 0.95), (1171, 0.92), (1207, 0.89), (1244,

PAGE 89

77 0.86), (1280, 0.83), (1317, 0.8), (1354, 0.77), (1390, 0.74), (1427, 0.71), (1463, 0.68), (1500, 0.65) SEED GERMINATION AND SEEDLING SURVIVAL seedlings[AllFoci](t) = seedlings[AllFoci](t dt) + (G[AllFoci] transition[AllFoci] transition_dw[AllFoci]) dt INIT seedlings[AllFoci] = 0 INFLOWS: G[AllFoci] = IF BG < 1 THEN 0 ELSE ((Sf[AllFoci]*Gr)) OUTFLOWS: transition[AllFoci] = ((seedlings[AllFoci]*non_dead_wood_proportion[AllFoci])) transition_dw[AllFoci] = seedlings[AllFoci]*dw_proportion[AllFoci] seedling_bank[AllFoci](t) = seedling_bank[AllFoci](t dt) + (transition[AllFoci] growth[AllFoci] Dseedling[AllFoci]) dt INIT seedling_bank[AllFoci] = 0 INFLOWS: transition[AllFoci] = ((seedlings[AllFoci]*non_dead_wood_proportion[AllFoci])) OUTFLOWS: growth[AllFoci] = (IF BG < 1 THEN 0 ELSE seedling_bank[AllFoci]/12)+(Dseedling[AllFoci]*0) Dseedling[AllFoci] = (seedling_mortality)*seedling_bank[AllFoci] seedling_bank_on_nurselogs[AllFoci](t) = seedling_bank_on_nurselogs[AllFoci](t dt) + (transition_dw[AllFoci] DCWDseedling[AllFoci] CWD_Growth[AllFoci]) dt INIT seedling_bank_on_nurselogs[AllFoci] = 0 INFLOWS: transition_dw[AllFoci] = seedlings[AllFoci]*dw_proportion[AllFoci] OUTFLOWS: DCWDseedling[AllFoci] = ((seedling_mortality/CWD_death_rate))*seedling_bank_on_nurselogs[AllFoci] CWD_death_rate = 12 dw_proportion[AllFoci] = total_dw_in_foci[AllFoci]*2/(Af[AllFoci] +.000000000001) Gr = 0

PAGE 90

78 non_dead_wood_proportion[AllFoci] = (1-dw_proportion[AllFoci]) seedling_mortality = .04 CREATION OF NEW RECRUITMENT FOCI random_chance_new_foci_dead_wood[AllFoci] = (IF Af[AllFoci] = 0 THEN MONTECARLO((0.103/12)*DW*0.23*((L/100)*(Db/0.15625)*r)) ELSE 0) INFLOW TO: sapling (IN SECTOR: Tree Growth and Survival) random_chance_of_new_foci_appearing[AllFoci] = (IF Af[AllFoci] = 0 THEN MONTECARLO((0.063/12)*BG*0.02*((L/100)*(Db/0.15625)*r)) ELSE 0) INFLOW TO: sapling (IN SECTOR: Tree Growth and Survival) DW = 0 r = .1 TREE GROWTH AND SURVIVAL dead_trees[AllFoci](t) = dead_trees[AllFoci](t dt) + (D1[AllFoci] decay[AllFoci]) dt INIT dead_trees[AllFoci] = 0 INFLOWS: D1[AllFoci] = death_modifier*.89*SC1[AllFoci] OUTFLOWS: decay[AllFoci] = dead_trees[AllFoci]/120 dead_trees_2[AllFoci](t) = dead_trees_2[AllFoci](t dt) + (D2[AllFoci] decay_2[AllFoci]) dt INIT dead_trees_2[AllFoci] = 0 INFLOWS: D2[AllFoci] = death_modifier*.94*SC2[AllFoci] OUTFLOWS: decay_2[AllFoci] = dead_trees_2[AllFoci]/120 dead_trees_3[AllFoci](t) = dead_trees_3[AllFoci](t dt) + (D3[AllFoci] decay_3[AllFoci]) dt INIT dead_trees_3[AllFoci] = 0 INFLOWS: D3[AllFoci] = death_modifier*1.06*SC3[AllFoci] OUTFLOWS: decay_3[AllFoci] = dead_trees_3[AllFoci]/120

PAGE 91

79 dead_trees_4[AllFoci](t) = dead_trees_4[AllFoci](t dt) + (D4[AllFoci] decay_4[AllFoci]) dt INIT dead_trees_4[AllFoci] = 0 INFLOWS: D4[AllFoci] = (death_modifier)*1.11*SC4[AllFoci] OUTFLOWS: decay_4[AllFoci] = dead_trees_4[AllFoci]/120 dead_wood_in_foci[AllFoci](t) = dead_wood_in_foci[AllFoci](t dt) + (decay[AllFoci] decay_of_wood[AllFoci]) dt INIT dead_wood_in_foci[AllFoci] = 0 INFLOWS: decay[AllFoci] = dead_trees[AllFoci]/120 OUTFLOWS: decay_of_wood[AllFoci] = dead_wood_in_foci[AllFoci]/120 dead_wood_in_foci_2[AllFoci](t) = dead_wood_in_foci_2[AllFoci](t dt) + (decay_2[AllFoci] decay_of_wood_2[AllFoci]) dt INIT dead_wood_in_foci_2[AllFoci] = 0 INFLOWS: decay_2[AllFoci] = dead_trees_2[AllFoci]/120 OUTFLOWS: decay_of_wood_2[AllFoci] = dead_wood_in_foci_2[AllFoci]/120 dead_wood_in_foci_3[AllFoci](t) = dead_wood_in_foci_3[AllFoci](t dt) + (decay_3[AllFoci] decay_of_wood_3[AllFoci]) dt INIT dead_wood_in_foci_3[AllFoci] = 0 INFLOWS: decay_3[AllFoci] = dead_trees_3[AllFoci]/120 OUTFLOWS: decay_of_wood_3[AllFoci] = dead_wood_in_foci_3[AllFoci]/120 dead_wood_in_foci_4[AllFoci](t) = dead_wood_in_foci_4[AllFoci](t dt) + (decay_4[AllFoci] decay_of_wood_4[AllFoci]) dt INIT dead_wood_in_foci_4[AllFoci] = 0 INFLOWS: decay_4[AllFoci] = dead_trees_4[AllFoci]/120

PAGE 92

80 OUTFLOWS: decay_of_wood_4[AllFoci] = dead_wood_in_foci_4[AllFoci]/120 decayed[AllFoci](t) = decayed[AllFoci](t dt) + (decay_of_wood[AllFoci]) dt INIT decayed[AllFoci] = 0 INFLOWS: decay_of_wood[AllFoci] = dead_wood_in_foci[AllFoci]/120 decayed_2[AllFoci](t) = decayed_2[AllFoci](t dt) + (decay_of_wood_2[AllFoci]) dt INIT decayed_2[AllFoci] = 0 INFLOWS: decay_of_wood_2[AllFoci] = dead_wood_in_foci_2[AllFoci]/120 sapling[AllFoci](t) = sapling[AllFoci](t dt) + (growth[AllFoci] + input[AllFoci] + CWD_Growth[AllFoci] + random_chance_new_foci_dead_wood[AllFoci] + random_chance_of_new_foci_appearing[AllFoci] G1[AllFoci] sapling_death[AllFoci]) dt INIT sapling[AllFoci] = 0 INFLOWS: growth[AllFoci] (IN SECTOR: Seed Germination and Seedling Survival) input[1] = annual_sapling_addition/180 input[2] = annual_sapling_addition/180 input[3] = annual_sapling_addition/180 input[4] = annual_sapling_addition/180 input[5] = annual_sapling_addition/180 input[6] = annual_sapling_addition/180 input[7] = annual_sapling_addition/180 input[8] = annual_sapling_addition/180 input[9] = annual_sapling_addition/180 input[10] = annual_sapling_addition/180 input[11] = annual_sapling_addition/180 input[12] = annual_sapling_addition/180 input[13] = annual_sapling_addition/180 input[14] = annual_sapling_addition/180 input[15] = annual_sapling_addition/180 input[16] = annual_sapling_addition*0 input[17] = annual_sapling_addition*0 input[18] = annual_sapling_addition*0 input[19] = annual_sapling_addition*0 input[20] = annual_sapling_addition*0 input[21] = annual_sapling_addition*0 input[22] = annual_sapling_addition*0 input[23] = annual_sapling_addition*0

PAGE 93

81 input[24] = annual_sapling_addition*0 input[25] = annual_sapling_addition*0 input[26] = annual_sapling_addition*0 input[27] = annual_sapling_addition*0 input[28] = annual_sapling_addition*0 input[29] = annual_sapling_addition*0 input[30] = annual_sapling_addition*0 CWD_Growth[AllFoci] (Not in a sector) random_chance_new_foci_dead_wood[AllFoci] (IN SECTOR: Chance of new foci appearing) random_chance_of_new_foci_appearing[AllFoci] (IN SECTOR: Chance of new foci appearing) OUTFLOWS: G1[AllFoci] = (IF BG < 1 THEN 0 ELSE sapling[AllFoci]/292.75)*growth_modifier sapling_death[AllFoci] = .0083*sapling[AllFoci] SC1[AllFoci](t) = SC1[AllFoci](t dt) + (G1[AllFoci] G2[AllFoci] D1[AllFoci]) dt INIT SC1[AllFoci] = 0 INFLOWS: G1[AllFoci] = (IF BG < 1 THEN 0 ELSE sapling[AllFoci]/292.75)*growth_modifier OUTFLOWS: G2[AllFoci] = (IF BG < 1 THEN 0 ELSE SC1[AllFoci]/342.75)*growth_modifier D1[AllFoci] = death_modifier*.89*SC1[AllFoci] SC2[AllFoci](t) = SC2[AllFoci](t dt) + (G2[AllFoci] G3[AllFoci] D2[AllFoci]) dt INIT SC2[AllFoci] = 0 INFLOWS: G2[AllFoci] = (IF BG < 1 THEN 0 ELSE SC1[AllFoci]/342.75)*growth_modifier OUTFLOWS: G3[AllFoci] = (IF BG < 1 THEN 0 ELSE SC2[AllFoci]/687.5)*growth_modifier D2[AllFoci] = death_modifier*.94*SC2[AllFoci] SC3[AllFoci](t) = SC3[AllFoci](t dt) + (G3[AllFoci] G4[AllFoci] D3[AllFoci]) dt INIT SC3[AllFoci] = 0

PAGE 94

82 INFLOWS: G3[AllFoci] = (IF BG < 1 THEN 0 ELSE SC2[AllFoci]/687.5)*growth_modifier OUTFLOWS: G4[AllFoci] = (IF BG < 1 THEN 0 ELSE SC3[AllFoci]/1075)*growth_modifier D3[AllFoci] = death_modifier*1.06*SC3[AllFoci] SC4[AllFoci](t) = SC4[AllFoci](t dt) + (G4[AllFoci] D4[AllFoci]) dt INIT SC4[AllFoci] = 0 INFLOWS: G4[AllFoci] = (IF BG < 1 THEN 0 ELSE SC3[AllFoci]/1075)*growth_modifier OUTFLOWS: D4[AllFoci] = (death_modifier)*1.11*SC4[AllFoci] annual_sapling_addition = 0 death_modifier = 0.00181 number_of_foci = (IF trees_in_foci[1] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[2] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[3] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[4] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[5] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[6] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[7] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[8] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[9] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[10] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[11] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[12] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[13] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[14] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[15] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[16] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[17] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[18] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[19] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[20] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[21] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[22] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[23] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[24] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[25] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[26] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[27] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[28] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[29] > 0 THEN 1 ELSE 0) + (IF trees_in_foci[30] > 0 THEN 1 ELSE 0) total_dw_in_foci[AllFoci] = ((dead_wood_in_foci[AllFoci]*.5)+dead_wood_in_foci_2[AllFoci]+(dead_wood_in_foci_3[AllFoci]*2)+(dead_wood_in_foci_4[AllFoci]*3)) trees_in_foci[AllFoci] = ((dead_trees[AllFoci]+dead_wood_in_foci[AllFoci]+SC1[AllFoci]+decayed[AllFoci])*.5)+((dead_trees_2[AllFoci]+dead_wood_in_foci_2[AllFoci]+decayed_2[AllFoci]+SC2[AllFoci]))+((dead_trees_3[AllFoci]+dead_wood_in_foci_3[AllFoci]+decayed_3[AllFoci

PAGE 95

83 ]+SC3[AllFoci])*2)+((dead_trees_4[AllFoci]+dead_wood_in_foci_4[AllFoci]+decayed_4[AllFoci]+SC4[AllFoci])*3) Not in a sector decayed_3[AllFoci](t) = decayed_3[AllFoci](t dt) + (decay_of_wood_3[AllFoci]) dt INIT decayed_3[AllFoci] = 0 INFLOWS: decay_of_wood_3[AllFoci] (IN SECTOR: Tree Growth and Survival) decayed_4[AllFoci](t) = decayed_4[AllFoci](t dt) + (decay_of_wood_4[AllFoci]) dt INIT decayed_4[AllFoci] = 0 INFLOWS: decay_of_wood_4[AllFoci] (IN SECTOR: Tree Growth and Survival) CWD_Growth[AllFoci] = (IF BG < 1 THEN 0 ELSE seedling_bank_on_nurselogs[AllFoci]/12)+(DCWDseedling[AllFoci]*0) OUTFLOW FROM: seedling_bank_on_nurselogs (IN SECTOR: Seed Germination and Seedling Survival) INFLOW TO: sapling (IN SECTOR: Tree Growth and Survival) Fc[AllFoci] = (2*PI)*(SQRT(Af[AllFoci]/PI) growth_modifier = 1 P = (Fa/Ga) total_attractive_trees = (ARRAYSUM(SC1[*])*.5)+ ARRAYSUM(SC2[*]) + (ARRAYSUM(SC3[*])*2) + (ARRAYSUM(SC4[*])*3) + (ARRAYSUM(dead_trees[*])*.5) + ARRAYSUM(dead_trees_2[*]) + (ARRAYSUM(dead_trees_3[*])*6) + (ARRAYSUM(dead_trees_4[*])*9) + (ARRAYSUM(dead_wood_in_foci[*])*.5) + ARRAYSUM(dead_wood_in_foci_2[*]) + (ARRAYSUM(dead_wood_in_foci_3[*])*2) + (ARRAYSUM(dead_wood_in_foci_4[*])*3) + (ARRAYSUM(decayed[*])*.5) + ARRAYSUM(decayed_2[*]) + (ARRAYSUM(decayed_3[*])*2) + (ARRAYSUM(decayed_4[*])*3) year = TIME/12

PAGE 96

LIST OF REFERENCES Alcantara, J. M., P. J. Rey, F. Valera, and A. M. Sanchez-Lafuente. 2000. Factors shaping the seedfall pattern of a bird dispersed plant. Ecology 81:1937-1950. Antos, J. A., R. Parish, and K. Conley. 2000. Age structure and growth of the tree-seedling bank in subalpine spruce-fir forests of south-central British Columbia. American Midland Naturalist 143:342-354. Armesto, J., J. C. Aravena, C. Villagran, C. Perez, and G. Parker. 1996. Bosques templados de la cordillera de la costa. In J. Armesto, C. Villagran, and M. Kalin, editors. Ecologa de los Bosques Nativos de Chile. Editorial Universitaria Santiago, Santiago. Armesto, J. J., I. Daz, C. Papic, and M. F. Willson. 2001a. Seed rain of fleshy and dry propagules in different habitats in the temperate rainforests of Chiloe Island, Chiloe. Austral Ecology 26:311-320. Armesto, J. J., and J. Figueroa. 1987. Stand structure and dynamics in the temperate rain forests of Chiloe Archipelago, Chile. Journal of Biogeography 14:367-376. Armesto, J. J., and R. Rozzi. 1989. Seed dispersal syndromes in the rain forest of Chiloe: evidence for the importance of biotic dispersal in a temperate rain forest. Journal of Biogeography 16:219-226. Armesto, J. J., R. Rozzi, C. Smith-Ramirez, and M. T. K. Arroyo. 1998. Conservation targets in South American temperate forests. Science 282:1271-1272. Armesto, J. J., C. Smith-Ramirez, and R. Rozzi. 2001b. Conservation strategies for biodiversity and indigenous people in Chilean forest ecosystems. Journal of the Royal Scoiety of New Zealand 31:865-877. Beckage, B., J. S. Clark, B. D. Clinton, and B. L. Haines. 2000. A long-term study of tree seedling recruitment in southern Appalachian forests: the effects of canopy gaps and shrub understories. Canadian Journal of Forest Research 30:1617-1631. Bell, S. S., M. S. Fonseca, and L. B. Motten. 1997. Linking restoration and landscape ecology. Restoration Ecology 5:318-323. Bewley, J. D., and M. Black 1982. Physiology and Biochemistry of Seeds in Relation to Germination Viability, Dormancy, and Environmental Control. SpringerVerlag, NewYork. 84

PAGE 97

85 Bibby, C. J., N. D. Burgess, D. A. Hill, and S. H. Mustoe 2000. Bird Census Techniques. Academic Press, London. Brewer, J. S. 2002. Disturbances increase seedling emergence of an invasive native shrub in pitcher-plant bogs. Natural Areas Journal 22:4-10. Brown, S., and A. E. Lugo. 1994. Rehabilitation of tropical lands: a key to sustaining development. Restoration Ecology 2:97-111. Bustamante, R. O., and J. J. Armesto. 1995. Regeneration dynamics in canopy gaps of a montane forest of Chiloe Island, Chile. Revista Chilena de Historia Natural 68:391-398. Cardoso da Silva, J. M., C. Uhl, and G. Murray. 1996. Plant succession, landscape management, and the ecology of frugivorous birds in abandoned amazonian pastures. Conservation Biology 10:491-503. Carmona, M. R., J. J. Armesto, J. C. Aravena, and C. A. Perez. 2002. Coarse woody debris biomass in successional and primary temperate forests in Chiloe Island, Chile. Forest Ecology and Management 164:265-275. Cespedes, C. L., A. Uchoa, J. R. Salazar, F. Perich, and F. Pardo. 2002. Plant growth inhibitory activity of p-hydroxyacetophenones and tremetones from Chilean endemic Baccharis species and some analogous: A comparative study. Journal of Agricultural and Food Chemistry 50:2283-2292. Charlesdominique, P. 1995. Plants-Frugivorous Animals Interactions Consequences on seed dispersal and forest regeneration. Revue D Ecologie-La Terre Et La Vie 50:223-235 Chanasyk, D.S., I.R. Whitson, E. Mapfumo, J.M. Burke, and E.E. Prepas. 2003. The impacts of forest harvest and wildfire on soils and hydrology in temperate forests: A baseline to develop hypotheses for the Boreal Plain. Journal of Environmental Engineering and Science 2:S51-S63. Christie, D. A., and J. Armesto. 2003. Regeneration microsites and tree species coexistence in temperate rain forests of Chilo Island, Chile. Journal of Ecology 91:776-784. Connell, J. H. 1971. On the role of natural enemies in preventing competitive exclusion in some marine mammals and in rain forest trees. Pages 298 312 in P. J. den Boer and G. Gradwell, editors. Dynamics of Populations. Centre for Agricultural Publishing and Documentation (PUDOC), Wageningen, The Netherlands Connell, J. H., and R. O. Slatyer. 1977. Mechanisms of succession in natural communities and their role in community stability and organization. The American Naturalist 111:1119-1144.

PAGE 98

86 Darnell, T. M. and K. E. Sieving. In Press. Landscape connectivity is not defined by corridors alone: A translocation experiment with an endemic understory bird in south-temperate rainforest. Conservation Biology. Debussche, M., and P. Isenmann. 1994. Bird-dispersed seed rain and seedling establishment in patchy Mediterranean vegetation. Oikos 69:414-426. Denslow, J. S., E. Newell, and A. Ellison. 1991. The effect of understory palms and cyclanths on the growth and survival of Inga seedlings. Biotropica 23:225-234. Daz, I., J. Armesto, S. Reid, K. E. Sieving, and M. Willson. 2005. Linking forest structure and composition: avian diversity in successional forests of Chiloe Island, Chile. Biological Conservation 123:91-101. Daz, I., C. Papic, and J. J. Armesto. 1999. An assessment of post-dispersal seed predation in temperate rain forest fragments in Chiloe Island, Chile. Oikos 87:228-238. Donoso, 1996. Ecology of Nothofagus Forests in Central Chile. Pages 271-292 In Thomas T. Veblen, Robert S. Hill, Jennifer Read editors. The Ecology and Biogeography of Nothofagus Forests New Haven: Yale University Press Donoso, P. J., and R. D. Nyland. 2005. Seedling density according to structure, dominance and understory cover in old-growth forest stands of the evergreen forest type in the coastal range of Chile. Revista Chilena de Historia Natural 78:51-63. Duncan, R. S., and C. A. Chapman. 2003. TreeShrub Interactions During Early SecondaryForest Succession in Uganda. Restoration Ecology 11:198. Ferguson, R. N., and D. R. Drake. 1999. Influence of vegetation structure on spatial patterns of seed deposition by birds. New Zealand Journal of Botany 37:671-677. Ferro, A., M. Gefell, R. Kjelgren, D. S. Lipson, N. Zollinger, and S. Jackson 2003. Maintaining hydraulic control using deep rooted tree systems. Advances in Biochemical Engineering/Biotechnology 78:125-156. Figueroa, J., J. J. Armesto, and J. F. Hernandez. 1996. Seed germination and dormancy strategies of temperate rain forest species in Chiloe, Chile. Revista Chilena De Historia Natural 69:243-251. Figueroa, J., and C. H. Lusk. 2001. Germination requirements and seedling shade tolerance are not correlated in a Chilean temperate rain forest. New Phytologist 152:483-489. Figueroa, J. A. 2003. Seed germination in temperate rain forest species of southern Chile: chilling and gap-dependency germination. Plant Ecology 166:227-240.

