The population and feeding ecology of tortoises and feral burros on Volcan Alcedo, Galapagos Islands


Material Information

The population and feeding ecology of tortoises and feral burros on Volcan Alcedo, Galapagos Islands
Physical Description:
xii, 150 leaves : ill., map ; 28 cm.
Fowler, Lynn Elizabeth, 1953-
Publication Date:


Subjects / Keywords:
Turtles -- Galapagos Islands   ( lcsh )
Galapagos tortoise   ( lcsh )
Donkeys -- Galapagos Islands   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1983.
Includes bibliographical references (leaves 142-149).
Statement of Responsibility:
by Lynn Elizabeth Fowler.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000427103
notis - ACH5845
oclc - 11082729
System ID:

Full Text








I would like to thank the scientists and personnel

of the Charles Darwin Research Station and Parque

Nacional Galapagos for providing advice and logistical

support during the course of my field research. Special

thanks are extended to former CDRS director Hendrik Hoeck

and SPNG Superintendent Miguel Cifuentes. I thank

Ulrike Eberhardt and Henk van der Werff for their help

with botanical identifications, and Marsha Cox at the

Smithsonian Institution for straightening out my confused

accounts. Without the generous financial assistance of

friends and folks that I met while in the Islands, this

study would not have been possible. I thank these con-

tributors; Mrs. Vera Stangl in particular receives my

gratitude for her support.

I am grateful for the friendship of the DeRoy and

Moore families who provided me with a home in the

Islands, and of the captains and mariners who trans-

ported me to and from Isabela Island and brought me my

mail. I want to express my deepest appreciation to

assistants John Roe and my sister, Ada Fowler, who pro-

vided invaluable help and moral support in the field.

Thanks go also to Dr. Patti Moehlman for her advice in

the field and afterwards.

I have greatly appreciated the guidance and encour-

agement that my chairman, Dr. John H. Kaufmann, has given

during my field work and since my return to the States.

Thanks are also extended to my committee members, Dr.

Carmine Lanciani and Dr. George Tanner for their advice

and review of this manuscript. Paloma Ibarra did my


Dr. Mark K. Johnson, Louisiana State University,

kindly aided me with diet quantification and Gary

Matson's lab in Montana analyzed the burro teeth I col-

lected. Dr. Frank Martin and Dr. John Cornell helped

with statistics.

Finally, I would like to express my deepest thanks

to my mother, Margaret Fogg, Eduardo and Ella Neira and

my family for their patience, loving support and con-

tinuing confidence throughout my student years.

This research was supported, in part, by funds from

Sigma Xi, the National Geographic Society and the

Explorer's Club.





LIST OF TABLES .. .. ... vi

LIST OF FIGURES . ........ .. ix


ONE INTRODUCTION ............... 1


Volcan Alcedo, Isabela Island 5
Climate . .. .. 10
Vegetation ........ .. 18


Methods . 23
Rim Census and Count Results .. 28
Burro Birth Season and Group
Size/Composition Results 47
Discussion . .. 59


Methods ..... .. 69
Results ..... 71
Discussion ...... .. 75


Methods . .. .. 83
Results . .. 85
Discussion . 92




Methods .
Results .
Discussion .


Methods .
Results .
Discussion .





. 101

. 103
. 106
. 119

. 122

. 123
S. .. ... 125
. 130

. 135

. 142

. 150



2-1 Monthly Rainfall . ... 11

2-2 Daily Air Temperatures
Isabela Island . 12

2-3 Monthly Frequency of Garua Days
Rim Camp, Alcedo . 15

2-4 Average Daily Garua Catch Measurements 17

3-1 Burro and Tortoise Censuses and Counts 27

3-2 Average Number of Burros on Rim Censuses
by Month and Area . .. 31

3-3 Average Number of Tortoises on Rim Censuses
by Month and Area . .. .31

3-4 Rim Camp Censuses . 37

3-5 Southeast Slope Censuses . 38

3-6 South Floor Censuses . 40

3-7 Sulfur Slope Counts . 40

3-8 Midcamp Censuses . ... .46

3-9 North Plateau Counts . 46

3-10 Seasonal Distribution of Young and
Pregnant Burros . 50

3-11 Percent of Burro Groups of Different
Sizes by Season .. .. 51

3-12 Percent of Burro Groups of Different
Sizes by Area . ... 52


4-1 Burro Sex and Age at Death .