PAGE 99

87 Fjelds, J., and N. Krabbe 1990. Birds of the High Andes. Zoological Museum, Svendborg, Denmark. Gude, J. 2000. Avian community structure in two oldfield types on Isla Grande de Chiloe, Chile, and the potential for avian aid in forest restoration. Wildlife Ecology and Conservation. University of Florida, Gainesville. Gunnarsson, U., and H. Rydin. 1998. Demography and recruitment of Scots pine on raised bogs in eastern Sweden and relationships to microhabitat differentiation. Wetlands 18:133-141. Hallgren, W. S., and J. Pitman. 2000. The uncertainty in simulations by a Global Biome Model (BIOME3) to alternative parameter values. Global Change Biology 6:483-495. Halpern, B. S., S. D. Gaines, and R. R. Warner. 2005. Habitat size, recruitment, and longevity as factors limiting population size in stage-structured species. American Naturalist 165:82-94. Hanley, M. E. 1998. Seedling herbivory, community composition and plant life history traits. Perspectives in Plant Ecology, Evolution and Systematics 1:191-205. Harmon, M. E., J. F. Franklin, F. J. Swanson, P. Sollins, S. V. Gregory, J. D. Lattin, N. H. Anderson, S. P. Cline, N. G. Aumen, J. R. Sedell, G. W. Lienkaemper, K. Cromack, and K. W. Cummins. 1986. Ecology of coarse woody debris in temperate ecosystems. Advances in Ecological Research 15:133-302. Harvey, C. 2000. Windbreaks enhance seed dispersal into agricultural landscapes in Monteverde, Costa Rica. Ecological Applications 10:155-173. Hill, J. D., C. D. Canham, and D. M. Wood. 1995. Patterns and causes of resistance to tree invasion in rights-of-way. Ecological Applications 5:495. Hobbs, R. J., and D. R. Norton. 1996. Towards a conceptual framework for restoration ecology. Restoration Ecology 4:93-110. Hofgaard, A. 1993. Structure and regeneration patterns in a virgin Picea-Abies forest in northern Sweeden. Journal of Vegetation Science 4:601-608. Holl, K. D. 1998. Do bird perching structures elevate seed rain and seedling establishment in abandoned tropical pasture? Restoration Ecology 6:253-261. Holl, K. D. 1999. Factors limiting tropical rain forest regeneration in abandoned pasture: seed rain, seed germination, microclimate, and soil. Biotropica 31:229-242. Holl, K. D. 2002. Effect of shrubs on tree seedling establishment in an abandoned tropical pasture. Journal of Ecology 90:179-187.

PAGE 100

88 Holl, K. D., M. E. Loik, E. H. Lin, and I. A. Samuels. 2000. Tropical montane forest restoration in Costa Rica: overcoming barriers to dispersal and establishment. Restoration Ecology 8:339-349. Hostetler, M. 1999. Scale, birds, and human dimensions: a potential approach for integrative research in urban ecosystems. Landscape and Urban Planning 45(1),15-19. Houle, G. 1992. Spatial relationship between seed and seedling abundance and mortality in a deciduous forest of north-eastern North-America. Journal of Ecology 80:99-108. Houle, G. 1995. Seed dispersal and seedling recruitment the missing link(s). Ecoscience 2:238-244. Howe, H. F., and J. Smallwood. 1982. Ecology of seed dispersal. Annual Review of Ecology and Systematics 13:201-228. Howe, H. F., and M. N. Mirti. 2004. When seed dispersal matters. BioScience 54:651-660. Hubbell, S. P. 1979. Tree dispersion, abundance, and diversity in a tropical dry forest. Science 203:1299-1309. Hubbell, S. P. 1980. Seed predation and the coexistence of tree species in tropical forests. Oikos 35:214-229. Hulme, P. E. 2002. Seed-eaters: seed dispersal, destruction and demography. Pages 257-273 in D. J. Levey, W. R. Silva, and M. Galetti, editors. Seed Dispersal and Frugivory : Ecology, Evolution, and Conservation. CABI Publishing, New York. Jansen, A. 1997. Terrestrial invertebrate community structure as an indicator of the success of a tropical rainforest restoration project. Restoration Ecology 5:115-124. Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests. American Naturalist 104:501-528. Kirkman, L. K., K. L. Coffey, R. J. Mitchell, and E. B. Moser. 2004. Ground cover recovery patterns and life-history traits: implications for restoration obstacles and opportunities in a species-rich savanna. Journal of Ecology 92:409-421. Kirschbaum, M. U. F. 1999. CenW, a forest growth model with linked carbon, energy, nutrient and water cycles. Ecological Modelling 118:17-59. Kollman, J., and M. Pirl. 1995. Spatial pattern of seed rain of fleshy-fruited plants in a scrubland-grassland transition. Acta Aecologica 16:313-329.

PAGE 101

89 Laycock, W. A. 1991. Stable sates and thresholds of range condition on North American rangelands: A view point. Journal of Range Management 44:427-433. Lee, P. C., S. Crites, M. Nietfield, H. Van Nguyen, and J. B. Stelfox. 1997. Characteristics and origins of deadwood material in aspen-dominated boreal forests. Ecological Applications 7:691-701. Lewontin, R. C. 1969. The meaning of stability. Pages 13-24. Diversity and Stability in Ecological Systems. Brookhaven National Laboratory, Brookhaven, New York, USA. Li, X. D., and S. D. Wilson. 1998. Facilitation among woody plants establishing in an old field. Ecology 79:2694-2705. Lusk, C. H. 1995. Seed size, establishment sites, and species coexistance in a Chilean rain-forest. Journal of Vegetation Science 6:249-256. Lusk, C. H. 1999. Long-lived light-demanding emergents in southern temperate forests: the case of Weinmannia trichosperma (Cunoniaceae) in Chile. Plant Ecology 140:111-115. Lusk, C. H., and A. Del Pozo. 2002. Survival and growth of seedlings of 12 Chilean rainforest trees in two light environments: Gas exchange and biomass distribution correlates. Austral Ecology 27:173-182. Mallik, A. U. 2003. Conifer regeneration problems in boreal and temperate forests with ericaceous understory: Role of disturbance, seedbed limitation, and keytsone species change. Critical Reviews in Plant Sciences 22:341-366. McClanahan, T. R., and R. W. Wolfe. 1993. Accelerating forest succession in a fragmented lansdscape: the role of birds and perches. Conservation Biology 7:279-288. McDonnell, M. J., and E. W. Stiles. 1983. The structural complexity of old field vegetation and the recruitment of bird-dispersed plant species. Oecologia. 56:109-116. McGee, G. G., and J. P. Birmingham. 1997. Decaying logs as germination sites in northern hardwood forests. Northern Journal of Applied Forestry 14:178-182. Meyer, G. A., and M. C. Witmer. 1998. Influence of seed processing by frugivorous birds on germination success of three North American shrubs. The American Midland Naturalist 140:129-139. Milchunas, D. G., and W. K. Lauenroth. 1995. Inertia in plant community structure: state changes after cessation of nutrient-enrichment stress. Ecological Applications 5:452-458.

PAGE 102

90 Mladenoff, D. J., G. E. Host, J. Boeder, and T. R. Crow. 1996. LANDIS: a spatial model of forest landscape disturbance, succession, and management. Pages 175-180 in M.F. Goodchild, L. T. Steyaert, and B. O. Parks, editors. GIS and Environmental Modeling: Progress and Research Issues. GIS World Books, Fort Collins. Navarro, C., C. Donoso, V. Sandoval, and C. Gonzalez. 1997. Evaluacion de raleos en un renoval de canelo (Drimys winteri (Forst.)) en la Cordillera de la Costa de Valdivia, Chile. Bosque 18:51-65. Naveh, Z. 1994. From biodiversity to ecodiversity: a landscape-ecology approach to conservation and restoration. Restoration Ecology 2:180-189. Nepstad, D. C., C. Uhl, C. A. Pereira, and J. M. C. daSilva. 1996. A comparative study of tree establishment in abandoned pasture and mature forest of eastern Amazonia. Oikos 76:25-39. Ohlson, M., R. H. Okland, J. F. Nordbakken, and B. Dahlberg. 2001. Fatal interactions between Scots pine and Sphagnum mosses in bog ecosystems. Oikos 94:425-432. Papic, C., and J. J. Armesto. 2000. Decaying logs as recruitment sites for woody seedlings in early forest succession in Chiloe, Chile. Ecology and Evolutionary Biology. Universidad de Chile, Santiago. Parker, T. V. 1997. The scale of successional models and restoration objectives. Restoration Ecology 5:301-306. Peterson, C. J., and B. J. Haines. 2000. Early successional patterns and potential facilitation of woody plant colonization by rotting logs in premontane Costa Rican pastures. Restoration Ecology 8:361-369. Peterson, C. J., and S. T. A. Pickett. 1990. Microsite and elevational influences on early forest regeneration after catastrophic windthrow. Journal of Vegetation Science 1:657-662. Putz, F. E., and C. D. Canham. 1992. Mechanisms of arrested succession in shrublands root and shoot competition between shrubs and tree seedlings. Forest Ecology and Management 49:267-275. Reay, S. D., and D. A. Norton. 1999. Assessing the success of restoration plantings in a temperate new zealand forest. Restoration Ecology 7:298-308. Reid, S., I. A. Daz, J. J. Armesto, and M. F. Willson. 2004. Importance of native bamboo for understory birds in Chilean temperate forests. Auk 121:515-525. Rey, P. J., and J. M. Alcantara. 2000. Recruitment dynamics of a fleshy-fruited plant (Olea europea): connecting patterns of seed dispersal to seedling establishment. Journal of Ecology 88:622-633.

PAGE 103

91 Rozzi, R., J. J. Armesto, A. Correa, J. C. TorresMura, and M. Sallaberry. 1996. Avifauna of primary temperate forests of unihabited islands of Chiloe archipelago, Chile. Revista Chilena de Historia Natural 69:125-139. Rozzi, R., J. Silander, J. J. Armesto, P. Feinsinger, and F. Massardo. 2000. Three levels of integrating ecology with the conservation of South American temperate forests: the initiative of the Institute of Ecological Research Chiloe, Chile. Biodiversity and Conservation 9:1199-1217. Ruthsatz, B., and C. Villagran. 1991. Vegetation pattern and soil nutrients of a magellanic moorland on the cordillera-de-piuchue, Chiloe island, Chile. Revista Chilena De Historia Natural 64:461-478. Sarmiento, L., L. D. Llambi, A. Escalona, and N. Marquez. 2003. Vegetation patterns, regeneration rates and divergence in an old-field succession of the high tropical Andes. Plant Ecology 166:63-74. Simard, M. J., Y. Bergeron, and L. Sirois. 1998. Conifer seedling recruitment in a southeastern Canadian boreal forest: the importance of substrate. Journal of Vegetation Science 9:575-582. Slocum, M. G. 2000. Logs and fern patches as recruitment sites in a tropical pasture. Restoration Ecology 8:408-413. Slocum, M. G. 2001. How tree species differ as recruitment foci in a tropical pasture. Ecology 82:2547-2559. Slocum, M. G., and C. C. Horvitz. 2000. Seed arrival under different genera of trees in a neotropical pasture. Plant Ecology 149:51-62. Stattersfield, A. J., M. J. Crosby, A. J. Long, and D. C. Wedge. 1998. Endemic Bird Areas of the World: Priorities for Biodiversity Conservation. Birdlife Conservation Series. Birdlife International, Cambridge, UK. Stewart, G. H., A. B. Rose, and T. T. Veblen. 1991. Forest development in canopy gaps in old-growth beech (Nothofagus) forests, New-Zealand. Journal of Vegetation Science 2:679-690. Stokes, M. A., and T. L. Smiley. 1968. An Introduction to Tree-Ring Dating. University of Chicago Press, Chicago. Streng, D. R., J. S. Glitzenstein, and P. A. Harcombe. 1989. Woody seedling dynamics in an east Texas floodplain forest. Ecological Monographs 59(2):177-204. Stylinski, C. D., and E. B. Allen. 1999. Lack of native species recovery following severe exotic disturbance in southern Californian shrublands. Journal of Applied Ecology 36:544-554.

PAGE 104

92 Suding, K. N., and D. Goldberg. 2001. Do disturbances alter competitive hierarchies? Mechanisms of change following gap creation. Ecology 82:2133-2149. Sun, G., H. Riekerk, and L. V. Kornhak. 2000. Ground-water-table rise after forest harvesting on cypress-pine flatwoods in Florida. Wetlands 20:101-112. Szwagrzyk, J., J. Szewczyk, and J. Bodziarczyk. 2001. Dynamics of seedling banks in beech forest: results of a 10year study on germination, growth and survival. Forest Ecology and Management 141:237-250. Takahashi, M., Y. Sakai, R. Ootomo, and M. Shiozaki. 2000. Establishment of tree seedlings and water-soluble nutrients in coarse woody debris in an old-growth PiceaAbies forest in Hokkaido, northern Japan. Canadian Journal of Forest Research 30:1148-1155. Tappeiner, J. C., D. Huffman, D. Marshall, T. A. Spies, and J. D. Bailey. 1997. Density, ages, and growth rates in old-growth and young-growth forests in coastal Oregon. Canadian Journal of Forest Research 27:638-648. Toh, I., M. Gillespie, and D. Lamb. 1999. The role of isolated trees in facilitating tree seedling recruitment at a degraded sub-tropical rainforest site. Restoration Ecology 7:288-297. Traveset, A., T. Bermejo, and M. Willson. 2001. Effect of manure composition on seedling emergence and growth of two common shrub species of Southeast Alaska. Plant Ecology 155:29-34. Traveset, A., and M. F. Willson. 1997. Effect of birds and bears on seed germination of fleshy-fruited plants in temperate rainforests of southeeat Alaska. Oikos 80:89-95. Van Breeman, N. 1995. How Sphagnum bogs down other plants. Trends in Ecology and Evolution 10:270-275. Veblen, T. T. 1979. Structure and dynamics of Nothofagus forests near timberline in south-central Chile. Ecology 60:937-945. Veblen, T. T. 1985. Forest development in tree-fall gaps in the temperate rain forests of Chile. National Geographic Research 1:162-183. Villagran, C., and F. Hinojosa. 1997. Historia de los bosques de Sudamrica II. fitogeografa. Revista Chilena de Historia Natural 70:241-267. Wallis, J., S. Peterson, K. Chichakly, and B. Richmond. 2001. Stella Research. High Performance Systems. Westcott, D. A., and D. L. Graham. 2000. Patterns of movement and seed dispersal of a tropical frugivore. Oecologia 122:249-257.

PAGE 105

93 Wills, C., R. Condit, R. B. Foster, and S. P. Hubbell. 1997. Strong densityand diversity-related effects help to maintain tree species diversity in a neotropical forest. PNAS 94:1252-1257. Willson, M., and A. Traveset. 2000. The ecology of seed dispersal. Pages 85-110 in M. Fenner, editor. Seeds: The Ecology of Regeneration in Plant Communities. CAB International, Wallingford. Willson, M. F., and J. J. Armesto. 1996. The natural history of Chiloe, Chile: On Darwin's trail. Revista Chilena De Historia Natural 69:149-161. Willson, M. F., T. L. De Santo, C. Sabag, and J. J. Armesto. 1994. Avian communities of fragmented south-temperate rainforests in Chile. Conservation Biology 8. Wilson, A. D., and D. J. Shure. 1993. Plant competition and nutrient limitation during early succession in the southern Appalachian mountains. American Midland Naturalist 129:1-9. Wunderle, J. M. J. 1997. The role of animal seed dispersal in accelerating native forest regeneration on degraded tropical lands. Forest Ecology and Management 99:223-235. Young, P., S. Parkinson, and M. Lees. 1996. Simplicity out of complexity in environmental modelling: Occam's razor revisited. Journal of Applied Statistics 23:165. Zahawl, R. A., and C. K. Auspurger. 1999. Early plant succession in abandoned pastures in Ecuador. Biotropica 31:540-552. Zimmerman, J. K., J. B. Pascarella, and T. M. Aide. 2000. Barriers to forest regeneration in an abandoned pasture in Puerto Rico. Restoration Ecology 8:350-360.

PAGE 106

BIOGRAPHICAL SKETCH Michael was born amongst the suburban sprawl that is otherwise known as West Palm Beach, Florida. It was there that his father, a wetland biologist, helped to instill a great love of the natural world in his young son. After 18 years of life in a drained wetland, Michael moved to Gainesville, Florida, to attend the University of Florida and become a rabid fan of Gator football. Despite his interests in writing and philosophy, as well as his steadfast desire not to follow in his fathers footsteps, Michael graduated in 1999 with a B.S. in wildlife ecology and conservation. A love of Gainesville, a great advisor, a noble research goal, and a chance to do research in Southern Chile all conspired to keep Michael in Gainesville to write this thesis. Now that this thesis is finally done, Michael hopes to pursue a career integrating ecological research with information technology, spend a lot of time thinking about as much as possible, have lots of fun, and continue learning amazing things. 94


Permanent Link: http://ufdc.ufl.edu/UFE0012043/00001

Material Information

Title: The Role of Birds and Microsites in the Regeneration of South-Temperate Rainforest
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0012043:00001

Permanent Link: http://ufdc.ufl.edu/UFE0012043/00001

Material Information

Title: The Role of Birds and Microsites in the Regeneration of South-Temperate Rainforest
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0012043:00001


This item has the following downloads:


Full Text












THE ROLE OF BIRDS AND MICROSITES IN THE REGENERATION OF SOUTH-
TEMPERATE RAINFOREST















By

MICHAEL P. MILLESON


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Michael P. Milleson















ACKNOWLEDGMENTS

I thank my parents first and foremost for helping to get where I am today. I thank

the Fundacion Senda Darwin, Biological Station, Senda Darwin. I also thank Traci M.

Darnell for all of her help and support, and Mary F. Willson and Juan J. Armesto for their

guidance. Finally I thank my awesome advisor, my wonderful girlfriend, my wacky dog,

and each and every one of my friends.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ......... .................................................................................... iii

LIST OF TABLES ........................ ....... ......... ............. vii

LIST OF FIGURES ............ .......... ............................... viii

ABSTRACT ........ .............. ............. ...... ...................... xi

CHAPTER

1 AVIAN SEED DISPERSER ACTIVITY AND AVAILABILITY OF
GERMINATION SUBSTRATE IN BACCHARIS-DOMINATED OLD-FIELDS
IN SO U T H E R N C H IL E .................................................................. ..................... 1

Introduction ...........................................................................................
Chilean South-tem operate Rainforest................................... ....................... 2
Disturbance, Arrested Succession, and Consequences................. ............ 3
Alternative Hypotheses for Arrested Succession: Seed Dispersal vs.
Germ nation Lim itation .................................. .....................................4
Seed dispersal limitation ......................... .....................5
Seed germination and seedling establishment limitation ............................8
R research D esign ................................................................. 9
Frugivore activity hypothesis ................................................... ....... ........ .11
G erm nation site hypothesis ...................................... ......................... .......... .11
M e th o d s ..............................................................................12
Study Site............................................. 12
S tu d y S p e c ie s ................................................................................................. 12
Frugivore A activity H ypothesis ........................................ ........................ 13
B ird cen su ses ................................................................... ............... 13
F ocal sam ples .................................................................. ............... 14
S e e d tra p s ................................................................... 1 5
G erm nation Site H ypothesis........... ................. ........................ ............... 15
Seedling transects ....... ........ .. ......... ..... .. ....... ......... ............ .... 15
Substrate availability transects ............................... ............... 16
Germination and survival experiment ................. ................................ 16
Results ................ ................ .......................17
Frugivore A activity H ypothesis ........................................ ........................ 17
B ird C ensuses .............. .... ......... ... ......... .......................................... 17









F ocal sam ples .................................................................. ............... 18
S e e d tra p s ............................................................................................... 1 9
G erm nation Site H ypothesis......................................... .......................... 20
Seedling transects ..................................................... .. ........ .... 20
Substrate availability transects .............. ............................................. 20
G erm ination............................................. 21
D isc u ssio n ............................................................................................................. 2 1
F ru g iv o re A ctiv ity ......................................................................................... 2 2
S eedling E stab lishm ent ................................................................................. 24
R e cru itm en t F o ci ........................................................................................... 2 6
C o n c lu sio n s ................................................................................................... 2 7

2 RAINFOREST RESTORATION SCENARIOS FOR BACCHARIS-
DOMINATED OLD-FIELDS IN SOUTHERN CHILE: A SIMPLE
ECOSYSTEM MODEL AS A DECISION MAKING TOOL...............................28

Intro du action ................. ..... ............ ........................ ................................ 2 8
Chilean South-temperate Rainforest: Natural Disturbance Regime and
Arrested Succession ................................. .......................... ... ......28
M modeling Restoration Scenarios ................................ .. ................ 31
M odel D description ...............................................................33
O v erv iew ....................................................... 3 3
O ld-field Characteristics....... .............. ..................... ............. ............. 35
S eed rain .................................................................. ............. 3 7
Seed Germination and Seedling Survival .....................................................40
Creation of N ew Foci ................................................................ ............. 41
Tree Growth and Survival ................. .................................43
Sensitivity A nalyses ................................................ ............... 45
M anagem ent Scenarios...................................... ......... 47
R results .................................48................................................
Initial Conditions ................................. ........................... ... .......48
Coarse woody debris ................. ...............................48
B a cch a ris ............................................................................... 4 9
Sphagnum and grass ...................... ............................49
T ree cover........................... ...................... ...... 50
Length of edge ......... ...................................... 50
Bird density ................................ ............................ 53
Sensitivity A naly ses ....................................... .................... ..................53
Germination and tree growth rate................. ................. ............ 53
Survival rate of seedlings on dead wood .............................................54
Rate of cover loss to forest......................................................... ............... 55
Rarity of new recruitment-foci formation ................................................... 56
M management scenarios .............................. ........ .......................57
D iscu ssio n ............... ...... .. ............ .......... ....... ....... ....................... 5 7
Relative Importance of Seed Dispersal and Germination Limitation .................57
Restoration Methods / Management Scenarios ............................................. 61
Assessment of model scenarios ....................... ........... .............. 61


v









Recommendations for efficient reclamation of Baccharis fields ...............62
A ssessm ent of the M odel......................................................................... ...... 68
Com plexity .................................... .......................... .... ........68
R elev ant scales ............................................... ................ 6 8
M odel im provem ent ......................................................... ............. 69

APPENDIX

A ICONOGRAPHIC REPRESENTATION OF MODEL ..........................................71

B M O D E L E Q U A T IO N S ................................................................... .....................75

L IST O F R E FE R E N C E S ........... ....................................................... ...........................84

B IO G R A PH IC A L SK E TCH ..................................................................... ..................94















LIST OF TABLES

Table p

1 Comparison of the number of seedlings found beneath non-tree cover types to
the number found beneath trees to estimate the odds that a seedling would be
established there. ......................................................................44

2 Parameters used for sensitivity analyses. Values represent mean of 6 fields..........46

3 Field conditions, input, and time required for 50% regeneration under 3 different
management scenarios and three different starting conditions. ............................60















LIST OF FIGURES


Figure p

1 Conceptual model of options for use of an arrested successional site ...................7

2 Average number of frugivores counted during 10 minute point counts at 12 sites
in C hiloe, C while. .................................................... ................. 18

3 Mean number of frugivore visits during 30-minute focal samples to clusters of
trees and adjacent single trees in degraded old-fields in Chiloe, Chile (N = 24).....19

4 Mean number of frugivore visits during 30-minute focal samples to fruiting and
non-fruiting trees in degraded old-fields in Chiloe, Chile (N = 14).........................20

5 Density of seedlings found growing on each substrate type in 9 degraded old-
fields in Chiloe, Chile.................... ..................... .. ......21

6 Percent cover of various substrate types across six degraded old-fields in Chiloe,
Chile. Error bars represent +/-1 SE. ............................................. ............... 22

7 Number of seeds germinating on each substrate type in 18 trials, in degraded
old-fields in C hiloe, C while. ........................................ ........................................23

8 Seedling density divided by substrate availability for various substrates in
Baccharis-dominated oldfields throughout Chiloe, Chile................... ............25

9 A simple representation of old-field regeneration as conceptualized by the
m odel. ................................................................................34

10 The value of S, or seed rain per meter of forest in the old-field, changes with
increasing tree cover in the old-field......................... ...... ...............43

11 Correlation between age of Drymis winterii (as determined by coring) and
diam eter at breast height (DBH). ........................................ ......................... 45

12 Correlation between the age of Drymis winterii (as determined by coring) and
tree h eig ht. ......................................................... ................ 4 7

13 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial amount of coarse
woody debris (CW D) is increased. ........................................ ....... ............... 49









14 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial amount of coarse
woody debris (CWD) is increased with zero initial trees in the old-field ..............50

15 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial coverage of
Baccharis is increased.................. ............................... ............ 51

16 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial coverage of
Sphagnum is increased. ...................... .. .................... ....................... .....51

17 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial coverage of
grasses and ferns is increased ....................................... ............... ............... 52

18 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial number of trees is
increased............... .. ....... ........................ ............ ......... 52

19 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the length of adjacent edge is
increased............................................................................................. .53

20 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial number of
frugivorous birds is increased. ............................................................................ 54

21 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the germination rate is
in cre a se d .................................................. ................. ......................5 5

22 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of tree growth is
increased or decreased .................. ............................ .... .. .. .. ........ .... 56

23 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the mortality of seedlings on
dead w ood is increased ......................................... .............. .. .. ........ .... 57

24 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of seedling mortality
increases. ............................................................................58

25 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of out-competition of
B accharis is increased ........................ .. ........................ .... ........ ................59









26 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of out-competition of
Sphagnum is increased. ................................................ ................................ 59

27 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of out-competition of
grass and fern is increased ............. ...................... ................... ............... 61

28 Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the value of "r" (additional
rarity of new recruitment foci establishment) is increased.................. ............63

29 Time predicted for regeneration of forest (to 5, 25, and 50% forest cover) with
low densities of seed dispersers, compared with time required for regeneration
with very few micro-sites suitable for germination, compared to default values
of each (see text). ................................................... ................. 65

30 Model prediction of the effect of the number of saplings planted annually on the
rate of forest regeneration (to 25, 50, 75, 100% forest cover). .............................67















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

THE ROLE OF BIRDS AND MICROSITES IN THE REGENERATION OF SOUTH-
TEMPERATE RAINFOREST

By

Michael P. Milleson

December, 2005

Chair: Kathryn E. Sieving
Major Department: Wildlife Ecology and Conservation

Determining the mechanisms of arrested succession at restoration sites can

influence understanding and management of landscape scale patterns and processes. On

Isla Grande de Chiloe, a large continental island in southern Chile, conversion of south-

temperate rainforest to pasture is occurring at a high rate. In some cleared sites

agricultural activity cannot be implemented due to invasion of persistent shrub fields

comprised ofBaccharis spp., a scrubby bush in the Asteraceae. Baccharis-dominated

fields serve no economic purpose and are of limited use to wildlife. If they could be

restored to native forest then landowners and endangered endemic wildlife species would

accrue greater benefit. The goal of this study was to identify the relative importance of

two likely limitations on natural regeneration of native forest in Baccharis-invaded sites

on Isla Grande de Chiloe, Chile.

I tested two alternative hypotheses for arrested succession in my system: that shrub

fields persist because of lack of seed dispersal, and lack of appropriate substrates for seed









germination and seedling establishment. In chapter 1 I present findings of field studies

(experiments and comparative observations) showing that both ample seed dispersal and

provision of suitable germination sites must occur to increase the rate of forest

regeneration in Baccharis-dominated shrub fields. Sites with few or scattered trees

received significantly less avian seed-disperser visitation than did sites with more than

approximately 30 trees/ha or with trees in clumps of several individuals. Additionally,

substrate types with the greatest seedling density were the rarest of the substrates

available.

In the second chapter, I develop a dynamic systems model that incorporates

processes actuating both hypothesized mechanisms of arrested succession to address

realistic scenarios for restoration given constraints and goals relevant to local landowners.

I use field data and findings reported in the literature to parameterize the model. Based

on the model, the following suggestions are made. First, areas selected for forest

regeneration should be adjacent to mid-successional or old growth forest stands in order

to ensure sufficient avian seed dispersal. Second, the focus should be on providing

sufficient germination substrates and making the field attractive to avian dispersers. The

model also showed that given both avian seed dispersers and suitable germination sites,

the most important deterrents to forest succession are the competitive effects of

Sphagnum and Baccharis cover in the fields. Finally, the model demonstrates how these

deterrents to regeneration can be overcome simply by planting a few trees each year in

old fields that have minimal germination sites a realistic recommendation given

landowner constraints.














CHAPTER 1
AVIAN SEED DISPERSER ACTIVITY AND AVAILABILITY OF GERMINATION
SUBSTRATE IN BACCHARIS-DOMINATED OLD-FIELDS IN SOUTHERN CHILE

Introduction

A common goal of ecological restoration is to quicken the natural pace of

secondary succession (Hobbs & Norton 1996). However, in cases of arrested succession,

when a damaged or degraded ecosystem does not return to the original state (Brown &

Lugo 1994), the causes of the arrest must first be identified and removed, if possible,

before secondary succession can proceed (Parker 1997). Three general types of factors

can cause arrested succession. Sites can be colonized by species that inhibit the growth

or spread of species more characteristic of the desired ecological state (Connell & Slatyer

1977); abiotic factors pushed outside the local species' ranges of tolerance by the

disturbance can prevent establishment by representative colonizers (Milchunas &

Lauenroth 1995); or, finally, disturbance can bring about conditions (biotic or abiotic)

that favor the influx of an entirely different set of species to the site (Suding & Goldberg

2001).

Determining the mechanisms of arrested succession at small scales can influence

understanding and management of larger landscape scale patterns and processes (Bell et

al. 1997). Once succession becomes arrested, site characteristics may change

considerably, pushing the site over a threshold toward the "basin of attraction" of an

alternate state (Laycock 1991; Lewontin 1969). When undesirable alternate stable states

arise in landscapes, restoration ecological approaches can be used to identify and release









the mechanisms that generate and maintain them. The goal of this research was to

identify mechanisms that may be inhibiting forest succession in previously cleared fields

in the Valdivian temperate rainforest region of southern Chile. The persistent shrub

fields that may develop following clearing of Chilean temperate rainforest can occupy

significant areas (30% or more in regions where forest clearing is advanced) and provide

little to no ecological or economic productivity (Gude, 2000; personal observation).

Therefore, these persistent shrub fields are manifest as an undesirable alternate ecosystem

state at the landscape scale. In this study I examine alternative causes of arrested forest

succession at a community-scale, in shrub fields at forest edges, with the goal of

identifying factors maintaining persistent shrub lands that, via restoration work, could be

released or altered to allow forest succession to proceed.

Chilean South-temperate Rainforest

The South temperate rainforest of the Valdivian region of Chile (35- 480 S)

receives between 1,000 and 6,000 mm of rain per year, and is characterized by emergent

evergreen broad-leaved trees (e.g., Nothofagus oblique, N. alpine, andN. dombeyi) and

conifers (e.g., Podocarpus nubigina). Typical canopy and understory species include

Drimys winterii, Weinmannia trichosperma, and several trees in the family Myrtaceae

(Willson et al. 1994). Endemism is very high (Stattersfield et al. 1998), ranging from

45% in vertebrates to 90% in seed plants (Armesto et al. 1996; Villagran & Hinojosa

1997). The typical disturbance regime is characterized by periodic catastrophic

disturbances such as earthquakes, volcanic activity, and fire. Windthrow and treefall gap

creation are common occurrences (Veblen 1979). This disturbance regime prevents

shade tolerant tree species such as Laurelia phillipiana and Saxegothea conspicua from

out-competing the shade intolerant Nothofagus (Bustamante & Armesto 1995). Due to









human settlement and agricultural land uses, much of the remaining Valdivian rainforest

exists as fragmented patches in a matrix of pastoral, agricultural, and industrial forestry

land uses, where increased gap creation, fire, and windthrow frequencies along forest

edge have intensified disturbances in remaining forests (Willson & Armesto 1996).

Forest clearing and associated human activities in remaining forest have resulted in

global endangerment of endemic flora (Armesto et al. 1998) and fauna of the region

(Stattersfield et al. 1998).

Disturbance, Arrested Succession, and Consequences

On Isla Grande de Chiloe, a large continental island serviced by a system of ferries,

forest conversion has been slower than on adjacent mainland (Rozzi et al. 2000) due to its

greater economic isolation. Here, the rural life style is characterized by pasture creation

(via tree cutting followed by fire) for milk and meat cows and sheep, by non-mechanized

row crop production (oxen teams are often used to till the soil), and by fuel wood

acquisition in the most accessible forest patches (Armesto et al. 1998). Large scale

forestry (via clear-cutting) also occurs in the island's southern sectors, and plantations of

pine and eucalyptus are increasing throughout Chiloe's rural landscapes (Armesto et al.

2001b). In rural communities, economic productivity for the people is partly determined

by availability of pasture for livestock and wood fiber for cooking and building materials.

This can be limited by the development of persistent shrub fields (dominated by

Baccharis magellanica) following forest clearing that cannot be used for livestock or row

crop production. Moreover, native forest succession does not occur readily in Baccharis-

dominated sites.

While native forest frequently reinvades logged sites that are not further disturbed

by fire, and in some agricultural old fields left fallow, Baccharis fields frequently









develop after forest clearing. It appears that Baccharis takes over especially where the

water table may be higher than elsewhere, and significant soil inundation prevents

establishment of both native trees (Bewley & Black 1982) and cultivars. The shrubs can

be burned back but, without prohibitively expensive ditching and draining, fire alone

does not often improve site utility for either agriculture or forest regeneration. Baccharis

overstory may prevent establishment by forest tree species (Cespedes et al. 2002; Putz &

Canham 1992), and is underused by the local avifauna (Gude, 2000). Therefore this

shrub land formation is of low economic and ecological value, and it appears to be highly

persistent. In the landscape of NE Chiloe, Baccharis fields comprise around 30% of the

land cover and are virtually unused by native wildlife species (T. M. Darnell, K. E.

Sieving, unpublished data). Since farmers that clear forest for pasture or wood products

and get Baccharis development in the cleared area usually move to clear a different site,

if available, I view these fields as a restoration priority. Regeneration of forest on

arrested successional sites would provide wildlife habitat and at least minimal economic

benefit (forest products) for people, and this might protect forest in other sites from

additional clearing (Fig. 1). In this study I focused on understanding factors limiting

natural forest regeneration in sites dominated by Baccharis shrubs.

Alternative Hypotheses for Arrested Succession: Seed Dispersal vs. Germination
Limitation

Forest regeneration can be limited by several factors, including competition, lack of

nutrients, irregular disturbance regime, or allelopathy (Brewer 2002; Connell & Slatyer

1977; Kirkman et al. 2004; Mallik 2003; Wilson & Shure 1993; Cespedes et al. 2002).

While all of these factors likely play a role in creating and maintaining a state of arrested









succession, this study focuses on dispersal and germination due to their importance in this

system (Armesto & Rozzi 1989; Papic & Armesto 2000).

Seed dispersal limitation

Insufficient seed dispersal can limit opportunities for establishment and growth of

diverse plant species and, thereby, reduce vegetative structural heterogeneity. Given that

more than 70% of all trees, shrubs, and vines in the Valdivian rainforests are bird

dispersed (Armesto & Rozzi 1989), access to Baccharis fields by frugivorous birds and

their activities in them are likely to define many parameters of regeneration. In a study

by Armesto et al. (2001a), only 10% of the fleshy fruits collected in seed traps placed in

rainforest fragments reached the margins of the forest patch, suggesting that even fewer

would reach beyond forest edges and into shrub fields. Moreover, since the principal

seed dispersing birds are forest species (Willson et al. 1994), the absence of suitable

habitat for them in cleared fields could contribute to arrested succession. Seed dispersal

is important in this site if seeds are able to establish in open areas or if they are deposited

on suitable microsites for germination (Howe & Mirti 2004).

Three possible factors have been identified that might make a site, such as an

anthropogenic shrub field, unsuitable for use by frugivorous birds: a lack of perches, a

lack of structural diversity, and a lack of food. Several studies have found that seed rain

is positively correlated with vegetation that offers natural perching sites (Debussche &

Isenmann 1994; Ferguson & Drake 1999; Harvey 2000; Kollman & Pirl 1995) and with

the availability of manmade perches (but see Holl 1999; McClanahan & Wolfe 1993;

McDonnell & Stiles 1983). In this study, I assessed the importance of natural (tree)

perch availability in Baccharis fields on avian frugivore activity.









A lack of vegetative structural complexity may also result in decreased visitation to

a site by seed dispersers. Holl (1998) and McDonnell and Stiles (1983) found that

perches that are more complex received more seed rain. Cardoso da Silva, et al. (1996)

also found a positive correlation between structural complexity and bird use of a site.

Structural complexity, and its positive effect on bird use, is also increased when trees are

found in close proximity to one another (Toh et al. 1999). Complex vegetative structure

may appeal to birds for reasons such as increased cover and more diverse microhabitats.

In sum, enhancing complexity in target degraded sites may be an important factor in

restoration where bird dispersal is a central constraint on inputs of seeds. To address this

aspect of seed dispersal limitation into shrub fields, I assessed the effect of simple natural

perches (lone trees) versus more complex perching and cover for frugivorous birds,

represented by clumps of trees.

A third factor limiting seed dispersal may be lack of food; fruiting vegetation in a

restoration site can attract birds to make more visits during which they are more likely to

defecate seeds. Although Holl (1998) found that using fruit as bait did not increase bird

visitation, Cardoso da Silva, et al. (1996) found an increase in the use of an abandoned

field when naturally occurring fruit resources were higher. Moreover, Slocum and

Horvitz (2000) found greater seed dispersal beneath fleshy fruit producing trees in Costa

Rica. Wunderle (1997) also suggests that, in general, the presence of fruit plays an

important role in attracting seed dispersers. Thus, a potential consideration in restoration

efforts is the availability of fruit to frugivores in the target sites. In this study, I addressed

this possibility by assessing the relative influence of fruiting and non-fruiting trees in

shrub fields on frugivore visitation and activity.

































Purview of this study


Figure 1. Conceptual model of options for use of an arrested successional site. The left
path helps alleviate continued fragmentation pressure on temperate rainforests
in S. Chile, but is expensive. The right path enhances regeneration of forest
and wildlife habitat in areas that would otherwise have no wildlife or
economic values. Both have the same end result from a landscape level
perspective; however, fully identifying how to enhance regeneration is the
purview of this study.









Seed germination and seedling establishment limitation

Once successfully dispersed to a site, tree establishment and growth can be limited

at various life stages. Germination of dispersed seeds can be prevented by physical

characteristics of the microsite insolationn, moisture, litter depth, pH, etc.; Houle 1992;

Peterson & Pickett 1990; Streng et al. 1989), and by seed predation (Hulme 2002). Other

critical life-stages include the seedling and sapling stages, at which point plants are

especially vulnerable to herbivory (Hanley 1998) and fungal attack (Rey & Alcantara

2000). In Baccharis fields, inundation, desiccation, and nutrient limitation may all affect

seed germination. Year round surface water inhibits germination of most tree species

(Bewley & Black 1982), and is likely to do so under the hydric conditions that

characterize Baccharis fields (Papic & Armesto 2000). Moreover, as Baccharis becomes

established in cleared sites development of a surface layer of Sphagnum is commonly

observed (Ruthsatz & Villagran 1991) and could reduce germination of seeds in at least

two ways. When Sphagnum accumulates in a disturbed site, it creates acidic and nutrient

poor conditions that are known to exclude colonization by forest species (Van Breeman

1995, J. Armesto, K. Clark, pers. comm.). Seeds may also desiccate readily when

deposited on top of Sphagnum mats, because while this lichen requires standing water to

grow, the top layers can be well above available moisture (Van Breeman 1995).

It has been suggested that all of these extreme conditions limiting seed germination

in hydric forests and old-fields can be alleviated by the availability of dead wood on the

ground (Papic & Armesto 2000; Takahashi et al. 2000). Woody debris in Baccharis

fields fluctuates less in water content relative to soils (Papic & Armesto 2000), and

therefore, could ameliorate both low and high water stresses on seeds falling in bare

ground or Sphagnum-laden sites. Moreover, large pieces of woody debris from tree









trunks (commonly called nurse logs) can collect organic debris, preventing nutrient

limitations (Harmon et al. 1986). After germinating on dead wood, seedlings

subsequently become established by growing down into the soil. With sufficient maturity

attained while supported by nurse logs, tree seedlings/saplings of wet forest can then

tolerate, and even alter conditions of the soil. In other wet temperate forest systems,

conifer germination is largely dependent on the availability of nurse logs (Hofgaard 1993;

Simard et al. 1998), and in northern hardwood forests, yellow birch and red spruce

densities are 24 and 5 times greater, respectively, on nurse logs than on the forest floor

(McGee & Birmingham 1997). In premontane Costa Rican pastures, four of the most

common woody species were found significantly more often on logs than on surrounding

microsites (Peterson & Haines 2000). In a temperate Chilean forest system, seedlings of

eight tree species common to the forests of Chiloe, were found growing primarily on

dead wood on the forest floor (Christie & Armesto 2003; Lusk 1995). Additionally,

Papic and Armesto (2000) showed that survivorship of one-year-old seedlings of the five

most dominant tree species found in the region is higher on woody debris in logged

fields. Thus, it is likely that post-clearing fire applied by land owners in my study system

reduces the availability of deadwood and that this is a limitation on germination success

of seeds arriving into these fields.

Research Design

I considered two potentially interacting hypotheses to better understand the

processes inhibiting succession in south temperate old-fields and the potential for

manipulating them during restoration to forest. The first hypothesis, that a lack of avian

frugivore activity is limiting forest regeneration, is based on the possibility that a lack of

perches, structural diversity providing cover, and/or a lack of food is making the









Baccharis fields unsuitable for use by frugivorous birds. Rather than use artificial

perches, I conducted three comparative observational studies to examine the influence of

naturally occurring variation in tree density and fruiting activity on bird activity at the

scale of 0.5 ha old-fields. To test the second hypothesis, that the limited availability of

suitable germination microsites is limiting forest regeneration, I conducted one

comparative-observational study examining the relationship between the presence of

seedlings on various substrates and the availability of these sites, and one experimental

study examining germination rates on different substrates.

This study was conducted along forest/old-field boundaries at sites occurring over a

400-km2 area. At the scale of my study plots, the phenomena of interest are localized

bird movements, microhabitat choices (e.g., perches and feeding sites), and small-scale

changes in substrate availability, which influence individual tree growth and distribution.

At larger scales, other forces such as economics, large-scale disturbances, and patch

context are operating to determine overall extent of forest versus other land uses in the

regional landscape. But the purpose of this study was to examine small-scale processes

potentially under the control of individual landowners seeking recommendations for

managing their small parcels comprising the total area (sensu Hostetler 1999). With-in

the context of a given field, the study was conducted close to the forest edge because

successional processes fostering forest intrusion into hydric fields (e.g., frugivore

activity, seed fall, and deadwood accumulation) occur from the forest edge outward

(Armesto et al. 2001a; Armesto & Rozzi 1989). Data generated here were used to

identify limitations on seedling establishment and to parameterize a systems model for

comparison of different forest restoration scenarios (Chapter 2).