4-2 Young Burro Age at Death .

4-3 Years Since Death Based on Weathering
of Burro Bones . .

5-1 Success of South Caldera Floor Nests .

5-2 Success of North Caldera Floor
1979/1980 Nests . .

5-3 Fates of South Caldera Floor Nests .

5-4 Fates of Eggs in Burro Damaged Nests
South Caldera Floor . .

5-5 Fates of North Caldera Floor Nests .

5-6 Success of Undisturbed Nests
South and North Floor . .

5-7 A Comparison of Fertility, Hatching
and Emergence Success of Undisturbed
Natural Nests of Geochelone elephantopus
porteri, ephippium and vandenburghi .

6-1 Numbers of Burro and Tortoise Feeding
Plots Examined . .

6-2 Numbers of Burro and Tortoise Fecal
Samples Collected . .

6-3 Plant Species Eaten by Burros
in Feeding Plots ... .

6-4 Common Plant Species Eaten by Tortoises
in Feeding Plots . .

6-5 Uncommon Plant Species Eaten by
Tortoises in Feeding Plots .

6-6 Plant Species Eaten by Burros and
Tortoises, Volcan Alcedo .

6-7 Percent Relative Density of Plant
Fragments in Burro and Tortoise Feces,
Volcan Alcedo . .


. 105

. 105

. 107

. 109

. 110

. 111

. 115



7-1 Summary of Tortoise Activity . 126

7-2 Seasonal Comparison of Daily Time
Budgets of Alcedo Tortoises . 128





2-1 Galapagos Islands . 7

2-2 Volcan Alcedo Study Sites .. .. 9

3-1 Around-the-Rim Censuses of Burros
and Tortoises . 29

3-2 Burros on Around-the-Rim, South Floor
and Midcamp Censuses . ... 33

3-3 Tortoises on Around-the-Rim and South
Floor Censuses . .. 34

3-4 Distribution of Small Tortoises
on Alcedo . . 42

3-5 Distribution of Medium Tortoises
on Alcedo . . 43

3-6 Distribution of Large Tortoises
on Alcedo .... .. 44

3-7 Reproductive Periodicity of Burros 48

3-8 Seasonal Changes in Burro Group
Composition . .. 54

3-9 Burro Group Composition on Different
Areas . .. . 57

4-1 Months in Which Burros Died ... 73

Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy




April 1983

Chairman: John H. Kaufmann
Major Department: Zoology

Feral burros (Equus asinus) were introduced to the

Galapagos Archipelago in the 1830s. Volcan Alcedo,

Isabela Island, harbours 500 to 700 burros in addition to

the largest remaining population of Galapagos tortoises,

Geochelone elephantopus vandenburghi. Burro and tortoise

population and feeding ecologies were studied on Alcedo

to investigate the possible impact of burros on


There is no permanent source of fresh water on

Alcedo; during the wet season (January to June) rain

water collected in pools and was readily available. Peak

burro natality coincided with the rainy season, as did

tortoise breeding.

During the dry season water was occasionally avail-

able in drip-puddles along the southeastern crater rim.

Temporary water availability influenced distributions and

behavior of burros and tortoises. Both species showed

a tendency to congregate along the moist southeastern

section in the dry season.

Apparently water shortages result in an unusually

high level of mortality among young, sexually mature

burros. Forty percent of 136 burro carcasses and skele-

tons were of animals between three and six years old.

Water shortages probably limit burro population growth.

Burro and tortoise diets were investigated using

direct observation of feeding animals and fecal analysis

techniques. Seventy-two percent of 92 plant species con-

sumed by burros and/or tortoises were eaten by both

animals. Burro and tortoise wet season and early dry

season diets were different, but in the late dry season

both animals consumed Sida and competition for food is a


An investigation of seasonal tortoise feeding

behavior demonstrated, however, that in late dry season

months tortoises spent little time feeding. Late in the

dry season feeding occupied only nine percent of tortoise

daily activity time. In wet and early dry season months

tortoises fed during 40 percent of their active hours.

Even during the dry season competition for food may be

insignificant because tortoises scarcely feed.

Burros trampled some tortoise nests; eighteen

percent of 88 monitored nests were damaged by burros.

Entire clutches were destroyed in 4.5 percent of the

nests. Natural emergence success for Geochelone

elephantopus vandenburghi was 64.9 percent.




About twelve thousand years ago man began to domes-

ticate selected animal species. As he spread across the

globe, he took his domesticated animals with him. In

time, domestic animals escaped or were abandoned. Some

successfully established feral breeding populations in

their new homelands.

The problems created by feral mammals are diverse

and widespread. Island ecosystems are particularly frag-

ile and vulnerable to the ecological disturbances that

are created when domestic animals become feral. Goats,

cattle, pigs, and sheep are among the destructive, large

herbivores that have been widely introduced onto islands

across the world and have been the subject of much

research (on goats, Yocom [1967], Sykes [1969], Williams

and Rudge [1969], Spatz and Mueller-Dombois [1973, 1975a],

Coblentz [1976], Bullock [1977], Rudge and Campbell [1977],

Wardel et al. [1978]; on sheep and cattle, Wilson and

Orwin [1964], Taylor [1971]; on pigs, Taylor [1971],

Spatz and Mueller-Dombois [1975b]).

The Galapagos Islands, 960 kilometers off the coast

of Ecuador, are unique in their flora and fauna and in


their historical role in the origin of Darwin's theory of

evolution by natural selection. As on many of the

world's island systems, various endemic Galapagos species

are threatened by populations of exotic plants and

animals. Feral horses, burros, cattle, goats, pigs,

dogs, cats, rats and mice inhabit the Islands. Research

projects are being conducted to investigate the ecologi-

cal impact of these introduced species and recommenda-

tions are being made concerning methods of control or


In an effort to determine the effect that feral

burros have on the endangered Galapagos tortoise,

Geochelone elephantopus, I studied feral burro and tor-

toise feeding ecologies, population distributions and

interactions on Volcan Alcedo, Isabela Island. Research

began in October 1979 and was completed in December 1980.

Feral burros occur on all five of the major islands

in the Galapagos Archipelago. The exact date of intro-

duction is not known. Colonists first settled on Isla

Floreana in the 1830s and brought with them a variety of

domestic animals. Since burros are utilized by farmers

on the mainland and are preadapted to arid climates, they

were probably among the first animals taken to the

Galapagos by early settlers. Burros were soon dispersed

to even the uninhabited regions by oil seekers

(R. H. Beck in Van Denburgh, 1914) who used them to trans-

port kegs of tortoise oil to ships and settlements, and

by miners who were in search of sulfur in the deposits

around the volcanic craters.

Estimates of the feral burro populations on the

major islands are as follows: 300 on San Cristobal,

200-300 on Santa Cruz, 500-700 on Santiago (Lucho

Calvopina, pers. comm.), and 2,000-4,000 on Floreana

(Tina Beach and Felipe Cruz, pers. comm.). In addition,

burros occur on three of the five volcanoes that make up

the largest island, Isabela. Volcan Cerro Azul and

Sierra Negra on southern Isabela have relatively small

burro populations; Volcan Alcedo, to the north, has a

population of between 500 and 700 animals (this study).