Frugivore activity hypothesis

Assuming that increasing frugivore activity correlates strongly with seed movement

(Westcott & Graham 2000), I studied frugivore visitation rates to old-fields with varying

numbers of trees (Debussche & Isenmann 1994; Ferguson & Drake 1999; Harvey 2000;

Kollman & Pirl 1995), clusters versus single trees (Cardoso da Silva et al. 1996; Holl

1998; McDonnell & Stiles 1983; Toh et al. 1999), and fruiting versus non-fruiting trees

(Cardoso da Silva et al. 1996).

I predicted that bird activity would be influenced by type (fruiting vs. non-fruiting),

occurrence, and distribution (solitary vs. clumped) of trees in fields and proximity to

forest. Specifically, I predicted that avian frugivore abundance would be greatest in old-

fields with greatest numbers of remnant trees, that avian frugivores would be most often

associated with clusters of trees rather than single trees, that avian frugivores would be

more often associated with fruiting trees rather than non-fruiting trees, and that avian

seed deposition would be greatest beneath trees.

Germination site hypothesis

To test whether germination sites for forest trees are limiting in old fields, I

surveyed fields for tree seedlings and identified 6 relevant micro-site types; beneath trees,

coarse woody debris (CWD), CWD beneath a tree, beneath Baccharis, Sphagnum, and

bare ground. I then surveyed fields for the availability of micro-sites that were relevant

to seedling establishment, and tested to see if sites promoting seedling establishment

were limited. I predicted that I would most often find seedlings growing on dead wood

beneath trees, and least often on Sphagnum moss. I also tested actual establishment rates

by planting seeds on three different substrates (decaying wood, bare dirt, and Sphagnum)

and comparing their germination and survival success. I expected to see greater rates of









establishment on dead wood when compared to Sphagnum moss and bare ground (Papic

& Armesto 2000). The major assumption is that sites that are suitable for germination are

also suitable for further survival, which is not always true (Gunnarsson & Rydin 1998).

This assumption was not tested, however Papic and Armesto (2000) found that seedling

survival is higher on coarse woody debris than on bare ground.

Methods

Study Site

The study was conducted during the months of January and February 2001-2002, at

and near Estaci6n Biol6gica Senda Darwin, a field station located on Isla Grande de

Chiloe (9,600 km2 ) roughly 10 km from the coast of Chile (4155'S, 73035'W). The

main woody tree species colonizing disturbed habitat are the avian dispersed D. winterii,

and E. cordifolia, and the wind dispersed N. nitida (Veblen 1985). The study fields are

located in the northeastern part of the island, near the towns of Manao and Linao. Mean

annual rainfall is 1906 mm (peaks in Austral winter; June-September) and the mean

annual temperature is 11 C (Armesto & Figueroa 1987).

Study Species

The primary seed dispersers in this system are Elaenia albiceps (white crested

elaenia or fio fio) and Turdusfalklandi (austral thrush or zorzal; Willson et al. 1994).

The white crested elaenia occupies forest interior, edges, and clearings in Nothofagus

forest. Its breeding season is from November to February (Fjeldsi & Krabbe 1990). The

austral thrush makes use of a variety of habitats ranging from Nothofagus understory to

gardens, parks, or brushy country. Its breeding season begins in October and ends in

February (Fjeldsi & Krabbe 1990). Nothofagus nitida was the numerically dominant tree









species in the post-disturbance shrub fields, followed by Drimys winterii, Amomyrtus

meli, Eucryphia cordifolia, and Podocarpus nubigena.

Frugivore Activity Hypothesis

Bird censuses

I conducted bird censuses during the breeding season (January and February 2001-

2002) between the hours of 06:30 and 09:30 to examine the effect of trees in fields on the

number of avian visits to fields. A total of 12 0.5 ha rectangular old field sites were

selected, including 4 sites in each of the following categories based on the density of

emergent forest trees (> 10cm dbh); low (with zero to 18 trees/ha), medium (30- 48

trees/ha), and high (more than 58 trees/ha). Site selection was constrained by proximity

to the field station and thus was non-random. However, sites were at least 200m apart,

and in most cases greater than 1000m apart, limiting the chances of non-independence.

For each sample, I delimited a 100m by 50m section of old-field adjacent to a forest edge

that contained Baccharis magellanica. Nine sites were sampled 3 times each in 2001,

and an additional three sites were added in 2002. In order to avoid confusing temporal

effects on frugivore abundance with site effects, all sites were censused once before any

site was censused a second time (with two exceptions, due to travel restrictions).

Censuses were only conducted on non-rainy mornings. For each sample, I recorded each

frugivore seen moving from forest into the 0.5 ha section during a 10 minute period.

Ten-minute point counts probably allowed double counting to occur. However, I was not

concerned with movement from forest to field per individual bird, but rather total number

of field visits per unit time. Whether by one or by several birds, each visit has an equal

probability of resulting in a defecated seed. Since all of my sites were located near forest

patches large enough to support many individuals of my study species, I assumed that









linear densities of these bird's territories along forest edges were comparable among

sites. Thus, my census plots sampled visits by an equal number of individuals. The

effect of field type (high, medium, or low number of trees) on number of avian frugivore

visits was subjected to a Kruskal-Wallace one way analysis of variance.

Focal samples

I used a sub-set of the sites described above to conduct 30-censuses of two specific

trees or tree clusters between the hours of 07:00 and 10:00. Sites were selected based on

the availability of trees that fit the following design. To compare clusters of trees to

single trees, I selected a cluster of trees, usually mixed species, and a single tree,

equidistant from the forest edge and within 20m of one another. I defined a cluster as a

group of two or more trees where each tree was within 0.5m of the foliage edge of its

nearest neighbor. Mean cluster size was 10.18m (+/- 1 S.E. = 0.8430m) circumference at

the outer edge of the crown. The mean crown circumference of a single tree was 4.00m

(+/- 0.4624m). After selecting the trees, I placed myself in an inconspicuous location that

provided an unobstructed view of both trees and clusters and counted the number of avian

frugivore visits to either the cluster of trees or the single tree during a 30-minute period.

24 single-cluster pairs were censused. I also conducted focal samples comparing fruiting

tree species (Drimys winterii) to non-fruiting tree species, using the same methods

described above. Drimys winterii were not presently bearing fruit in four of the 14 pairs

sampled. The effect of cluster type (cluster or single) and tree type (fruiting or non-

fruiting) on the number of avian frugivore visits was analyzed using Mann-Whitney U

tests.









Seed traps

To test for a difference in seed rain with distance from the forest edge, I placed 120

seed traps in clusters of three throughout four fields located at and near Senda Darwin.

Seed traps were placed in clusters of three to increase the area sampled at each location.

Seed traps were modeled after those used by Amesto et al. (2001a). The traps were

constructed from a metal ring 30cm in diameter, supported by three metal stakes

approximately 50cm above the ground. Seed catching area for each trap was

approximately 0.07 m2. Plastic netting (mesh size = 2mm) was attached to each ring to

collect the seeds. I placed 20 of the trap clusters within 25 meters of the forest edge, and

20 from 25 50 m from the edge. The design was slightly unbalanced, because there

were insufficient trees within 25 meters of the edge. To determine whether avian seed

dispersal was higher beneath trees, I placed 19 of the traps directly beneath a tree, and the

other 21 at randomly chosen, non-tree locations. Traps were placed at the end of January

2001, and were checked at the end of February 2001, the beginning of January 2002, and

the end of February 2002. Only seeds that were of a different species than the tree above

the trap were counted, unless there was good evidence that the seed had been dropped by

a bird (i.e., fecal material evident). I tested for the effect of distance from edge on seed

rain and the effect of location (tree or non-tree) on seed rain using Mann-Whitney U tests.

Germination Site Hypothesis

Seedling transects

In 2001 I set up five transects in each of nine fields to determine where seedlings

were actually growing. Fields were chosen at random from the 12 fields that I sampled

for frugivore activity. I placed 100-m transects parallel to the edge at 10, 20, 30, 40, and

50 meters from the edge, in order to control for distance from edge. Along each transect,









I searched at three randomly selected points for each of the following substrates: bare

ground, dead wood, bare ground beneath a tree, dead wood beneath a tree, beneath

Baccharis magellanica, and on Sphagnum moss for a total of 18 randomly chosen points

on each transect. At each point, I searched a 5m radius for the presence of the substrate,

and counted the presence and number of seedlings growing on the nearest one-meter

square area of the substrate. All trees less than one centimeter in diameter were

considered seedlings. Seedlings were not identified to species, but were differentiated

from non-woody species.

Substrate availability transects

In 2002, I randomly chose six of the 12 fields to analyze substrate/cover type

availability. In each field, I established five 50-m transects perpendicular to the forest

edge, in order to incorporate any variability along this gradient. Using a measuring tape I

paced the transects, visually estimating the percentage cover of each substrate type for

each meter of distance, arriving at a percent cover for each substrate type over the total

50 meters. A Chi-squared test was then used to compare the number of seedlings found

at each substrate or cover type to the percent occurrence of each type.

Germination and survival experiment

In February 2001, I planted 1080 seeds of two species; Drimys winterii and

Amomyrtus meli, on three substrates; dead wood, bare ground, and Sphagnum moss to

compare rates of germination and survival on different substrates. Drimys and

Amomyrtus were chosen based on availability and their prevalence in successional fields.

At each of 18 locations, in two large Baccharis fields at Senda Darwin, I placed 10 seeds

of each species on 0. m2 area of each substrate. Only locations where all three substrates

occurred within 2m of each other were chosen, to control for localized site effects. Due









to this constraint sites were chosen based on availability, and thus were non-random. At

each location, I placed seeds in three microhabitats, each of which constituted a

treatment: (1) dead wood, (2) bare ground, (3) and Sphagnum moss. Prior to planting,

any existing seedlings were removed, and the substrate surface was lightly scraped to

remove any existing seeds. Seeds were lightly pressed into exposed soil or Sphagnum

moss and dropped onto crevasses or pressed into soft portions of CWD. In an actual

defecation event, seeds would be accompanied by fecal material, and would have been

subjected to intestinal acids, which may enhance germination (Traveset et al. 2001).

Each location was marked with small metal stakes and flagging tape. In January of

2002,, I recorded the percentage of seeds that had germinated and survived. The effect of

substrate type on germination and survival was evaluated using a Kruskal_Wallace test.

Results

Frugivore Activity Hypothesis

Bird Censuses

The mean number of frugivores counted during 10-minute point counts for fields

with low, medium, and high numbers of trees was highest for the sites with the most trees

(> 29) and lowest for the sites with less than nine trees (Fig. 2). In total, 75 frugivore

visits were recorded. Elaenia albiceps accounted for 64% of the observations, and

Turdusfalklandi made up the 36%. Mean number of frugivore visits at high sites was -4

times higher than and medium sites, where mean frugivore visits was -5 times higher

than low sites. Sites with low and medium numbers of trees were both found to be

significantly different from high tree sites, but not from each other (X2(2)= 7.78, P =0.02

N=12; Fig. 3).














O


O


I-





0-


high med low
Number of trees in field


Figure 2. Average number of frugivores counted during 10 minute point counts at 12
sites in Chiloe, Chile. Sites are grouped as high (>29 trees), medium (15-24
trees), and low (0-9 trees) numbers of trees (N = 12). Error bars represent +/-
1 SE.

Focal samples

During focal tree sampling, clusters of trees received significantly more bird visits

on average than single trees (Z(46)= -3.38, P = 0.001, N=24; Fig. 3). When

circumference was controlled for, clusters still received significantly more visits (Z(44)=

-2.396, P=0.02, N=46). Clusters received 6.6 times more visits on average than single

trees. However, there was no significant difference in the number of visits to fruiting and

non-fruiting trees, though the sample size was lower (Z(26)= -0.48, P =0.73, N=14; Fig.

4) Fruiting tree species received approximately twice as many visits on average as non-

fruiting tree species.






19





n 3-


0
2-


0_

I .
O-




Cluster Single
Tree configuration


Figure 3. Mean number of frugivore visits during 30-minute focal samples to clusters of
trees and adjacent single trees in degraded old-fields in Chiloe, Chile (N=
24). Error bars represent +/- 1 S.E.

Seed traps

Four hundred and ninety four seeds from fruiting tree species were collected in the

seed traps over the course of one year. Of these, 482 were found beneath trees (out of 16

trap clusters) and 12 were found in random non-tree locations (out of 2 non-tree trap

clusters). After discounting seeds that I was unable to verify as being bird dispersed (i.e.

seeds of the same species as the tree above the trap and not embedded in fecal material),

there were 110 seeds dispersed to locations beneath trees and 12 dispersed to random

non-tree locations. Seed rain of verifiably dispersed seeds within 25 meters of the edge

did not vary significantly from seed rain between 25 and 50 meters from the edge

(Z(38)=-0.03, P=0.98, N=20). Dispersed seeds were significantly more common beneath

trees than in random non-tree locations (Z(38)=-3.33, P = 0.004, N=20).









Germination Site Hypothesis

Seedling transects

I found the highest density of seedlings beneath trees, followed by dead wood

beneath trees, Baccharis, bare ground, dead wood, and Sphagnum (Fig.5).



1.0-
0
0.8-


0.6-
0
O4
.0 0.4-


S0.2-



Fruiting Non-fruiting
Tree category


Figure 4. Mean number of frugivore visits during 30-minute focal samples to fruiting
and non-fruiting trees in degraded old-fields in Chiloe, Chile (N = 14). Error
bars represent +/- 1 S.E.

Substrate availability transects

Baccharis was the most common substrate type, covering 34.57% of the sampled

area followed by ferns, Sphagnum and bare soil. Trees, trees with dead wood, and dead

wood were the least common substrate types (Fig. 6). Seedlings were found in

significantly higher numbers than would be expected beneath trees, and on dead wood

beneath trees, and in significantly lower numbers than expected on Sphagnum moss

(X2(5)=192, P < 0.001, N=157).











2 I



1.6


E
1.2 -



S0.8



0.4



0-
Tree + Dead Tree Dead Wood Bare Ground Baccharus Sphagnum
Wood
Substrate/Cover Type


Figure 5. Density of seedlings found growing on each substrate type in 9 degraded old-
fields in Chiloe, Chile.

Germination

Of 1080 planted seeds (N=360 on each substrate type), only 7 Drimys seeds

germinated and survived for one year. Five of these were on dead wood, two were on

bare ground, and no seeds placed on Sphagnum moss germinated and survived (Fig. 7).

Discussion

I conclude that seed dispersal limitation is the most important factor limiting

regeneration of Baccharis fields for the following reasons. Fields with no or very few

trees received almost no frugivore visits. Since very few seeds were found in traps

placed in open areas, some sort of perching structure appears to be necessary to

"jumpstart" regeneration. Availability of suitable germination substrate, while also

important, seems to be secondary to seed dispersal, since some seedlings were found in










all substrate types. The latter finding suggests that although some substrates might be

better for germination, increased overall dispersal into any sites could increase

germination rates. While the history of the study fields is not fully known, most were

logged at least 30-50 years ago (I. Diaz, Pers. Comm.). Presence of seedlings and trees in

some of my study plots suggests that succession to forest is not actually arrested close to

forest edges, but slowed enough to be experiencing limitations on regeneration

representative of truly arrested sites.




L 0.40-

0

O 0.30-
o

0
0.20-
o.


0.10-



0.00-
Tree DW Dead Wood Baccharis Fern
Tree Bare Ground Sphagnum
Substrate / Cover Type



Figure 6. Percent cover of various substrate types across six degraded old-fields in
Chiloe, Chile. Error bars represent +/-1 SE.

Frugivore Activity

Following my predictions, fields with more trees received more visitations from

frugivorous birds than fields with fewer trees. However, lack of a significant difference

between low and medium tree density suggests the existence of a threshold value for the

number of trees required to significantly increase the number of visitations, alternatively,









my sample size may have been too low to detect a difference. Based on my study design,

the threshold could be around 15 trees per hectare. Therefore, at minimum, my results

imply that if forest is logged and the land owner wants forest to regrow, then leaving at

least 15 trees/ha will positively influence the number of frugivorous birds frequenting the

area (and enhance the rate of regrowth; see Chapter 2). Due to the limited area of my

study plots, however, point counts on a larger set of fields varying more systematically in

the number of remnant trees should be conducted to obtain a more accurate determination

of numbers of trees needed to attract seed dispersers.


1200 1080

1000

800

0 600
I-

E 400
Z
200

2 5 0 7
0 -
Bare Ground Dead Wood Sphagnum Total Total Planted

Substrate type



Figure 7. Number of seeds germinating on each substrate type in 18 trials, in degraded
old-fields in Chiloe, Chile.

Focal sampling revealed that trees growing in large clusters attract more frugivores

than trees growing singly. This may be due to the fact that a cluster of trees provides

greater cover from predators and is more likely to contain multiple food resources than a

single tree (Debussche & Isenmann 1994; Ferguson & Drake 1999; Holl 1999;









McClanahan & Wolfe 1993; Slocum & Horvitz 2000). It is also possible that clusters of

trees simply receive more visitors due to their greater volume and likelihood of being in

the path of birds moving across open areas. But in any case, it may be more effective,

when leaving remnant trees during forest clearing, to leave them in small clumps.

Although, this may seem counterintuitive for pasture creation, trees can be useful in

providing shade and windbreaks for cattle. Additionally, if pasture creation is successful,

remaining trees can be used for fuel wood when it has become apparent that Baccharis

and Sphagnum are not becoming established.

Data from the seed trap study suggest that frugivores overwhelmingly deposit seeds

beneath trees. Although density dependent mortality can be greater beneath trees (Howe

& Smallwood 1982; Janzen 1970; but see Hubbell 1979; 1980), in this system it may be

compensated by conditions under trees that are significantly better for germination and

survival (i.e., elevated and drier soils, reduced daily variations in temperature and

humidity, and a lack of competition from shrubs and grasses; Nepstad et al. 1996). This

combination of seed attraction and favorable conditions for germination and growth

characterizes what has been called "recruitment foci", or points in old fields from which

regenerating forest grows outward (McDonnell & Stiles 1983; Slocum & Horvitz 2000;

Toh et al. 1999).

Seedling Establishment

The location of seedlings in fields further suggests that some of the seeds deposited

beneath trees survive to the seedling stage. Seedlings were significantly less prevalent on

other substrate types. After trees and trees with dead wood, the next most common

location for seedlings was beneath Baccharis. Because Baccharis is so common relative

to colonizing or remnant trees, on a per plant basis, Baccharis shrubs receive much less










seed input than trees in old-fields. However, Baccharis is occasionally used as a

perching site by the smaller bird species, and some seedling establishment might occur in

the absence of trees, albeit at an extremely slow rate. When seedling density is divided

by substrate availability, it becomes apparent that the rarest sites represent the highest

seedling establishment (Fig. 8).


S120


100


80-
IM


60


40-


c 20


0 0-B
Tree Tree + Dead Dead wood Bare soil Baccharis Sphagnum
wood
Substrate Type


Figure 8. Seedling density divided by substrate availability for various substrates in
Baccharis-dominated oldfields throughout Chiloe, Chile.

Testing germination success on different substrates yielded little data. Low

germination and survival rates in my study (0.97 %) may accurately reflect an extremely

low germination rate under natural conditions. In a study of germination of the same

species under green house conditions, however, a much greater percentage of seeds

germinated (-75 %; Figueroa et al. 1996). Actual germination rates may have been

higher, however, because I was unable to check for germination between field seasons. It

is quite possible that germination rates were much higher, and that subsequent survival









was low. Moreover, it is not yet known for this system how gut passage affects the seed

germination process (Traveset & Willson 1997). Further testing of seed germination on a

variety of substrate types after frugivore gut processing would be beneficial.

Recruitment Foci

I found more seedlings growing beneath trees than any other location. This is

somewhat expected a priori (Nepstad et al. 1996; Slocum & Horvitz 2000), and my work

also suggests that this is due to some combination of increased seed input and better

conditions for survival of seeds and seedlings beneath trees (Figs 2, 5). From the seed

trap portion of the study, we know that seeds are being dispersed, and that the area

beneath trees is the main recipient of these dispersed seeds. What we don't know is

whether this area is also better for seed germination. However, the presence of a tree

suggests that the area was historically a favorable microsite, particularly because most

trees used for this study probably came in after the fields were created; the oldest tree that

I cored was 54 years old. Typically, dispersal away from a parent tree is thought to

increase seed survival by decreasing intraspecific competition (Hubbell 1979), as well as

by providing escape from predators and pathogens (Connell 1971; Janzen 1970).

However, in the case ofBaccharis dominated old-fields, survival away from the parent

tree is likely much lower, unless the seed is dispersed beneath the canopy of another tree.

For future work it will be important to examine how distance from a tree influences

germination success in sites where succession is arrested, as the rate of survival should

reach a maximum closer to the tree, then drop off drastically, reflecting the unfavorable

conditions of the site (Houle 1995). Ascertaining how close to a tree a seed needs to fall

to maximize its germination success would facilitate restoration planning, by allowing

estimation of optimal tree planting density.









Conclusions

In order to speed-up or initiate forest regeneration in Baccharis dominated shrub

fields, lack of seed dispersal and a lack of suitable germination sites must both be

overcome. My study suggests most strongly that seed dispersal is the main limiting

factor. Moreover, I can conclude that the simplest means of overcoming arrested

succession will be to plant tree saplings in fields where forest regeneration is desired.