Feral burros had become established on Isabela by

the 1860s (S. Habel, 1868 in Salvin, 1876). By the

1880s, they were very numerous on Isabela as well as on

San Cristobal, Floreana and Santa Cruz Islands (T. Wolf

in Baur, 1891). Old literature makes no specific mention

of exactly how and when burros arrived on Volcan Alcedo.

Because Alcedo had both a large tortoise population and

sulfur deposits, however, oil seekers and sulfur miners

with pack burros surely visited its slopes.

The Charles Darwin Research Station (CDRS) and

Galapagos National Park personnel have long feared that

feral burros damage the flora and fauna of Alcedo and the

other islands where they occur. Thorton (1971), Wiggins

and Porter (1971), MacFarland et al. (1974a), and van der

Werff (1978) expressed these fears. Prior to this study,

however, no investigation of the impact of feral burros

in the Galapagos had been undertaken. Similar studies

have been completed of feral burros in the southwestern

United States and their impact on flora, their competi-

tion with bighorn sheep, and their effect on birds and

small mammals (Moehlman, 1974, 1979; Woodward, 1976;

Woodward and Ohmart, 1976; USDI, 1977; Hanley and Brady,

1977 a, b; Norment and Douglas, 1977; Seegmiller and

Ohmart, 1981).

In addition to a large burro population, Volcan

Alcedo has the largest remaining population of Galapagos

Geochelone tortoises. In the past, the Galapagos giant

tortoises were heavily exploited; first by pirates,

sealers and whalers in the 1600-1800s, then by colonists

and oil seekers, and during the early 1900s, by scien-

tific collecting expeditions. Well over 100,000 tor-

toises were taken from the Galapagos Archipelago

(Townsend, 1925).

Of the original fourteen races of Geochelone

elephantopus only ten races remain; seven of these are

severely threatened due to decreased populations and

introduced mammalian competitors and predators

(MacFarland et al., 1974a, b). Alcedo's Geochelone

elephantopus vandenburghi population is the least en-

dangered of the tortoise races. However, introduced

rats, cats, and burros on Alcedo pose a potential threat

from both predation and competition.



Volcan Alcedo, Isabela Island

Isabela is by far the largest island in the

Galapagos group. Its land surface area of 4,670 square

kilometers (Wiggins and Porter, 1971) includes more than

half of the total land area of all the islands in the

Archipelago combined. Six volcanoes, connected by exten-

sive lava flows, form this J-shaped island (Figure 2-1).

Volcan Alcedo, in the middle of Isabela and 1128 meters

high, has a large central caldera which is between seven

and eight kilometers wide (Banfield et al., 1956 and

Parque Nacional Galapagos, 1980).

The Galapagos volcanoes are typical, gently sloping

shield volcanoes; several of them are active. Sierra

Negra, to the south of Alcedo, erupted as recently as

1979. Volcan Wolf erupted in September, 1982. Alcedo

last erupted in 1954, from a small fissure on its outer

southeastern flank (Thorton, 1971). There is an active

fumarole on the inner southeastern wall of the crater

which, until 1969, was surrounded by a bubbling pool of

mineral-laden water. The pool has since dried, but

the fumarole remains and emits hot sulfur steam









0 v1
m) u

a, ,

$ 5



continuously. Other small sulfur vents dot the inner

south and southwestern slopes of the caldera, testimony

to the incessant activity underground.

To reach Alcedo's crater, one lands on a small

beach towards the northeastern side of the volcano. From

there, a burro/tourist trail leads ten kilometers up its

flank to the base of the crater rim. My "Midcamp" study

site was located in this area, at the eastern foot of the

volcano (Figure 2-2). The rim of Alcedo rises abruptly

from its outer slopes. A few hundred meters high and rel-

atively flat-topped in some places and hilly in others,

the rim varies greatly in width. My "Rim Camp" study

site was approximately five kilometers southeast of the

ascent path from outer flanks to rim. Six kilometers

further south along the rim was a descent path to the

inner wall fumarole. From the fumarole, a path led to

the caldera floor and my "South Floor" study site. One

of the main tortoise nesting zones was located in this

area. During the rainy season several large pools of

water formed on the south floor. These were heavily used

and of great importance to animals since there is no

permanent source of fresh water on Alcedo. The "North

Plateau" study site was located on the widest, northern

section of the rim. Directly below the North Plateau, on

the caldera floor, was the north floor nesting area,

where many tortoises nested.




im Camp



Figure 2-2
Volcan Alcedo Study Sites

To facilitate data collection, I divided the

caldera rim into four areas of roughly equal length.

Area two encompassed the wide North Plateau and area four

included Rim Camp study site. Areas one and three were

relatively narrow sections of the rim with vegetation

somewhat similar to Midcamp and North Plateau, respec-


The Climate

The Galapagos Islands, though they straddle the

equator, are not typically "tropical" in their climate,

which is strongly influenced by the sea surrounding them.

The Humboldt current, sweeping up the western coast of

South America from Antarctica, turns west at the equator

and bathes the Islands in its chilling waters. Sea water

temperatures range from 19 to 23 degrees C during the

warmer months of Janeary through June and are usually

several degrees cooler during the remainder of the year

(Wellington, 1975).

There are two seasons in the Galapagos: a "warm

season" (also referred to as "wet season") from January

through June and a "cool, garua season" (or "dry season")

from July through December. Rains, though they primarily

fall during the warm, wet season, vary considerably from

one location to the next. During the dry season, coastal

areas may be several months without precipitation. High-

land areas, however, are often blanketed in a fine wet


a o

oa4= 0 D0

r Lno
D -w r- r- in







q E
O a




m rl E
3 C-



m.e U.
-4 1


0 40

U ,

o -


Sf. o c o


1 l 0 4

c0 3 C0
ri C s i2Oe < n o


z c

m ar
N ko

>4 M

3 0

0c i


.-I I
0 J
4-4 4.

C (

4 -I


* *-* *

.w NC oma -D
*bNoo'ro orwooin

C's, UO~('*N



P-) 4-





0 0

4- I 1

0 0


Table 2-2

Daily Air Temperatures (OC)
Isabela Island, 1980

Location Pto.Villamil Santo Tomas V.Alcedo*
Altitude 6 m 350 m 1100 m
Max. Min. Max. Min. Max. Min.

January 26.4 24.3 27.8 18.6 20.6 15.0
February 29.6 26.5 29'.2 19.0 -24.2 15.1
March 27.6 24.0 31.4 19.4 26.5 16.1
April 27.6 24.2 30.1 20.1 26.3 17.0
May 26.1 24.0 26.5 19.4 25.0 16.7
June 25.3 23.3 25.0 17.3 -
July 24.1 21.1 23.0 16.4 21.9 12.0
August 23.9 19.4 23.2 16.2 26.7 11.4
September 23.7 19.0 23.4 16.0 -
October 23.1 19.6 23.2 16.5 24.5 13.0
November 23.1 20.0 24.1 16.5 23.7 12.1
December 23.1 20.8 25.0 16.8 23.3 13.1

*Southeast Rim Camp.

mist, "garua," which condenses on the vegetation. The

prevailing winds are from the southeast, hence the

southern, windward slopes of islands are wetter than

their northern slopes. Wiggins and Porter (1971) and

van der Werff (1978) provide more detailed discussions of

the climate of the Galapagos Islands.

Meteorological data for 1980 on Volcan Alcedo is

presented in Tables 2-1 and 2-2. Monthly rainfall and

daily temperatures from Alcedo (Rim Camp) are compared

with data from other CDRS weather stations in the

Islands. As mentioned previously, rainfall is quite var-

iable and often localized. Areas of higher altitude tend

to receive more precipitation, but not necessarily during

the same months of the year, or as a result of the same

storms. During 1980, rainfall on Alcedo was greatest

between January and May; March was unusually dry. From

August through early November, very little rain fell.