And while my work did not find overwhelming evidence that the presence of dead wood

(nurse logs) is necessary to overcome germination limitation, evidence that this would be

important is mounting from this and other rainforest systems (Harmon et al. 1986; Papic

& Armesto 2000; Takahashi et al. 2000). Therefore, based on the field studies and

literature review presented here, I put forward the following conservative

recommendations for reducing Baccharis coverage where it exists, or fostering forest

development rather than shrub development where forest is freshly cleared. (1) Plant, or

preferably leave behind, a small number of forest trees after clearing; at least some in

clumps. (2) Leave a significant amount of coarse woody debris throughout the area of

interest to provide safe germination sites. In the next chapter I explore recommendations

in greater detail using an empirical systems model of forest regeneration under different

starting conditions in order to allow landowners greater certainty in applying these simple

recommendations to their own fields which may vary in starting conditions that could

affect the outcome.














CHAPTER 2
RAINFOREST RESTORATION SCENARIOS FOR BACCHARIS- DOMINATED
OLD-FIELDS IN SOUTHERN CHILE: A SIMPLE ECOSYSTEM MODEL AS A
DECISION MAKING TOOL

Introduction

Determining the underlying mechanisms of arrested succession at local restoration

sites can influence understanding and management of larger landscape scale patterns and

processes (Bell et al. 1997). For example, arrested succession is one way that stable and

widespread community types (or alternate stable states; (Laycock 1991; Lewontin 1969)

can arise that, in turn, define the mosaic of ecological and socioeconomic characteristics

of human landscapes (Naveh 1994). Therefore, when undesirable alternate stable states

arise in landscapes, restoration ecological approaches can be used to identify and release

mechanisms that generate and maintain them. In this study, I identified potential

mechanisms of arrested succession in mesic old-fields resulting from anthropogenic

clearing of south-temperate rainforest in Chile. The community type produced that is

undesirable, from both socioeconomic and ecological perspectives, is a persistent shrub

community dominated by Baccharis spp.

Chilean South-temperate Rainforest: Natural Disturbance Regime and Arrested
Succession

The South temperate rainforest in the Valdivian region of Chile (35- 480 S)

receives between 1,000 and 6,000 mm of rain per year, and is characterized by emergent

evergreen broad-leaved trees (e.g., Nothofagus oblique, N. alpine, andN. dombeyi) and

evergreen conifers (e.g., Podocarpus nubigina). Canopy species include Drimys winterii,









Weinmannia trichosperma, and several trees in the family Myrtaceae (Willson et al.

1994). Vertebrate dispersed species make up approximately 70% of the flora (Armesto &

Rozzi 1989) and include Drimys winterii, Amomyrtus luma, A. meli, Eucryphia

cordifolia, Weinmannia trichosperma, Podocarpus nubigena, Laurelia philippiana, and

N. nitida. Two bird species, Eleania albiceps, and Turdusfalklandii disperse the

majority of seeds. Endemism is very high in this biome (Stattersfield et al. 1998),

ranging from 45 % in vertebrates to 90% in some groups of seed plants (Armesto et al.

1996; Villagran & Hinojosa 1997).

The natural forest disturbance regime is characterized by large-scale periodic

catastrophes including earthquakes, volcanic activity, and fire that occur relatively rarely

in a given site. Wind throw and tree fall gap creation occur much more frequently at any

given location and at smaller scales (Veblen 1979). The natural disturbance regime

prevents shade tolerant tree species such as Laurelia phillipiana and Saxegothea

conspicua from out competing the shade intolerant Nothofagus (Bustamante & Armesto

1995) and helps maintain a typically diverse tree canopy composition across scales and

throughout the Valdivian region. Human disturbances include widespread clearing of

forest followed by burning and then conversion to agriculture (especially in the lowlands)

and to industrial forestry plantations of exotic pine and eucalyptus. Typical farm and

plantation plots are much larger, and more highly altered, than natural clearings.

Productivity of cleared lands can be limited by the development of persistent

shrubs (dominated by Baccharis magellanica) following forest clearing that prevent

livestock or row crop production and natural forest succession. Apparently the Baccharis

shrub canopy develops where the water table is, or becomes, elevated following forest









clearing. While soil inundation alone can prevent establishment of native trees (Bewley

& Black 1982), Baccharis overstory may also prevent establishment by desirable species

(Cespedes et al. 2002); Putz and Canham 1992). Landowners can bum back the shrubs

but without prohibitively expensive ditching and draining, fire alone does not reliably

improve site utility for either agriculture or forest regeneration (Pers. Obs.). Baccharis

species are common disturbance-related site invaders in many regions where they are

native, as in this study (Stylinski and Allen 1999; Sarmiento et al. 2003). A natural

ecotonal vegetation type called Magellanic moorland, occurring at higher elevations in

Chile, has similar plant community properties to the Baccharis shrublands that represent

arrested succession in the lowlands (Ruthsatz and Villagren 1991). But monospecific

Baccharis shrublands are not naturally widely distributed in the study area and are a

direct result of clearing at a larger scale than the natural disturbance regime.

On Isla Grande de Chiloe, up to 30% or more of local landscapes currently persist

in this arrested state, and extensive areas have been dominated by uniform Baccharis

stands for more than 50 years (T. M. Darnell, unpublished data). Baccharis shrublands

are resource poor and not productive of native wildlife (Sieving unpubl. data) or

agricultural produce. Farmers that clear forest and then see the development of

Baccharis shrub in the cleared area will move on to clear another site, if available, to try

again. Therefore, from the perspective of maximizing biotic productivity in this

landscape, restoration of Baccharis fields to either pasture or native forest would be

beneficial. Regeneration of forest would provide wildlife habitat and forest products for

people, and viable pasture creation would stave off further clearing of native forest (Fig.

1). Given that forest clearing and human activity in southern Chile have resulted in









global endangerment of endemic flora (Armesto et al. 1998) and fauna of the region

(Stattersfield et al. 1998), native forest restoration is a high priority. In this chapter I

explore two likely causes of arrested forest succession in through systems modeling of

restoration scenarios I developed based on field data (from Chapter 1).

In Chapter 1 I tested two alternative hypotheses for arrested succession in my

system: that shrub fields persist because of lack of seed dispersal into them, and that they

persist because associated conditions prevent seed germination and seedling

establishment. I found evidence that both mechanisms are operating to suppress

succession in Baccharis fields on Chiloe Island, in addition to supporting justification

based on the work of others. Therefore, in this chapter I develop a dynamic systems

model that incorporates processes actuating both mechanisms of arrested succession in

order to address realistic scenarios for restoration given constraints and goals relevant to

local landowners.

Modeling Restoration Scenarios

Several restoration scenarios are possible for the patchwork of cleared old-fields

that exist in the temperate rainforests of southern Chile. One possibility involves the

creation of drainage ditches followed by controlled burning, to dry out the shrub fields

and create pasture. While this does not replace forest, it does reduce the amount of

additional forest that needs to be cleared. A second potential approach is to plant rows of

trees with high rates of evapotranspiration (Ferro et al. 2003); trees can function as

"pumps" in this scenario, reducing the hydric conditions and allowing easier conversion

to pasture or forest. However, both of these approaches are costly, and depend on

monetary resources of individual landowners; many farmers in the region do not have the

resources to employ these techniques (Personal Observation). A third general approach,









and the focus of this modeling exercise, is the restoration of native forest via restoration

of secondary forest succession in shrub-dominated fields. Forest regeneration on arrested

sites could lead to some economic gain for landowners (from forest products or provision

of livestock shelter and understory browse during winter) and may, thereby, lessen

pressures to clear more forest.

In Chapter 1 I presented evidence that seed dispersal can be limited if birds are not

attracted to shrub fields. Attractiveness is determined by the number of trees, whether

they are fruiting, and their spatial distribution. Planting trees can help attract seed-

dispersing birds into shrub fields, and potentially provide better microsites for

germination and seedling establishment. It is also clear that because ground-level

substrates where seeds might fall can be either too wet (bare ground) or too dry or acidic

(Sphagnum), coarse woody debris (CWD) coarse woody debris is the best germination

substrate because it will catch and germinate seeds that would otherwise be inundated in

water or desiccated in the Sphagnum layer (Lusk 1995; Papic & Armesto 2000).

Therefore, suitability of germination and seedling establishment sites are determined by

relative availability of CWD versus adverse substrate conditions. Both planting trees and

adding CWD cost money and labor. Therefore, the goal of the model is to assess the gain

(in rate of forest regeneration) against the estimated relative costs of inputs.

This chapter presents a dynamic systems model of seed disperser response to field

characteristics, seed germination processes, and tree establishment and growth in the old-

fields adjacent to south-temperate rainforest on Isla Grande de Chiloe, Chile (41055'S,

73035'W). The purpose of this model is threefold. The first is to expand our

understanding of the relative importance of seed dispersal by birds and the limitations of









seed germination as key components of forest succession. The second is to explore

methods for restoring rainforest by simulating removal of limitations to the process of

succession (i.e. lack of trees, insufficient germination sites). The third objective is to

assess the model as a predictive tool that can be used by managers and landowners who

want to restore native forest, given certain constraints (i.e., time, money, labor). I use

field data (Chapter 1) and information from other studies, to parameterize the model,

which necessarily simplifies certain aspects of the regenerative process. For example,

interactions between tree species, soil properties, and nutrient levels are ignored.

Sensitivity analyses were applied in cases where model parameters are estimated or

derived from the literature to analyze their importance, to provide a range of outcomes

when a parameter is not well known, and to generate useful hypotheses for further

testing.

Model Description

Overview

The model is comprised of five subsystems: old-field characteristics (size, cover

and substrate types, length of edge), seed rain, seed germination and seedling survival,

creation of new recruitment foci, and tree growth and survival (Fig. 9). The underlying

premise of the model is that succession acts as a positive feedback loop during early

stages such that trees present in the old-field act as an attractant for avian seed dispersers.

As seeds are dispersed to suitable germination sites, the number of trees increases as does

area covered by tree canopy and, in turn, this attracts more seed dispersers and alters the

habitat to create more germination sites (Alcantara et al. 2000; Slocum 2001). At later

stages, as tree density increases, homeostatic properties slow forest growth

(e.g.,Tappeiner et al. 1997; Wills et al. 1997). Rather than attempt to explicitly model









every aspect of the ecosystem, select attributes are modeled incorporating available data

from field studies. For example, instead of modeling all processes that could influence

tree growth rates; I incorporated rates of tree growth recorded in nearby systems (see

below). Stella 7.0 (Wallis et al. 2001), an icon based modeling environment, was used to

create this model. Following is a brief description of key model equations. The model is

presented in its entirety in Appendix 1 iconographicc form) and Appendix 2 (equation

form).


Figure 9. A simple representation of old-field regeneration as conceptualized by the
model. Dark arrows represent the flow of material (through space or time
i.e. seeds or tree growth), dashed arrows represent the influence of one
component on another.









Old-field Characteristics

The parameters I defined to reflect old-field characteristics include: area of ground

cover of 5 different types; area of tree canopy cover; and the rate of spread of tree cover

in the face of competition with other ground cover types. Cover types defined here (and

observed in the field; Chapter 1) are trees, Sphagnum moss, course woody debris (CWD),

Baccharis, grasses and ferns, and bare ground. The category "grasses and ferns" refers

to vegetation that is greater than 0.25 m in height and that is sufficiently dense to limit the

amount of light striking the soil. The category "bare ground" refers to either exposed soil

or patches of sparse grasses and soil. Each of these cover-types either promotes or

inhibits tree growth. Tree growth in this instance refers to the increase in total tree

canopy area throughout the old-field. Other trees, CWD, and bare ground act as

promoters of tree growth (i.e., area of canopy cover can increase in and around areas with

these cover types). On the other hand, tree seedlings have difficulty getting established in

areas with Sphagnum, Baccharis, and grasses and ferns. Therefore, growth of forest area

is inhibited where these substrate types occur (Chapter 1).

To represent proportion of canopy coverage, the canopy cover of trees in the old-

field is calculated relative to the size of the old-field. This is shown in Eq. 1


Ca= Fa/Oa (1)

where Fa is the area of ground covered by tree or tree canopy (also referred to as a focus),

and Oa is the area of the old-field. When determining the actual area available for tree

growth, the various substrate types are subtracted from the total size of the old-field, as in

Eq. 2


Ga = Oa (SPcov + Bcov + GFcov)









where Ga is the area available for trees to grow, SPcov is the area of old-field covered by

Sphagnum moss, Bcov is the area of old-field covered by Baccharis, and GFcov is the

area of old-field covered by grasses and ferns. This means that trees are in competition

for resources with grasses, shrubs, and mosses. In other words, when shrubs and mosses

dominate an old-field, the area available for new tree growth is much lower than in an

old-field without competing cover types because there is less available space and

resources for the trees to grow.

Little data are available concerning the rate of spread of Sphagnum moss.

However, it has been shown that Sphagnum out-competes small plants by creating acidic,

anoxic, and nutrient poor conditions (Van Breeman 1995). This model incorporates

Sphagnum as a static feature that is slowly out-competed by trees, which, once rooted

with a canopy over the top of the shrub layer, can grow uninhibited. Although the actual

mechanisms underlying tree-shrub and tree-Sphagnum competition are poorly understood

at present, it is assumed that as forest trees grow, microclimatic changes occur, making

the habitat more suitable for trees and less suitable for Baccharis and Sphagnum.

Figueroa and Lusk (2001), for example, found that Baccharis shrub canopy is highly

shade intolerant. Additionally, Ohlson et al. (2001) describe the impedance of Sphagnum

growth by Pinus sylvestris in a boreal bog ecosystem. Eq. 3 calculates the rate at which

growing trees can out-compete Sphagnum moss.

SPcov(t) = SPcov(t dt) (IF P > 0.99 THEN I Fc *(1/180) ELSE 0) dt (3)

where SPcov is the area covered by Sphagnum at time t, P is the proportion of the old-

field covered by any substrate type other than bare ground (including forest), Fc is the

circumference of the recruitment foci, and dt is equal to one month. The value in









parentheses, 1/180, is an estimate, subjected to sensitivity analyses, as little data

regarding Sphagnum-tree competition is available. Therefore, as the edge of the

recruitment focus meets a patch of Sphagnum moss, it takes 15 years to shade out each

square meter of Sphagnum moss that is adjacent to the regenerating patch of trees. The

time required for forest to out-compete opposing cover types is unknown, and requires

further research. Therefore, very conservative values were chosen.

Eq. 4 and 5 show how Baccharis and grasses and ferns are similarly handled.

Bcov(t) = Bcov(t dt) (IF P > 0.99 THEN I Fc *(1/288) ELSE 0) dt (4)

GFcov(t) = GFcov(t dt) (IF P > 0.99 THEN I Fc *(1/48) ELSE 0) dt (5)

It takes 24 years to shade out 1m2 of Baccharis, and four years to shade out 1m2 of

grass or ferns.

Other old-field parameters include the length of forested edge bordering the field,

and the number and size of existing trees in the old-field, and are chosen to represent the

field of interest.

Seed rain

Based on data collected (Chapter 1), it is assumed that avian-dispersed seed rain

occurs with equal frequency within 50m of the forest edge, given appropriate conditions

(this was the scale of the field studies in Chapter 1). Hence, the dynamics of this model

are relevant to forest growth within 50m of the forest edge at larger distances into shrub

fields, certain rates may differ.

Two measurements were used to calculate the potential for seed dispersal into a

field; the number of avian frugivores inhabiting forest adjacent to the old-field, and the

seed rain beneath trees in old-fields. Bird censuses were conducted in forest edge habitat

adjacent to Baccharis dominated old fields to estimate the number of frugivores in









adjacent forests. Between 07:00 and 10:00 during January and February 2002, I walked

two 200m line transects into each of three different forest patches and counted frugivores

seen or heard within 40m of the transect. Due to the dense nature of the forest, distance

was estimated, not measured. Forest patches were chosen based on proximity to pre-

existing study sites and transect direction was chosen randomly. The average number of

frugivores counted was 12.5 per 80m wide transect. Elaenia albiceps will travel

hundreds of meters from the forest and into the adjacent old-field, if there is sufficient

structure in the field (Willson et al. 1994). This means that for each meter of edge there

are 0.16 birds with access to the old-field. This method is prone to under-counting

(Bibby et al. 2000), but to avoid over predicting rates of forest regeneration, the outcome

was not modified. This is high relative to linear densities of most breeding birds, but the

primary seed disperser, Elaenia albiceps, occurs at very high densities in this system

(Rozzi et al. 1996; Willson et al. 1994). To simplify the model, Turdusfalklani was not

included.

To estimate seed rain, 60 seed traps of roughly 0.07 m2 in surface area were placed

beneath trees in 2 old fields (see Chapter 1 for details). Over a one-year period, 92

verifiably dispersed seeds were found in 60 traps. Hence, 4.2m2 of old-field received 92

seeds in one year, generating a rate of 21.9 seeds/ m2*year, or 1.83 seeds/ m2*month

beneath trees.

Eq. (6) shows how the potential seed rain (PSR) was calculated.

PSR = (Sr *Oa)*(L/100)*(Db/.015625) (6)

where Sr is the number of seeds dispersed per unit of attractive habitat in the old field, L

is the length of the edge, and Db is the density of birds in the forest edge with access to









the old-field. As the number of trees in the old-field increases, the ability of the birds to

disperse seeds equally to all areas decreases. However, this levels off to some degree

because as the old-field becomes converted to forest it begins to provide suitable habitat,

increasing the seed disperser population (Figure 10 shows how the value of S changes

with the number of trees present in the old-field). The second part of this equation,

(L/100)*(N/16), adjusts the amount of seed rain for the length of edge, based on the bird

survey data suggesting that there are 16 frugivorous birds per 100 meters of edge.

Because the value of S is derived from real-world sampling efforts where the edge and

density were 100m and -0.16 birds/m of edge, respectively, these values allow the model

to adjust S to varying conditions (i.e. when there is 100m of edge N = 16, if there are

more or less than 100m of edge the model adjusts N).

Actual seed rain is based on the proportion of old-field covered by trees. The

portion of the potential seed rain that is dispersed is equal to the ratio of area beneath

trees to total field size (0.5 ha). However, most of this seed rain does not contribute to

the expansion of recruitment foci. In many systems, seeds that fall under existing canopy

tend to die off or form a seedling bank where they replace adults that die (Antos et al.

2000; Stewart et al. 1991; Szwagrzyk et al. 2001). From the tree's perspective, avian

frugivores both bring seeds from other species and individuals as well as disperse seeds

for the tree. In any case, this results in a seed shadow that typically results in a

leptokurtic distribution conforming to a negative exponential curve (Willson & Traveset

2000). Long distance dispersal events are dealt with below via the creation of new foci.

For the purposes of the model, it is the short distance dispersal that we are mainly

interested in, as seeds dispersed too far from the tree are likely to be dispersed in poor









quality habitat unsuitable for germination, and would not influence regeneration in focal

sites at this scale. Based on this traditional model of seed shadows, I expect seed rain to

be highest beneath the tree canopy (where it contributes to replacement of existing trees)

and, due to the lack of alternative perching sites, to drop off quickly as distance from the

canopy increases. Additionally, I expect conditions in most of the old-field to be

unsuitable for germination such that with the exception of rare conditions (discussed

below) survival is only possible in the immediate vicinity of existing tree canopy,

resulting in a process of"nucleation" (Debussche & Isenmann 1994), where the only seed

rain that contributes to tree growth occurs in the outer edge of the tree canopy.

Therefore, only the seeds that fall in the outer 0.5 meters of a tree canopy actually

contribute to the growth of recruitment foci in the model. This is represented by Eq. (7).

Fg = Af- (\(Af /t)-0.5)2*" ) / Af (7)

where Fg is the part of the focus where seed rain contributes to growth in area and Af is

the area of the focus (m2). For simplicity, recruitment foci are assumed to be circular in

shape. Eq. (8) describes the actual seed rain (ASR).

ASR = PSR*(ZAf/Oa)* YFg (8)

Seed Germination and Seedling Survival

Germination rates were determined by planting seeds of two species in various

substrates throughout an old field (see Chapter 1). Because the rate of germination and

survival was 0.97% and may not have adequately mimicked natural conditions, I

compared it to the germination rates obtained by Figueroa and Lusk (2001) for Drimys

winterii; these were observed to be -75% annually in gap conditions, and as low as 16%

in low light conditions. However, in this study, seeds were protected from predation. In

a study on seed predation in the same area, as many as 65% of seeds were removed by









predators (Diaz et al. 1999). Taking a conservative approach, I chose 6% (65% of 16%)

as the germination rate for the model. To account for the large difference, I ran a

sensitivity analysis to determine the impact of changes in the germination rate (see results

section). At rates below 1%, changes in the germination rate have a large impact on rates

of forest growth; above 1% the difference is much less. Eq. (9) shows how germination

is calculated.

G = Sf*Gr (9)

where G is germination, Sf is the number of seeds in the recruitment focus, and Gr is the

germination rate.

The proportion of seedlings that become established as saplings is based on a

mortality of 1% per week (for Drimys winterii) under controlled conditions in high light

and 4% per week in low light (Lusk & Del Pozo 2002). I took the conservative approach

and chose 4% per week to arrive at a number for the model. This value may be low

because it was determined under controlled conditions, but I explored this also with

sensitivity analyses. Due to the prevalence of standing water in arrested old-fields, the

presence of dead wood, acting as "nurse-logs", increases the chance of seedling survival.

Papic and Armesto (2000) found that survival was 12 times greater on dead wood that on

surrounding terrain. This is reflected in the model by setting the death rate for seedlings

on dead wood 12 times lower than that of other seedlings.

Creation of New Foci

In addition to the growth of any existing recruitment foci in the old-field, there is

some chance that new foci will be created as time passes. Since birds perch on Baccharis

or other small shrubs in the field (Personal Observation) or occasionally defecate while

flying, there is a small possibility that a seedling will become established at a new









location in the old-field each month. Determining the value of this likelihood requires

further research, and estimating it is quite difficult. However, I made the assumption that

there is a 100% chance of a seedling becoming established below a tree in any given year

(Slocum 2000), and that the proportion of the various cover types influences this

probability of occurrence. Adjusting for area, I then compared the number of seedlings

that I found beneath non-tree cover types (Chapter 1) to the number found beneath trees

to calculate the odds that a seedling would be established there annually (Table 1).

However, the presence of a seedling does not necessarily guarantee a surviving sapling.

To further represent the likelihood of sapling establishment, I included the transition

probabilities (e.g., Rey & Alcantara 2000) determined by Papic and Armesto (2000) for

survival of seedlings. However, because the process of survival from seed to established

sapling in the harsh conditions of a cleared old-field is poorly understood, I erred on the

side of caution and assumed that only 10% of the sites were suitable for new foci

establishment to occur. Eq. (11) shows how the random chance of new foci appearing is

handled.

IF Af= 0

(on dead wood) THEN MONTECARLO1 ((0.10/12)*DW)*0.23)*r)

or (11)

(on bare ground) THEN MONTECARLO ((0.06/12)*BG)*0.02)*r)

ELSE 0

where DW is the percent cover of dead wood in the old-field but not part of an existing

focus, BG is the percent cover of bare ground in the old-field, and r is the additional

1 Monte Carlo is a function that returns either a 1 or a 0 each time step. The number in
parentheses determines the frequency that a 1 is generated.









rarity of new focus formation (r=0.1 in this case). Values are divided by 12 to convert

the annual value to the monthly time step used in the model. (To adequately represent

multiple foci with the Stella software, it is necessary to "array" many of the variables in

the model. The if-then-else structure of the equation ensures that new foci do not appear

where one already exists.) This equation generates a random chance of a seed falling,

germinating and surviving, which increases with the presence of dead wood and bare

ground.


2
O 1.8
S 1.6
1.4
e 1.2

I
C 0.8
C 0.6
0.4
S0.2
Q)Df 0.2 --------------------------------
0
0 200 400 600 800 1000 1200 1400 1600

Area of tree cover in old-field (square meters)



Figure 10. The value of S, or seed rain per meter of forest in the old-field, changes with
increasing tree cover in the old-field. This is because the existing seed
disperser population becomes less able to distribute seeds as more attractive
habitat opens up. Eventually, seed disperser population should increase,
allowing the curve to level out.

Tree Growth and Survival

Ages of 16 trees in old fields were determined by coring in order to estimate the

tree growth rate. Tree cores were extracted as close to the ground as possible, then









sanded with progressively finer sand paper. After repeated sanding, growth rings were

visible and were counted under a binocular microscope (Stokes et al. 1968). Cores were

not cross-dated, so missing and false rings may be a potential source of error (Lusk

1999).

Table 1. Comparison of the number of seedlings found beneath non-tree cover types to
the number found beneath trees to estimate the odds that a seedling would be
established there.
Number of
seedlings in
r # of Proportion of field if entire Seedlings
Cover cover type to 5000m2 field Compared to
seedlings 5000m2 field per meter2
type total area tree
te found (x) reed () were same as (n/5000)
surveyed (p)
cover type
([1/p]*x)=n
Tree 89.00 0.04 2225 0.44 1.00
Dead
d 7.00 0.03 233 0.05 0.10
Wood
Bare
Bare 17.00 0.12 141 0.03 0.06
Ground
Baccharis 27.00 0.35 77 0.01 0.03
Sphagnum 8.00 0.15 53 0.02 0.04


Tree growth rate is based on the positive correlation found between age and DBH

(Fig. 11) and age and height (Fig. 12). The standard "tree" for this model is a 10cm DBH

tree with a canopy covering 1.68 m2 of old-field. This was determined by measuring the

canopy cover of 15 adult trees and averaging the area covered. Trees in this model are

broken down into 5 size classes; saplings with a DBH < 5cm, and trees with a DBH of

5cm, 10cm, 20cm, and 30cm or more. Between the size classes, canopy cover area is

assumed proportional to DBH. The growth rate for D. winteri to each size class is based

on commercial forestry data (Navarro et al. 1997).












R-Square = 0.67

O\


O 0


0
40- 0 o
0 0

30-
0
.- / o


20-

/ 0
10-
0

4 8 12 16
DBH (cm)


Figure 11. Correlation between age of Drymis winterii (as determined by coring) and
diameter at breast height (DBH).

Navarro et al. (1997) reported an annual mortality rate for D. winterii of 2.17%,

which equals 0.181% mortality per month. The number of trees in each size class is

governed by Eq. 12

SCx(t) = SCx(t dt) + (Gin Gout Dx) dt (12)

where SCx is the size class at time t, Gin is the input of trees growing to that size class,

Gout is the trees growing out of the size class, and Dx is the death of trees.

Sensitivity Analyses

I conducted sensitivity analyses in order to examine the impacts of changes in key

components of the model on model predictions (e.g., Halpern et al. 2005). A base set of

conditions representing an "average" field were used for all sensitivity runs (Table 2).

These values were determined by line transect sampling. The proportion of









cover/substrate type was visually estimated at each meter along five 50 meter transects in

each of six fields. The proportions of cover and substrate types were then averaged.

Table 2. Parameters used for sensitivity analyses. Values represent mean of 6 fields.

Initial value (m2) used in
Parameter
sensitivity analyses

Cover of Baccharis 1728.5 m2

Cover of Sphagnum 762.5 m2

Cover of Bare soil 603.5 m2

Cover of Tree (30 trees of size class 2)

Cover of Deadwood 312 m2

Cover of Ferns and 13
1513 m2
grasses

Field size 5000 m2

Edge length 100 m

Bird density 0.16 birds with access/m of edge

Germination rate 0.60% of seeds per month


Each parameter in question was then run repeatedly through an incremental series

of values centered on the mean value reported in Table l(e.g., Hallgren & Pitman 2000).

Sensitivity analyses that were run are listed in Table 2, and include characteristics of the

field and model parameters such as germination rate and rates of competition between

forest trees and other cover types

Given that field data were collected in old-fields with less than 50% tree cover, it is

doubtful the model, parameterized with such data, will sufficiently reflect the population

and community dynamics of largely forested sites. Controls on, and relationships among,









bird density, inter-specific competition, and microclimatic conditions are likely to change

over time, especially as forest grows and changes the quality of the habitat. Therefore, I

do not discuss the ramifications of model predictions regarding regeneration greater than

50%, and I present 75% and 100% forest regeneration model results as points of

comparison only. Due to the stochastic nature of the model I report average times until

regeneration. Finally, because of the sheer volume of computing time required to run

multiple analyses, minimum, median and maximum values were averaged based on 6

runs each.


o

50- R-Square = 0.57 /

0 o

M 40- /o o



30-
o- 0/




0
10-
0

4 6 8 10
Height of tree (m)

Figure 12. Correlation between the age of Drymis winterii (as determined by coring) and
tree height.

Management Scenarios

I applied the model to three management scenarios; a high cost, medium cost, and a

do-nothing scenario. The high cost scenario assumes that the land owner has sufficient









labor and money available to add dead wood to the field, plant saplings, and perhaps even

burn back some of the Baccharis and Sphagnum. In this scenario, non-forest cover types

were reduced by 5% each, CWD cover was increased by 5%, and 25 saplings are planted

each year. The medium cost model assumes that the farmer has an active interest in

restoring the old-field, but has little resources to do so. This farmer may plant a few trees

when the time is right, but will not be able to burn back the shrubs and lichen or drag

abundant logs or cut pieces of CWD into the field. In this scenario CWD cover is

increased by 1% and 10 saplings are planted annually. The do-nothing, or low cost,

scenario assumes that the landowner lacks either resources or interest, and leaves the land

as it is. Each management scenario was run with 3 starting conditions poor, medium,

and high quality fields. These three field types represent points along the continuum

from very low quality fields (no trees, a lot of Baccharis and Sphagnum cover, no CWD),

to very high quality fields (many trees remaining, no Baccharis or Sphagnum cover, good

CWD coverage, Table 3). In all, 9 starting conditions (3 cost by 3 starting condition

scenarios) were run 5 times each.

Results

Initial Conditions

Coarse woody debris

On average, increasing the amount of dead wood has a positive impact on the rate

of forest regeneration (Fig. 13). However, as more CWD is added, diminishing returns

occur, particularly after 1-2% cover of CWD. In fields with zero initial trees, the impact

of CWD is more pronounced (Fig. 14).










250



200
A A A
A A A

150



100 -75%
100%

50




0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00% 7.00%
Initial CWD in Oldfield (percent cover)


Figure 13. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial amount of
coarse woody debris (CWD) is increased.

Baccharis

As the initial coverage of Baccharis increases, the time required for forest

regeneration increases (Fig. 15). In fields with less than 30% Baccharis cover, little

effect is seen, but this is because I am measuring time until 50% regeneration, and at

these low values there is little obstacle to regeneration of 50% of the old-field. Above

30% Baccharis cover, the impact is drastic, but is dependent on the values chosen for the

rate at which forest "shades out" Baccharis (see below).

Sphagnum and grass

Changes in the initial coverage of Sphagnum and of ferns and grasses show a

similar pattern to that exhibited by changes in the coverage of Baccharis (Fig. 16 and Fig.

17). However, the percent cover required to slow forest growth corresponds with the rate










at which forest is able to out-compete either cover type. Thus, Sphagnum has less impact

than Baccharis, and grass less than Sphagnum.


250


200


150 .. : ... 25%
U. ."*" 50%
M P %PM** m6 0 m 75%
100 ----= 100%


50


0
0 ----------------------------
0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00%
CWD in Oldfield (percent cover)

Figure 14. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial amount of
coarse woody debris (CWD) is increased with zero initial trees in the old-
field.

Tree cover

Initial tree cover has some effect on rates of subsequent forest regeneration (Fig

18). The time required for 25% and 50% forest regeneration is somewhat reduced by

adding more trees.

Length of edge

Increasing the length of the border with adjacent forest proportionately decreases

the time required for forest regeneration because of the increased availability of seed

dispersers and seeds. For example, the difference between 125 and 225 meters of edge is

far less than the difference between 25 and 125 meters of edge (Fig. 20).






















100


50


0
0.00%


25%
* 50%
75%
100%


10.00% 20.00% 30.00% 40.00% 50.00% 60.00%

Initial Baccharis cover (percent cover)


Figure 15. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial coverage of
Baccharis is increased.


0 -
0.00%


25%
* 50%
75%
100%


10.00% 20.00% 30.00% 40.00% 50.00% 60.00%

Initial Sphagnum cover (percent cover)


Figure 16. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial coverage of
Sphagnum is increased.





















100

50

0
0.00%


25%
* 50%
75%
100%


10.00% 20.00% 30.00% 40.00% 50.00% 60.00%
Initial Grass and Fern (percent cover)


Figure 17. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial coverage of
grasses and ferns is increased.


25%
- 50%
75%
100%


100 200 300 400 500

Initial Number of trees (20cm>DBH >5cm) in old-
field


Figure 18. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial number of trees
is increased.










300

250

200 "
,. 25%
S150 50%
a N 75%
> 100 ..-M*.. ..% 100%
1 0 0 -- -"- E "---*--m *

50

0
0 50 100 150 200 250
Length of Edge

Figure 19. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the length of adjacent edge
is increased.

Bird density

As expected, forest regeneration rate is positively correlated with bird density (Fig.

20). The range in bird density values correspond with roughly 150 years in forest

regeneration times.

Sensitivity Analyses

Germination and tree growth rate

The variation among rates lower than 20% is much greater than among rates

between 20 and 70% (Fig. 21). The value for the germination rate was difficult to obtain,

and additional research is needed before the model should be used to make predictions.

Field observations (Chapter 1) revealed a low germination rate of 0.98%. Conversely,

controlled experiments (Figueroa & Lusk 2001; Figueroa 2003) found a germination rate

for the same tree species of 75%. A reasonable compromise would return a germination

rate of 35.49%, but as the model demonstrates, this is too high because at such high










germination rates, forest regenerates much faster than historical observations suggest. I

conservatively chose a 6% rate of germination for the model. When subjected to

sensitivity analysis, little change was seen for germination rates above 1% (Fig. 21).

There is a -125 year variation in time until 50% regeneration between 0% and 1% (Fig.

21). This extreme variation demonstrates that getting more accurate estimates of this

value is critical to the successful prediction of forest regeneration. As seen in Figure 22,

regeneration times vary proportionately to tree growth rates.


350

300 ,-

250

u ) 2 0 0 --- 2 5 %
S200 -* 50%
S. 75%
>- 150 100%

100 ---"'--------i r-". 100%
100

50

0 -
0.05 0.1 0.15 0.2 0.25 0.3
Initial Density of frugivorous birds in adjacent
forest


Figure 20. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the initial number of
frugivorous birds is increased.

Survival rate of seedlings on dead wood

According to Papic and Armesto (2000), survival of seedlings is 12 times higher on

CWD than on adjacent ground. However, variation in this variable had no impact on the

outcome of the model (Fig. 23). Overall seedling mortality had a minor impact on time










until forest regeneration (Fig. 24). The higher the seedling mortality the longer the time

required until forest regeneration.


300

250

200 -.
200 25%
; 50%
W 150 50%
S* 75%
U. 100%
100 --

50


0 -
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Germination Rate (annual proportion of seeds)


Figure 21. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the germination rate is
increased.

Rate of cover loss to forest

I parameterized the model to allow for the possibility that forest trees modify the

microclimate such that it becomes unsuitable for Baccharis, Sphagnum, or grasses.

However, at high rates of competition with other substrate types, the forest does not grow

fast enough to take advantage of the improved habitat conditions (Figs. 25-27). On the

other hand, if the time required for forest tree species to out-compete cover types such as

Sphagnum and Baccharis is very high (i.e. very low rates of cover loss), it has profound

effects on the model, therefore, this relationship could dictate whether or not old-fields

are in a state of arrested succession. For example, ifBaccharis is lost any faster than

0.008m2 per month, there is no corresponding increase in forest cover because forest









does not grow any faster to take advantage of the increased Baccharis loss (Fig. 25). At a

value of 0.008 m2 for Baccharis reduction, it takes roughly 12 years for a tree to shade

out 1 m2 of adjacent Baccharis. Although these rates of loss have no effect on time until

50% regeneration in the sensitivity analyses (i.e. average levels), if a particular field has a

very large area covered with Baccharis or Sphagnum, changes in competition rates can be

very important.


300


250 .


200
"" 25%
50%


100 -


50


0 -
0.01 0.015 0.02 0.025 0.03 0.035 0.04
Rate of tree canopy growth (square meters /
year)


Figure 22. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of tree growth is
increased or decreased.

Rarity of new recruitment-foci formation

For the model runs I chose an r (rarity of new focus formation) value of 0.1. As

this value was chosen purely to keep the model estimations conservative, it is important









to subject it to sensitivity analyses. Figure 28 demonstrates that this value directly

influences the rate of forest regeneration, especially at lower values.


300

250 a

200 .
S. 25%
M 150 ."" 50%
15 75%
100%
100

50

0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18
Seedling mortality on CWD (proportion of
seedlings / month)

Figure 23. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the mortality of seedlings
on dead wood is increased. Default mortality of seedlings is 0.16% seedlings
/ month.

Management scenarios

High quality fields with high cost input (additions of both CWD and saplings) show

the fastest regeneration times, followed by medium quality fields with high-cost input.

All field types show significant benefits from increasing the cost of input, however low

quality fields require high-cost inputs to achieve any significant regeneration at all (Table

3).

Discussion

Relative Importance of Seed Dispersal and Germination Limitation

When either seed dispersal or suitable germination microsites were set to very low

values (limiting), the outcomes suggest that germination site availability is potentially









more limiting of forest regeneration. At very low rates of seed dispersal forest

regenerates significantly faster with adequate availability of suitable germination sites

than in situations with very few germination sites. My results suggest that during the

initial stages of reclaiming established Baccharis fields, germination sites may be the

more important limiting factor (Fig. 29). My results agree generally with similar studies

in that both dispersal and germination are limiting factors (Costa Rican abandoned

pasture Holl et al, (2000); Puerto Rico Zimmerman et al. (2000). Unlike these

studies, however, my modeling effort allowed me to explore the relative importance of

one factor versus the other over a range of realistic values of starting conditions.


300

250

200 --- --- *
1 M M.* MMa 6 O._ ;."M 25%
150 -- i .iI = 5 -
uE M M M .75%
100%
100

50
e 10 --------------------------



0
0.07 0.12 0.17 0.22

Seedling mortality rate
(proportion of seeds that die / month)

Figure 24. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of seedling
mortality increases.


















200


150 -


S. m. .m- m m m M


100 -


0 0.005 0.01 0.015 0.02 0.025 0.03

Square meters of baccharis cover lost per month per
meter of new forest edge


Figure 25. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of out-competition
of Baccharis is increased.


0 0.002


0.004


0.006


0.008


0.01


25%
* 50%
75%
100%


0.012


Square meters of sphagnum lost per meter of new
forest edge/month

Figure 26. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of out-competition
of Sphagnum is increased.


0.035


No M M
*









Table 3. Field conditions, input, and time required for 50% regeneration under 3 different
management scenarios and three different starting conditions. (HC= High
Cost, MC= Medium Cost, LC = Low Cost, HQ = High Quality, MQ =
Medium Quality, LQ = Low Quality).
High Cost Medium Cost Low Cost
HQ MQ LQ HQ MQ LQ HQ MQ LQ
Sphagnum 0% 5% 14.5% 1% 10% 19.5% 1% 10% 19.5%
(% cover)
Baccharis
0% 10% 25% 1% 15% 30% 1% 15% 30%
(% cover)

Grass
Gra 1% 10% 0% 1% 10% 0% 1% 10% 0%
(% cover)
Dead
Wood 10% 6% 5% 6% 2% 1% 5% 1% 0%
(% cover)
Saplings 25 25 25 10 10 10 0 0 0
Planted/yr
Initial
Initial 100 50 0 100 50 0 100 50 0
trees
Average
time until
no No
50% forest 42 43 65 48 53 103 140
cover regen regen
cover
(years)

I found avian frugivores to be critical for sustaining regeneration of old-fields

throughout the progression of forest accretion to recruitment foci (Fig. 20). In this way,

the model suggests that the most rapid regeneration will not be achieved without

constant, ample seed inputs from birds visiting the sites. As fragmentation, forest loss,

and degradation of remnant forest patches intensifies, inadequate seed rain could become

limiting as frugivore density and activity declines. Currently, however, the high

abundance ofElaenia albiceps (et al. 2005) and the species' apparent tolerance for

disturbance, ensure they are likely to be available at sites I studied (and most disturbed

sites in Chiloe) to provide seed input, as long as there are recruitment foci available to










attract them into fields. In the case of reliable dispersal, the importance of both

germination rate and availability of germination substrates (Figs. 15-17) including coarse

woody debris (Fig. 13), may be relatively higher in determining variations in observed

regeneration rate. But if seed dispersal ceases, then reclamation ofBaccharis fields is

unlikely to occur without intense and costly manipulations (see below).


250


200


150 25%
(a* 50%
N 75%
100 -, 0- -100%


50 -


0
0 ----------------------------
0 0.01 0.02 0.03 0.04 0.05 0.06
Square meters of grass and fern lost per meter of
new forest edge/month

Figure 27. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the rate of out-competition
of grass and fern is increased.

Restoration Methods / Management Scenarios

Assessment of model scenarios

Recommendations for restoring old-fields will vary somewhat according to the

management scenario involved. The three field types modeled represent points along the

continuum from very low quality fields (no trees, a lot of Baccharis and Sphagnum cover,

no CWD), to very high quality fields (many trees remaining, no Baccharis or Sphagnum

cover, good CWD coverage). Very high and very low quality fields require a different

approach to ensure regeneration (Table 3). High quality fields (i.e. fields not fully









cleared or burned, or already undergoing succession) can be essentially left alone, and

regeneration will still occur, although according to the model it will still require roughly

100 years. In other words, high quality fields would represent fields in which succession

is not arrested. On the other hand, in low-quality fields, no regeneration occurs without

significant input, including a high level of sapling planting (Table 3). The model

suggests that the initial starting conditions have a large impact on the outcome of

restoration efforts. Whereas increasing the input from a medium cost scenario to a high

cost scenario has little impact on regeneration times for high and medium quality fields (5

or 10 year improvements), 50% regeneration does not occur within 250 years in poor

quality fields without high cost inputs. Therefore, if a farmer has no resources, and wants

to accelerate regeneration, he needs to avoid burning when clearing for pasture and leave

fruiting tree species scattered throughout the field. Whereas, if a farmer has sufficient

resources, quick regeneration can be obtained in a low quality field if labor and money

are expended to plant trees and drag wood into the field (as long as it is close to forest

with dense bird populations).

Recommendations for efficient reclamation of Baccharis fields

Based on this modeling effort and exploration of scenarios derived from

discussions with landowners regarding their experiences, an effective restoration strategy

would need to concentrate in the following areas in the order that I consider them here.

(1) Site location. Areas selected for forest regeneration should be adjacent to mid-

successional or old growth forest stands in order to ensure sufficient visitation and seed

dispersal by Elania albiceps (Diaz et al. 2005). Seed dispersal into old-fields clearly

helps sustain any attempts at restoration, whether a landowner is relying principally on

manipulations (additions of CWD and plantings) or natural processes (Figs. 19 and 20).









A landowner without resources for restoration, but with a variety of potential sites and

flexibility in choosing which sites can be allowed to go back to forest is not without

recourse. A high quality field (with trees and CWD) near a proper forest edge can be set

aside for forest regeneration and has a high likelihood of becoming a forest without any

manipulation, though even a small number of plantings could help speed the process

considerably (Table 3).


250


200
0

150 25%
S50%
0 pip 75%
100 %,, -- 100%


i 50


0 -- I I I
0 0.2 0.4 0.6 0.8 1 1.2
Proportion of new foci established

Figure 28. Model prediction of the number of years required to achieve 25, 50, 75 and
100% regeneration of degraded old-field habitat as the value of "r" (additional
rarity of new recruitment foci establishment) is increased.