Light rains began again in mid-November. Temperatures at

1100 meters on Alcedo were predictably lower than temper-

atures recorded on the southern coast of Isabela at

Puerto Villamil or on the slope of Sierra Negra at Santo


Sporadic rainfall and temperature records from the

South Floor and Midcamp study sites indicate that these

areas generally received less rainfall than did Rim Camp.

Likewise, at both sites, maximum and minimum temperatures

were usually several degrees higher than at Rim Camp.

Rim Camp, situated on the southeastern section of

Alcedo's rim, was in the path of the wet prevailing

winds. Consequently this area received a great deal of

precipitation in the form of garua, even during the cool,

dry season. Frequently, when the rest of the volcano was

dry and warm, and bathed in sunshine, Rim Camp was wet,

cold and windy, under a thick blanket of garua. Heavy

garua, condensing on the vegetation along the south-

eastern rim, would often form drip-pools under the moss

covered trees. Many of these pools had become quite deep

and enlarged after years of tortoise and burro use.

Daily weather profiles were kept for 121 days while

I was camped at Rim Camp (Table 2-3). On 92.6 percent of

these days there was garua. Rarely was the rim clear at

sunrise, and only 7.4 percent of the days were garua-free

from sunup to sundown. Most common were days with garua

at dawn, followed by some sunshine in the early after-

noon. Often, just before sunset, the garua clouds would

roll in again.

Seeking to compare the amount of moisture received

by the various sections of the volcano, I set up several

"garua catches." Four locations along the crater rim

were chosen and at each site two wire window-screens cut

into squares measuring 50 centimeters by 50 centimeters

were erected. Garua condensed on the screens, dripped

down into a slanted pipe-trough along the lower edge and

was collected in a holding container until measured.

Table 2-3

Monthly Frequency of Garua Days
Rim Camp, Alcedo

Garua/Rain Garua and Sun
Entire Day Sun Entire Day

January 3 2 0
February 5 9 0
March 3 5 3
April 3 11 0
May 8 2 0
July 7 6 0
August 0 6 3
October 1 5 1
November 6 14 2
December 11 5 0

Total 47 65 9
Percentage 38.8% 53.7% 7.4%

I positioned the two screens at each rim location at

slightly different angles, to insure that at least one

would be perpendicular to the direction of the winds for

maximum garua collection. Later, where one screen was

obviously more correctly oriented to the winds, I dis-

carded the data from the poorly positioned screen. Or,

if the two screens collected approximately equal amounts

of water and neither was consistently more efficient, I

averaged the amounts collected.

It was difficult to suspend containers for water

holding more than 3.8 liters, due to their weight when

filled, and to the destructive curiosity of tortoises.

Tortoises destroyed my garua catch number three. I was

forced to attach other screens high or over rockpiles

where tortoises could not reach. Unfortunately, because

I could only infrequently check the garua catches north

and south of Rim Camp, on extremely wet days the 3.8

liter containers overflowed. Therefore the data pre-

sented in Table 2-4 are an underestimate of garua precip-

itation collected by catches number one/two and four.

Catches five and seven never overflowed.

As seen from Table 2-4, the amount of garua mois-

ture condensing on the southeastern section of Alcedo was

much greater than the amount condensing on the eastern

and northeastern rim. I found that differences in amount

of moisture strongly influenced the distribution of both



I C4

>14 ,-4


a) C
(0 U c


10 rJ

1- -
*>- *
Q 3

r i

o -e

0 c II
4- 1 Z as

oE w 0) i
0 H

34 4





0a E-





0 0

3 0
v 0

o o
C C*

a) 0a)
.0 .Q0

a a)
> u
0 a






4 r=
a a

r7< I
!a e >
a) u-I
>u-. 3





o-I a)
(,I r

a) .-I
o o

M En

r u-
0 Ma

M 4 '
x, 3

0 0

a)U -4




tortoises and burros in the dry season. (See Chapter


The Vegetation

The heterogeneous pattern of soil types and mois-

ture availability on Alcedo has produced a diverse array

of vegetational zones. Van der Werff (1978), employing

a method of vegetation classification based on

physiognomical and structural criteria (after Fosberg,

1967) made a detailed study of the vegetation of Alcedo.