The best time to make the decision regarding placement of a potential regeneration

site is before complete forest clearing occurs. If, for example, only partial clearing and

no burning was done, followed by an observation period of a year; an assessment of a

site's best use could be accomplished while the site was still of high quality (in terms of

regeneration potential). If the physical conditions of the cleared site proved to foster

Baccharis rather than drier land use options (e.g., pasture) then the site should be left to

regenerate to forest while the landowner assessed other sites for non-forest land uses.









This kind of assessment would save a great deal of wasted labor; as landowners we work

with relate that certain Baccharis-dominated sites (apparently, the most mesic ones) can

be cleared and burned repeatedly without successful establishment of pasture or cropland.

After repeated burning and complete clearing, these sites rapidly reach states of stable

Baccharis-Sphagnum cover. My work suggests this occurs because such sites are devoid

of suitable germination substrates. Thereafter, these sites remain in an unproductive

stable state of no, or only marginal, use to humans and wildlife (Darnell and Sieving, in

press, Diaz et al. 2005). Therefore, intensive clearing and burning of such sites, without

timely assessment of future use, leads directly to more forest clearing and the spread of

Baccharis. To my knowledge, there is no way of determining, prior to beginning forest

clearing, what post-clearing site characteristics will be with respect to Baccharis

formation, although some research is beginning to address this idea in other systems

(Chanasyk et al. 2003). Therefore, I put forward the recommendation to partially clear

and then wait to see what the site tendencies appear to be.

(2) Germination sites. Given a proper location close to forest to insure adequate

dispersal, then the next focus should be on providing sufficient germination substrates.

Depending on the number of trees, amount of CWD, and shrub cover of the old-field, this

could entail the addition of CWD and/or the removal of Baccharis and Sphagnum. The

positive impacts of CWD are significant (Papic & Armesto 2000, suggested by model).

However, the degree of benefit obtained from adding dead wood to a site depends on

initial condition of the field. Fields show a slight benefit from an increase in the initial

amount of dead wood, likely because any impact that dead wood might have as sites for

new foci is swamped by positive germination conditions provided by existing recruitment









foci. However, in fields with no trees initially, the impact of CWD as nurse logs is

profound (Fig. 14). For example, with 0 trees, the time required for regeneration of 50%

forest cover decreased from 200 years with no dead wood, to roughly 100 years when 2%

of the field is covered with CWD. The importance of CWD in ecosystem function is no

longer underestimated (Carmona et al. 2002), especially in facilitation of recruitment of

new trees (McGee & Birmingham 1997; Slocum 2000).


250


200 D- Default values


150
150 t D Low bird density (.02
Birds in adjacent forest
S100 per meter of edge)
100
U Low substrate
availability (0 CWD,
50 -- 500 m bare ground)


0
Mean time to Mean time to Mean time to
5% 25% 50%

Figure 29. Time predicted for regeneration of forest (to 5, 25, and 50% forest cover) with
low densities of seed dispersers, compared with time required for regeneration
with very few micro-sites suitable for germination, compared to default values
of each (see text). Error bars represent 1 SD.

(3) Shrub and Sphagnum management. Even though, in some cases, shrubs can act

as facilitators to forest regeneration (Duncan & Chapman 2003; Holl 2002; Li & Wilson

1998; Zahawl & Auspurger 1999), they often have the opposite effect. Shrubs inhibit

seedling growth in Appalachian canopy gaps (Beckage et al. 2000). Similarly, Denslow

et al. (1991) found evidence for inhibition of tree seedling establishment by broad-leaved

understory plants in the tropics. Finally, Hill et al. (1995) found that shrub canopies









along utility rights of way in New York were very resistant to invasion. Therefore, in the

context of restoring Baccharis fields given the presence of both avian seed dispersers and

suitable microsites for germination, the single most important deterrent to forest

succession is the presence of Sphagnum and Baccharis. This is because seeds simply

cannot germinate and survive in direct competition with these two cover types (Table 3,

Figs. 15 and 16). Instead of burning away the shrub and moss layer which can only be

done during the two driest summer months and may destroy CWD and seedlings -

perhaps the most parsimonious approach is to reduce competition with these cover types

at the edges of recruitment foci by physical means (cutting shrubs and trampling at the

perimeter of recruitment foci). Providing a modest sized open buffer of suitable

germination substrate (e.g., bare ground, or modest CWD) around trees and tree clusters

would release one of the main inhibitors of the rate of regeneration suggested by the

model; competitive effects at the edges of recruitment foci (Figs. 15, 16, and 17), and

may be less labor-intensive than manipulations at the scale of whole fields.

(4) Planting trees. Finally, I would suggest actively planting new trees in old-fields.

The model demonstrates that planting just a few trees each year can have a large impact

on forest regeneration (Fig. 30). Reay and Norton (1999) found that without plantings,

restoration in a New Zealand temperate forest proceeded at a much slower pace.

According to the model, an average field (as defined above; see Methods: Sensitivity

Analyses) will require 157 years to reach 50% tree canopy cover. Planting 10 trees per

ha per year will reduce the time to 83 years, as predicted by my model. Further

increasing the rate of tree planting does little to decrease the regeneration rate (Fig. 30),

likely due to limitations on the growth rate of trees, and competition from Baccharis and










Sphagnum. However, as discussed in Chapter 1, there may be a threshold value for the

number of trees required to lure avian seed dispersers into old-fields. Further research is

required before this can be incorporated into the model. Although this model overlooks

potential differences between species, planting trees of a variety of forest species and

paying attention to the facilitative effects of various species may be very important

(Jansen 1997).


250


200


150 25%
50%
75%
100 100%


50 -


0
0 10 20 30 40 50 60
Number of saplings planted per year

Figure 30. Model prediction of the effect of the number of saplings planted annually on
the rate of forest regeneration (to 25, 50, 75, 100% forest cover).

Finally, another type of planting, not addressed by my work, is of the native

bamboo (Chusquea valdiviensis), which is very attractive to native understory birds (Reid

et al. 2004). Large patches of bamboo occur naturally associated with all seral stages of

south-temperate rainforest (Donoso 1996). It would be worthwhile investigating the

influence of bamboo on tree establishment in these fields due to its potential to provide

habitat immediately. However, the benefits of bamboo as a disperser attractant may be

offset by its negative impact on seedling establishment (Donoso & Nyland 2005).









Assessment of the Model

Complexity

While it is far from a complete description of the south temperate rainforest

ecosystem, this model acts as a good starting point for future research and hypothesis

generation. By necessity the model is far simpler than the ecosystem that it represents.

All tree-related data were obtained regarding only one species (Drimys winterii). This

species is not a poor choice, however, because it is one of the most common bird

dispersed colonizers of old-fields (Armesto et al. 2001a) and occurs in all seral stages of

forest in this region (Lusk & Del Pozo 2002). Models of greater complexity are in use

for similar studies. For example, LANDIS a spatially explicit landscape model

(Mladenoff et al. 1996) could incorporate a variety of tree and understory species with

unique demographic characteristics into regeneration scenarios. Additionally, a more

accurate (and much more complex) model might factor in such variables as carbon, other

nutrient and water cycles and energy flows (Kirschbaum 1999). However, by

maintaining simplicity, I was able to focus on parameters relevant to my original research

question concerning the relative importance of seedling establishment and seed dispersal.

Moreover, there is some merit in keeping the model simple and utilizing statistical

methods (e.g., Monte Carlo method) to model the stochastic properties inherent in

complex systems (Young et al. 1996).

Relevant scales

Because most key ecological processes of forest regeneration at edges take place

within 50 meters of the forest, the effects of larger spatial scales were ignored, even

though seed dispersers, for example, operate at larger scales. While I was constrained in

my field approaches (Chapter 1) to work on processes relevant to field-forest ecotones, I









do not think my conclusions concerning restoration of old-fields in general are inaccurate

because forest regeneration in old fields tends to occur through accretion at edges of

recruitment foci and remnant patches (Debussche & Isenmann 1994). In order to

incorporate larger-scale site characteristics, like attenuation in seed dispersal with

distance from forest, further study of avian frugivore movements would have been

required. I deemed this unnecessary because in the landscape where I worked, most

Baccharis fields were not more than 250-300 m wide. Frugivorous birds readily and

quickly crossed these open areas between patches of forest, so it is unlikely that dispersal

was dramatically different throughout the fields. In order to select sites that were

equivalent in proximity to forest patches and other land use types, the plot size I used to

collect data was easy to standardize from a design perspective. Perhaps the greatest

caution in using the model is relevant to the combined effects of the long time scales

involved and the potential that (over such long times) factors originating beyond the

spatial scale of the site could be significantly involved in determining final outcomes

(Parker 1997).

Model improvement

One major assumption of the model is that Baccharis, Sphagnum and grass are

static features of the old-field (i.e. they do not increase in area). In reality they are likely

to increase in cover area when not in direct competition with trees, but including this

dynamic was not feasible. However, this simplification could overlook important

dynamics of the system (Duncan & Chapman 2003; Holl 2002), and is deserving of

further work.

Finally, I suggest that the influence of standing water be incorporated into

modeling efforts, and that greater understanding of the hydrology of these sites is









necessary. Currently, standing water is assumed to be a factor leading to the increased

survival rate of seedlings on CWD. However, an analysis of historic weather trends,

combined with study of germination rates of immersed seeds may better illuminate the

role that the hydrology of the system plays in forest regeneration. Moreover, it also

seems important to establish the validity of a central assumption we make, based in part

on observations, that the water table is affected by forest clearing. We assume that one of

the reasons landowners clear so many areas that become wet and shrub-dominated is that

a high water table is not evident prior to clearing, but that the water table often rises

following clearing (e.g., Sun et al. 2000). This could occur in sites dominated by tree

species with high transpiration rates. Other sites, with lower rates of transpiration, may

be more easily assessed as to post-clearing hydrology. Further knowledge is needed

regarding site characteristics that can predict the outcome of forest clearing (e.g.,

Chanasyk et al. 2003). If arrested succession can be avoided by informed choices before

forest clearing, this would reduce the overall impact on an already fragmented ecosystem.

For fields already in an arrested state, it is hoped that this research and further research

will shed some light on returning old fields to forests that provide benefits to both

humans and wildlife.
















APPENDIX A
ICONOGRAPHIC REPRESENTATION OF MODEL

OLD-FIELD CHARACTERISTICS

L


Ga2 p


Circumference









SEED RAIN


total attractive trees

Dbn


undispersed seeds 2

SEED GERMINATION AND SEEDLING SURVIVAL


total dw in foci




dw proportion


non dead wood proportion


Dseedling


A








73



CREATION OF NEW RECRUITMENT FOCI






DW
random chance of
new foci appearing [dom chance new focl dead wood
seedling bank
growth


aplings

sapling death CWD Growth

seedling bank on nurselogs
G1







TREE GROWTH AND SURVIVAL
sapling





G1


decayed dead wood in focl dead trees S



decay of woo decay D1 G2



decayed 2 dead wood in foci 2 dead trees 2 SC2 BG
dec of wood 2 decay 2


D2

total dw in fo death rate


dead wood in focl 3 SC3
dead trees 3



ecay of wood 3 decay 3 D3 growth modifier






74














APPENDIX B
MODEL EQUATIONS

OLD-FIELD CHARACTERISTICS
L(t) = L(t dt)
INIT L = 0

Oa(t) = Oa(t dt)
INIT Oa = 0

Ca = Fa/Oa

Db = .15625

Fa = IF (total_attractivetrees+. 00000000000001)* 1.68 >Ga THEN Ga ELSE
(total_attractivetrees+.00000000000001)* 1.68

maxlength_of edge = (SQRT(Oa))*4

Bcov(t) = Bcov(t dt) + (- Bloss) dt

INIT Bcov = 0

OUTFLOWS:
Bloss = IF P > 0.999 THEN ARRAYSUM(Fc[*])*(Bloss_rate) ELSE 0

GFcov(t) = GFcov(t dt) + (- GFloss) dt

INIT GFcov = 0

OUTFLOWS:
GFloss = IF P > 0.999 THEN ARRAYSUM(Fc[*])*GFloss_rate ELSE 0

SPcov(t) = SPcov(t dt) + (- SPloss) dt
INIT SPcov = 0

OUTFLOWS:
SPloss = IF P > 0.999 THEN ARRAYSUM(Fc[*])*(SPlossrate) ELSE 0

BG = Oa-(Bcov+SPcov+GFcov+ (Ca*Oa))

B loss rate = 1/288










Ga = IF (Oa-(Bcov+GFcov+SPcov)) <= 0 THEN .000000001 ELSE (Oa-
(Bcov+GFcov+SPcov))

GF loss rate = 1/48

SP loss rate= 1/180

SEED RAIN
seed_pool(t) = seed_pool(t dt) + (PSR ASR[1] ASR[2] ASR[3] ASR[4] ASR[5]
- ASR[6] ASR[7] ASR[8] ASR[9] ASR[10] ASR[11]- ASR[12] ASR[13] -
ASR[14] ASR[15] ASR[16] ASR[17] ASR[18] ASR[19] ASR[20] ASR[21] -
ASR[22] ASR[23] ASR[24] ASR[25] ASR[26] ASR[27] ASR[28] ASR[29] -
ASR[30] ASR[AllFoci] undispersed_seeds) dt
INIT seed_pool = 0

INFLOWS:
PSR = (Sr*Oa)*(L/100)*(Db/. 15625)

OUTFLOWS:
ASR[AllFoci] = (seed_pool*((Af[AllFoci]/Oa)))*Fg[AllFoci]

undispersed_seeds = seed_pool-(ARRAYSUM(ASR[*]))

Sf[AllFoci](t) = Sf[AllFoci](t dt) + (ASR[AllFoci] G[AllFoci] seed_death[AllFoci])
* dt
INIT Sf[AllFoci] = 0

INFLOWS:
ASR[AllFoci] = (seed_pool*((Af[AllFoci]/Oa)))*Fg[AllFoci]

OUTFLOWS:
seed_death[AllFoci] = (Sf[AllFoci]-G[AllFoci])

Af[AllFoci] = trees in foci[AllFoci]*1.68

Fg[AllFoci] = (Af[AllFoci]- (((SQRT(Af[AllFoci]/PI))-
0.5)A2)*PI)/(Af[AllFoci]+. 0000000000000001)

Sr = GRAPH(total_attractivetrees)
(0.00, 1.82), (36.6, 1.82), (73.2, 1.82), (110, 1.79), (146, 1.76), (183, 1.73), (220, 1.70),
(256, 1.67), (293, 1.64), (329, 1.61), (366, 1.58), (402, 1.55), (439, 1.52), (476, 1.49),
(512, 1.46), (549, 1.43), (585, 1.40), (622, 1.37), (659, 1.34), (695, 1.31), (732, 1.28),
(768, 1.25), (805, 1.22), (841, 1.19), (878, 1.16), (915, 1.13), (951, 1.10), (988, 1.07),
(1024, 1.04), (1061, 1.01), (1098, 0.98), (1134, 0.95), (1171, 0.92), (1207, 0.89), (1244,









0.86), (1280, 0.83), (1317, 0.8), (1354, 0.77), (1390, 0.74), (1427, 0.71), (1463, 0.68),
(1500, 0.65)

SEED GERMINATION AND SEEDLING SURVIVAL
seedlings[AllFoci](t) = seedlings[AllFoci](t dt) + (G[AllFoci] transition[AllFoci] -
transition_dw[AllFoci]) dt
INIT seedlings[AllFoci] = 0

INFLOWS:
G[AllFoci] = IF BG < 1 THEN 0 ELSE ((Sf[AllFoci]*Gr))

OUTFLOWS:
transition[AllFoci] = ((seedlings[AllFoci]*non_deadwood_proportion[AllFoci]))

transition_dw[AllFoci] = seedlings[AllFoci]*dw_proportion[AllFoci]

seedling bank[AllFoci](t) = seedling bank[AllFoci](t dt) + (transition[AllFoci] -
growth[AllFoci] Dseedling[AllFoci]) dt
INIT seedling bank[AllFoci] = 0

INFLOWS:
transition[AllFoci] = ((seedlings[AllFoci]*non_deadwood_proportion[AllFoci]))

OUTFLOWS:
growth[AllFoci] = (IF BG < 1 THEN 0 ELSE
seedling bank[AllFoci]/12)+(Dseedling[AllFoci]*0)

Dseedling[AllFoci] = (seedlingmortality)*seedling bank[AllFoci]

seedling bankonnurselogs[AllFoci](t)= seedling bank on nurselogs[AllFoci](t dt)
+ (transition_dw[AllFoci] DCWDseedling[AllFoci] CWD_Growth[AllFoci]) dt
INIT seedling bankonnurselogs[AllFoci] = 0

INFLOWS:
transition_dw[AllFoci] = seedlings[AllFoci]*dw_proportion[AllFoci]

OUTFLOWS:
DCWDseedling[AllFoci] =
((seedlingmortality/CWD_deathrate))*seedlingbankonnurselogs[AllFoci]

CWD death rate = 12

dw_proportion[AllFoci] = totaldwinfoci[AllFoci]*2/(Af[AllFoci]
+.000000000001)


Gr=0










non_dead_wood_proportion[AllFoci] = (1-dw_proportion[AllFoci])

seedlingmortality = .04

CREATION OF NEW RECRUITMENT FOCI
random_chance new foci_deadwood[AllFoci] = (IF Af[AllFoci] = 0 THEN
MONTECARLO((0.103/12)*DW*0.23*((L/100)*(Db/0.15625)*r)) ELSE 0)

INFLOW TO: sapling (IN SECTOR: Tree Growth and Survival)
randomchance of newfoci_appearing[AllFoci] = (IF Af[AllFoci] = 0 THEN
MONTECARLO((0.063/12)*BG*0.02*((L/100)*(Db/0.15625)*r)) ELSE 0)

INFLOW TO: sapling (IN SECTOR: Tree Growth and Survival)
DW = 0
r= .

TREE GROWTH AND SURVIVAL
deadtrees[AllFoci](t) = deadtrees[AllFoci](t dt) + (D1[AllFoci] decay[AllFoci]) dt
INIT deadtrees[AllFoci] = 0

INFLOWS:
D [AllFoci] = deathmodifier*.89*SC1[AllFoci]

OUTFLOWS:
decay[AllFoci] = deadtrees[AllFoci]/120

deadtrees_2[AllFoci](t) = deadtrees_2[AllFoci](t dt) + (D2[AllFoci] -
decay_2[AllFoci]) dt
INIT deadtrees_2[AllFoci] = 0

INFLOWS:
D2[AllFoci] = deathmodifier*.94* SC2[AllFoci]

OUTFLOWS:
decay_2[AllFoci] = deadtrees_2[AllFoci]/120

deadtrees_3 [AllFoci](t) = deadtrees_3[AllFoci](t dt) + (D3[AllFoci] -
decay_3[AllFoci]) dt
INIT deadtrees_3[AllFoci] = 0

INFLOWS:
D3[AllFoci] = deathmodifier*1.06* SC3[AllFoci]

OUTFLOWS:
decay_3[AllFoci] = dead trees [AllFoci]/120









deadtrees_4[AllFoci](t) = deadtrees_4[AllFoci](t dt) + (D4[AllFoci] -
decay_4[AllFoci]) dt
INIT deadtrees_4[AllFoci] = 0

INFLOWS:
D4[AllFoci]= (deathmodifier)* 1.11 *SC4[AllFoci]

OUTFLOWS:
decay_4[AllFoci] = deadtrees_4[AllFoci]/120

deadwood in foci[AllFoci](t)= deadwood in foci[AllFoci](t dt) + (decay[AllFoci] -
decay_of wood[AllFoci]) dt
INIT deadwood in foci[AllFoci] = 0

INFLOWS:
decay[AllFoci] = deadtrees[AllFoci]/120

OUTFLOWS:
decay_of wood[AllFoci] = deadwood in foci[AllFoci]/120
deadwood in foci_2[AllFoci](t) = deadwood in foci_2[AllFoci](t dt) +
(decay_2[AllFoci] decay_of wood_2[AllFoci]) dt
INIT deadwood in foci_2[AllFoci] = 0

INFLOWS:
decay_2[AllFoci] = deadtrees_2[AllFoci]/120

OUTFLOWS:
decay_of wood_2[AllFoci] = deadwood in foci_2[AllFoci]/120

deadwood in foci_3[AllFoci](t) = deadwood in foci_3[AllFoci](t dt) +
(decay_3[AllFoci] decay_of wood_3 [AllFoci]) dt
INIT deadwood in foci_3[AllFoci] = 0

INFLOWS:
decay_3[AllFoci] = deadtrees_3[AllFoci]/120

OUTFLOWS:
decay_of wood_3[AllFoci] = deadwood in foci_3[AllFoci]/120
deadwood in foci_4[AllFoci](t) = dead wood in foci_4[AllFoci](t dt) +
(decay_4[AllFoci] decay_of wood_4[AllFoci]) dt
INIT deadwood in foci_4[AllFoci] = 0

INFLOWS:
decay_4[AllFoci] = deadtrees_4[AllFoci]/120









OUTFLOWS:
decay_of wood_4[AllFoci] = deadwood in foci_4[AllFoci]/120

decayed[AllFoci](t) = decayed[AllFoci](t dt) + (decay_of wood[AllFoci]) dt
INIT decayed[AllFoci] = 0

INFLOWS:
decay_of wood[AllFoci] = deadwood in foci[AllFoci]/120

decayed_2[AllFoci](t) = decayed_2[AllFoci](t dt) + (decay_of wood_2[AllFoci]) dt
INIT decayed_2[AllFoci] = 0

INFLOWS:
decay_of wood_2[AllFoci] = deadwood in foci_2[AllFoci]/120

sapling[AllFoci](t) = sapling[AllFoci](t dt) + (growth[AllFoci] + input[AllFoci] +
CWDGrowth[AllFoci] + randomchance newfocidead_wood[AllFoci] +
random_chance of newfoci_appearing[AllFoci] Gl[AllFoci] -
sapling_death[AllFoci]) dt
INIT sapling[AllFoci] = 0


INFLOWS:
growth[AllFoci]
input[l] = annual
input[2] = annual
input[3] = annual
input[4] = annual
input[5] = annual
input[6] = annual
input[7] = annual
input[8] = annual
input[9] = annual


input[10]
input[ll]
input[12]
input[13]
input[14]
input[15]
input[16]
input[17]
input[18]
input[19]
input[20]
input[21]
input[22]
input[23]


annual
annual
annual
annual
annual
annual
annual
annual
annual
annual
annual
annual
annual
annual


(IN SECTOR: Seed Germination and Seedling Survival)
sapling_addition/180
sapling_addition/180
sapling_addition/180
sapling_addition/180
sapling_addition/180
sapling_addition/180
sapling_addition/180
sapling_addition/180
sapling_addition/1 80


sapling_
sapling_
sapling_
sapling_
sapling_
sapling_
sapling_
sapling_
sapling_
sapling_
sapling_
sapling_
sapling_
sapling


addition/180
addition/180
addition/180
addition/180
addition/180
addition/180
addition*0
addition*0
addition*0
addition*0
addition*0
addition*0
addition*0
addition*0









input[24] = annual_sapling_addition*0
input[25] = annual_sapling_addition*0
input[26] = annual_sapling_addition*0
input[27] = annual_sapling_addition*0
input[28] = annual_sapling_addition*0
input[29] = annual_sapling_addition*0
input[30] = annual_sapling_addition*0

CWD_Growth[AllFoci] (Not in a sector)

random_chancenewfoci_deadwood[AllFoci] (IN SECTOR: Chance of new foci
appearing)

random_chance of newfoci_appearing[AllFoci] (IN SECTOR: Chance of new foci
appearing)

OUTFLOWS:
G1[AllFoci] = (IF BG < 1 THEN 0 ELSE sapling[AllFoci]/292.75)*growthmodifier

sapling_death[AllFoci] = .0083*sapling[AllFoci]

SC1[AllFoci](t)= SC1[AllFoci](t dt) + (G1[AllFoci] G2[AllFoci] D1[AllFoci]) dt
INIT SC1[AllFoci] = 0

INFLOWS:
G1[AllFoci] = (IF BG < 1 THEN 0 ELSE sapling[AllFoci]/292.75)*growthmodifier

OUTFLOWS:
G2[AllFoci] = (IF BG < 1 THEN 0 ELSE SC1[AllFoci]/342.75)*growthmodifier

D [AllFoci] = deathmodifier*.89*SC1[AllFoci]

SC2[AllFoci](t) = SC2[AllFoci](t dt) + (G2[AllFoci] G3[AllFoci] D2[AllFoci]) dt
INIT SC2[AllFoci] = 0

INFLOWS:
G2[AllFoci] = (IF BG < 1 THEN 0 ELSE SC1[AllFoci]/342.75)*growthmodifier

OUTFLOWS:
G3[AllFoci] = (IF BG < 1 THEN 0 ELSE SC2[AllFoci]/687.5)*growthmodifier

D2[AllFoci] = deathmodifier*.94* SC2[AllFoci]

SC3[AllFoci](t) = SC3[AllFoci](t dt) + (G3 [AllFoci] G4[AllFoci] D3[AllFoci]) dt
INIT SC3[AllFoci] = 0









INFLOWS:
G3[AllFoci] = (IF BG < 1 THEN 0 ELSE SC2[AllFoci]/687.5)*growthmodifier

OUTFLOWS:
G4[AllFoci] = (IF BG < 1 THEN 0 ELSE SC3[AllFoci]/1075)*growth_modifier

D3[AllFoci] = deathmodifier*1.06* SC3 [AllFoci]

SC4[AllFoci](t) = SC4[AllFoci](t dt) + (G4[AllFoci] D4[AllFoci]) dt
INIT SC4[AllFoci] = 0

INFLOWS:
G4[AllFoci] = (IF BG < 1 THEN 0 ELSE SC3[AllFoci]/1075)*growth_modifier

OUTFLOWS:
D4[AllFoci]= (deathmodifier)* 1.11 *SC4[AllFoci]

annual_sapling_addition = 0

death modifier =0.00181

number of foci = (IF trees in foci[l] > 0 THEN 1 ELSE 0) + (IF trees in foci[2] > 0
THEN 1 ELSE 0) + (IF trees_in foci[3] > 0 THEN 1 ELSE 0) + (IF treesin foci[4] > 0
THEN 1 ELSE 0) + (IF trees_in foci[5] > 0 THEN 1 ELSE 0) + (IF trees in foci[6] > 0
THEN 1 ELSE 0) + (IF trees_in foci[7] > 0 THEN 1 ELSE 0) + (IF trees in foci[8] > 0
THEN 1 ELSE 0) + (IF trees_in foci[9] > 0 THEN 1 ELSE 0) + (IF trees in foci[10] > 0
THEN 1 ELSE 0) + (IF trees in foci[ 1] > 0 THEN 1 ELSE 0) + (IF trees in foci[12] >
0 THEN 1 ELSE 0) + (IF trees in foci[13] > 0 THEN 1 ELSE 0) + (IF trees in foci[14]
> 0 THEN 1 ELSE 0) + (IF trees in foci[15] > 0 THEN 1 ELSE 0) + (IF
trees in foci[16] > 0 THEN 1 ELSE 0) + (IF trees in foci[17] > 0 THEN 1 ELSE 0) +
(IF trees infoci[18] > 0 THEN 1 ELSE 0) + (IF trees in foci[19] > 0 THEN 1 ELSE 0)
+ (IF treesinfoci[20] > 0 THEN 1 ELSE 0) + (IF trees in foci[21] > 0 THEN 1 ELSE
0) + (IF trees in foci[22] > 0 THEN 1 ELSE 0) + (IF treesinfoci[23] > 0 THEN 1
ELSE 0) + (IF trees in foci[24] > 0 THEN 1 ELSE 0) + (IF trees in foci[25] > 0 THEN
1 ELSE 0) + (IF trees in foci[26] > 0 THEN 1 ELSE 0) + (IF trees in foci[27] > 0
THEN 1 ELSE 0) + (IF trees in foci[28] > 0 THEN 1 ELSE 0) + (IF trees in foci[29] >
0 THEN 1 ELSE 0) + (IF trees in foci[30] > 0 THEN 1 ELSE 0)

total dw in foci[AllFoci] =
((deadwood in foci[AllFoci]*.5)+deadwood infoci_2[AllFoci]+(dead_wood infoci
_3[AllFoci]*2)+(deadwood in foci_4[AllFoci]*3))

trees in foci[AllFoci] =
((deadtrees[AllFoci]+dead_wood infoci[AllFoci]+SC 1[AllFoci]+decayed[AllFoci])*.
5)+((deadtrees_2[AllFoci]+deadwood in foci_2[AllFoci]+decayed_2[AllFoci]+SC2[
AllFoci]))+((dead trees_3 [AllFoci]+dead wood in foci_3 [AllFoci]+decayed_3 [AllFoci









]+SC3[AllFoci])*2)+((deadtrees_4[AllFoci]+deadwood infoci_4[AllFoci]+decayed_
4[AllFoci]+SC4[AllFoci])*3)

Not in a sector
decayed_3 [AllFoci](t) = decayed_3 [AllFoci](t dt) + (decay_of wood_3 [AllFoci]) dt
INIT decayed_3[AllFoci] = 0

INFLOWS:
decay_of wood_3 [AllFoci] (IN SECTOR: Tree Growth and Survival)

decayed_4[AllFoci](t) = decayed_4[AllFoci](t dt) + (decay_of wood_4[AllFoci]) dt
INIT decayed_4[AllFoci] = 0

INFLOWS:
decay_of wood_4[AllFoci] (IN SECTOR: Tree Growth and Survival)

CWD_Growth[AllFoci] = (IF BG < 1 THEN 0 ELSE
seedling bankonnurselogs[AllFoci]/12)+(DCWDseedling[AllFoci]*0)

OUTFLOW FROM: seedlingbank on nurselogs (IN SECTOR: Seed Germination
and Seedling Survival)

INFLOW TO: sapling (IN SECTOR: Tree Growth and Survival)
Fc[AllFoci] = (2*PI)*(SQRT(Af[AllFoci]/PI)

growthmodifier = 1

P = (Fa/Ga)

total_attractivetrees = (ARRAYSUM(SC1[*])*.5)+ ARRAYSUM(SC2[*]) +
(ARRAYSUM(SC3[*])*2) + (ARRAYSUM(SC4[*])*3) +
(ARRAYSUM(deadtrees[*])*.5) + ARRAYSUM(deadtrees_2[*]) +
(ARRAYSUM(deadtrees_3 [*])*6) + (ARRAYSUM(deadtrees_4[*])*9) +
(ARRAYSUM(deadwood in foci[*])*.5) + ARRAYSUM(deadwood infoci_2[*]) +
(ARRAYSUM(deadwood in foci_3 [*])*2) +
(ARRAYSUM(deadwood in foci 4[*])*3) + (ARRAYSUM(decayed[*])*.5) +
ARRAYSUM(decayed_2[*]) + (ARRAYSUM(decayed_3 [*])*2) +
(ARRAYSUM(decayed_4[*])*3)


year = TIME/12















LIST OF REFERENCES


Alcantara, J. M., P. J. Rey, F. Valera, and A. M. Sanchez-Lafuente. 2000. Factors
shaping the seedfall pattern of a bird dispersed plant. Ecology 81:1937-1950.

Antos, J. A., R. Parish, and K. Conley. 2000. Age structure and growth of the tree-
seedling bank in subalpine spruce-fir forests of south-central British Columbia.
American Midland Naturalist 143:342-354.

Armesto, J., J. C. Aravena, C. Villagran, C. Perez, and G. Parker. 1996. Bosques
templados de la cordillera de la costa. In J. Armesto, C. Villagran, and M. Kalin,
editors. Ecologia de los Bosques Nativos de Chile. Editorial Universitaria Santiago,
Santiago.

Armesto, J. J., I. Diaz, C. Papic, and M. F. Willson. 2001a. Seed rain of fleshy and dry
propagules in different habitats in the temperate rainforests of Chiloe Island,
Chiloe. Austral Ecology 26:311-320.

Armesto, J. J., and J. Figueroa. 1987. Stand structure and dynamics in the temperate rain
forests of Chiloe Archipelago, Chile. Journal of Biogeography 14:367-376.

Armesto, J. J., and R. Rozzi. 1989. Seed dispersal syndromes in the rain forest of Chiloe:
evidence for the importance of biotic dispersal in a temperate rain forest. Journal of
Biogeography 16:219-226.

Armesto, J. J., R. Rozzi, C. Smith-Ramirez, and M. T. K. Arroyo. 1998. Conservation
targets in South American temperate forests. Science 282:1271-1272.

Armesto, J. J., C. Smith-Ramirez, and R. Rozzi. 2001b. Conservation strategies for
biodiversity and indigenous people in Chilean forest ecosystems. Journal of the
Royal Scoiety of New Zealand 31:865-877.

Beckage, B., J. S. Clark, B. D. Clinton, and B. L. Haines. 2000. A long-term study of tree
seedling recruitment in southern Appalachian forests: the effects of canopy gaps
and shrub understories. Canadian Journal of Forest Research 30:1617-1631.

Bell, S. S., M. S. Fonseca, and L. B. Motten. 1997. Linking restoration and landscape
ecology. Restoration Ecology 5:318-323.

Bewley, J. D., and M. Black 1982. Physiology and Biochemistry of Seeds in Relation to
Germination Viability, Dormancy, and Environmental Control. Springer- Verlag,
NewYork.









Bibby, C. J., N. D. Burgess, D. A. Hill, and S. H. Mustoe 2000. Bird Census Techniques.
Academic Press, London.

Brewer, J. S. 2002. Disturbances increase seedling emergence of an invasive native shrub
in pitcher-plant bogs. Natural Areas Journal 22:4-10.

Brown, S., and A. E. Lugo. 1994. Rehabilitation of tropical lands: a key to sustaining
development. Restoration Ecology 2:97-111.

Bustamante, R. O., and J. J. Armesto. 1995. Regeneration dynamics in canopy gaps of a
montane forest of Chiloe Island, Chile. Revista Chilena de Historia Natural 68:391-
398.

Cardoso da Silva, J. M., C. Uhl, and G. Murray. 1996. Plant succession, landscape
management, and the ecology of frugivorous birds in abandoned amazonian
pastures. Conservation Biology 10:491-503.

Carmona, M. R., J. J. Armesto, J. C. Aravena, and C. A. Perez. 2002. Coarse woody
debris biomass in successional and primary temperate forests in Chiloe Island,
Chile. Forest Ecology and Management 164:265-275.

Cespedes, C. L., A. Uchoa, J. R. Salazar, F. Perich, and F. Pardo. 2002. Plant growth
inhibitory activity of p-hydroxyacetophenones and tremetones from Chilean
endemic Baccharis species and some analogous: A comparative study. Journal of
Agricultural and Food Chemistry 50:2283-2292.

Charlesdominique, P. 1995. Plants-Frugivorous Animals Interactions Consequences on
seed dispersal and forest regeneration. Revue D Ecologie-La Terre Et La Vie
50:223-235

Chanasyk, D.S., I.R. Whitson, E. Mapfumo, J.M. Burke, and E.E. Prepas. 2003. The
impacts of forest harvest and wildfire on soils and hydrology in temperate forests:
A baseline to develop hypotheses for the Boreal Plain. Journal of Environmental
Engineering and Science 2:S51-S63.

Christie, D. A., and J. Armesto. 2003. Regeneration microsites and tree species
coexistence in temperate rain forests of Chiloe Island, Chile. Journal of Ecology
91:776-784.

Connell, J. H. 1971. On the role of natural enemies in preventing competitive exclusion
in some marine mammals and in rain forest trees. Pages 298-312 in P. J. den Boer
and G. Gradwell, editors. Dynamics of Populations. Centre for Agricultural
Publishing and Documentation (PUDOC), Wageningen, The Netherlands.

Connell, J. H., and R. O. Slatyer. 1977. Mechanisms of succession in natural
communities and their role in community stability and organization. The American
Naturalist 111:1119-1144.









Darnell, T. M. and K. E. Sieving. In Press. Landscape connectivity is not defined by
corridors alone: A translocation experiment with an endemic understory bird in
south-temperate rainforest. Conservation Biology.

Debussche, M., and P. Isenmann. 1994. Bird-dispersed seed rain and seedling
establishment in patchy Mediterranean vegetation. Oikos 69:414-426.

Denslow, J. S., E. Newell, and A. Ellison. 1991. The effect of understory palms and
cyclanths on the growth and survival oflnga seedlings. Biotropica 23:225-234.

Diaz, I., J. Armesto, S. Reid, K. E. Sieving, and M. Willson. 2005. Linking forest
structure and composition: avian diversity in successional forests of Chiloe' Island,
Chile. Biological Conservation 123:91-101.

Diaz, I., C. Papic, and J. J. Armesto. 1999. An assessment of post-dispersal seed
predation in temperate rain forest fragments in Chiloe Island, Chile. Oikos 87:228-
238.

Donoso, 1996. Ecology of Nothofagus Forests in Central Chile. Pages 271-292 In
Thomas T. Veblen, Robert S. Hill, Jennifer Read editors. The Ecology and
Biogeography of Nothofagus Forests, New Haven: Yale University Press

Donoso, P. J., and R. D. Nyland. 2005. Seedling density according to structure,
dominance and understory cover in old-growth forest stands of the evergreen forest
type in the coastal range of Chile. Revista Chilena de Historia Natural 78:51-63.

Duncan, R. S., and C. A. Chapman. 2003. Tree-Shrub Interactions During Early
SecondaryForest Succession in Uganda. Restoration Ecology 11:198.

Ferguson, R. N., and D. R. Drake. 1999. Influence of vegetation structure on spatial
patterns of seed deposition by birds. New Zealand Journal of Botany 37:671-677.

Ferro, A., M. Gefell, R. Kjelgren, D. S. Lipson, N. Zollinger, and S. Jackson 2003.
Maintaining hydraulic control using deep rooted tree systems. Advances in
Biochemical Engineering/Biotechnology 78:125-156.

Figueroa, J., J. J. Armesto, and J. F. Hernandez. 1996. Seed germination and dormancy
strategies of temperate rain forest species in Chiloe, Chile. Revista Chilena De
Historia Natural 69:243-251.

Figueroa, J., and C. H. Lusk. 2001. Germination requirements and seedling shade
tolerance are not correlated in a Chilean temperate rain forest. New Phytologist
152:483-489.

Figueroa, J. A. 2003. Seed germination in temperate rain forest species of southern Chile:
chilling and gap-dependency germination. Plant Ecology 166:227-240.









Fjeldsa, J., and N. Krabbe 1990. Birds of the High Andes. Zoological Museum,
Svendborg, Denmark.

Gude, J. 2000. Avian community structure in two oldfield types on Isla Grande de
Chiloe, Chile, and the potential for avian aid in forest restoration. Wildlife Ecology
and Conservation. University of Florida, Gainesville.

Gunnarsson, U., and H. Rydin. 1998. Demography and recruitment of Scots pine on
raised bogs in eastern Sweden and relationships to microhabitat differentiation.
Wetlands 18:133-141.

Hallgren, W. S., and J. Pitman. 2000. The uncertainty in simulations by a Global Biome
Model (BIOME3) to alternative parameter values. Global Change Biology 6:483-
495.

Halpern, B. S., S. D. Gaines, and R. R. Warner. 2005. Habitat size, recruitment, and
longevity as factors limiting population size in stage-structured species. American
Naturalist 165:82-94.

Hanley, M. E. 1998. Seedling herbivory, community composition and plant life history
traits. Perspectives in Plant Ecology, Evolution and Systematics 1:191-205.

Harmon, M. E., J. F. Franklin, F. J. Swanson, P. Sollins, S. V. Gregory, J. D. Lattin, N.
H. Anderson, S. P. Cline, N. G. Aumen, J. R. Sedell, G. W. Lienkaemper, K.
Cromack, and K. W. Cummins. 1986. Ecology of coarse woody debris in temperate
ecosystems. Advances in Ecological Research 15:133-302.

Harvey, C. 2000. Windbreaks enhance seed dispersal into agricultural landscapes in
Monteverde, Costa Rica. Ecological Applications 10:155-173.

Hill, J. D., C. D. Canham, and D. M. Wood. 1995. Patterns and causes of resistance to
tree invasion in rights-of-way. Ecological Applications 5:495.

Hobbs, R. J., and D. R. Norton. 1996. Towards a conceptual framework for restoration
ecology. Restoration Ecology 4:93-110.

Hofgaard, A. 1993. Structure and regeneration patterns in a virgin Picea-Abies forest in
northern Sweeden. Journal of Vegetation Science 4:601-608.

Holl, K. D. 1998. Do bird perching structures elevate seed rain and seedling
establishment in abandoned tropical pasture? Restoration Ecology 6:253-261.

Holl, K. D. 1999. Factors limiting tropical rain forest regeneration in abandoned pasture:
seed rain, seed germination, microclimate, and soil. Biotropica 31:229-242.

Holl, K. D. 2002. Effect of shrubs on tree seedling establishment in an abandoned
tropical pasture. Journal of Ecology 90:179-187.






88


Holl, K. D., M. E. Loik, E. H. Lin, and I. A. Samuels. 2000. Tropical montane forest
restoration in Costa Rica: overcoming barriers to dispersal and establishment.
Restoration Ecology 8:339-349.

Hostetler, M. 1999. Scale, birds, and human dimensions: a potential approach for
integrative research in urban ecosystems. Landscape and Urban Planning 45(1),15-
19.

Houle, G. 1992. Spatial relationship between seed and seedling abundance and mortality
in a deciduous forest of north-eastern North-America. Journal of Ecology 80:99-
108.

Houle, G. 1995. Seed dispersal and seedling recruitment the missing link(s). Ecoscience
2:238-244.

Howe, H. F., and J. Smallwood. 1982. Ecology of seed dispersal. Annual Review of
Ecology and Systematics 13:201-228.

Howe, H. F., and M. N. Mirti. 2004. When seed dispersal matters. BioScience 54:651-
660.

Hubbell, S. P. 1979. Tree dispersion, abundance, and diversity in a tropical dry forest.
Science 203:1299-1309.

Hubbell, S. P. 1980. Seed predation and the coexistence of tree species in tropical forests.
Oikos 35:214-229.

Hulme, P. E. 2002. Seed-eaters: seed dispersal, destruction and demography. Pages 257-
273 in D. J. Levey, W. R. Silva, and M. Galetti, editors. Seed Dispersal and
Frugivory : Ecology, Evolution, and Conservation. CABI Publishing, New York.

Jansen, A. 1997. Terrestrial invertebrate community structure as an indicator of the
success of a tropical rainforest restoration project. Restoration Ecology 5:115-124.

Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests.
American Naturalist 104:501-528.

Kirkman, L. K., K. L. Coffey, R. J. Mitchell, and E. B. Moser. 2004. Ground cover
recovery patterns and life-history traits: implications for restoration obstacles and
opportunities in a species-rich savanna. Journal of Ecology 92:409-421.

Kirschbaum, M. U. F. 1999. CenW, a forest growth model with linked carbon, energy,
nutrient and water cycles. Ecological Modelling 118:17-59.

Kollman, J., and M. Pirl. 1995. Spatial pattern of seed rain of fleshy-fruited plants in a
scrubland-grassland transition. Acta Aecologica 16:313-329.