He listed the species present in 10 meter by 40 meter

quadrants and estimated species abundance using the

Tansley and Chipp categories (1926, as cited by van der

Werff, 1978). Following are summaries of van der Werff's

descriptions of the vegetation types at my four study


Rim Camp Study Site

Alliance Psychotrian rufipedis, association

Zanthoxylo-Polystichetum gelidi: an open, mossy ever-

green forest. The tree and shrub layers, three to seven

meters tall, were dominated by Zanthoxylum fagara,

Tournefortia pubescens and Tournefortia rufosericea. The

herb layer was nearly closed, with ferns and Verbena spp.

the most common tall herbs present. The low herb layer

was made up of several species which prefer open habi-

tats. These included Borreria laevis, Conyza

bonariensis, Cyperus brevifolius, Dichondra repens,

Hyptis rhomboidea, Mecardonia dianthera, Panicum

fasciculatum, Paspalum conjugatum, Plantago major and

Solanum nodiflorum. Exposure to winds and garua also

contributed to the lowness and openness of the vegeta-

tion. Vascular epiphytic plants were very common.

On the upper outer slopes of the southeastern rim

was the community of Pteris quadriaurita and Tropidia

polystachya. This community, not placed in an alliance

by van der Werff, consisted of two structurally different

vegetation types; one, an open evergreen forest dominated

by Scalesia microcephala and Croton scouleri with a shrub

layer of Tournefortia rufosericea and Psychotria rufipes;

the second, a closed evergreen scrub vegetation dominated

by Psychotria rufipes and Tournefortia rufosericea.

Ipomea alba, the most conspicuous herb, grew in dense

mats on the shrubs. For a complete list of the plant

species which occurred on Alcedo's southeastern rim and

upper slope, see van der Werff (1978). A partial species

list is included in Chapter Six of this report.

South Floor Study Site

Alliance Burserion graveolentis, association

Abutiletum depauperati, subassociation cyperetosum

aristati: a low deciduous forest. The most common tree

was Bursera graveolens and Walteria ovata was the most

common shrub. The tree layer was often only four meters

high; the shrub layer was one to three meters high. The

herb layer, with coverage of about 50 percent, was com-

posed of many species including Cyperus anderssonni, C.

aristatus, C. confertus, Cordia revoluta, Crotalaria

incana, Desmodium procumbens, Paspalum galapageium,

Portulaca oleracea, Phyllanthus caroliniensis and

Chrysanthellum pusillum. Epiphytes (except lichens) were


Several communities which van der Werff did not

place in alliances also occurred on the South caldera

floor. The community of Cyperus ligularis and

Scoparia dulcis was a sparse, open evergreen herb vegeta-

tion, found around fumaroles and sulfur vents. The com-

munity of Polygola anderssonni and Scalesia microcephala,

an open evergreen shrub savanna, was found on the south-

western floor in areas of pumice and obsidian deposits.

Walteria ovata and several herbs including Blainvillea

dichotoma, Cenchrus platyacanthus and Ophioglossum

reticulatum were common in this community. Finally, on

the bare lava flows of Alcedo, of which there were

several in the caldera, the community of Jasminocereus

thouarsii var. sclerocarpus and Pilea peploides, a sparse

vegetation dominated by cacti, occurred. See van der

Werff and Chapter Six of this report for more detailed

species lists.

Midcamp Study Site

Alliance Burserion graveolentis, association

Abutiletum depauperati, subassociation sidetosum rupo:

an open evergreen shrub community with a closed herb

layer. The dominant shrub species was Scalesia

microcephala. The most common herbs included Bidens

riparia, Blainvillea dichotoma, Cenchrus platyacanthus

and Sida rhombifolia. A few trees of Bursera graveolens

and Pisonia floribunda were present; epiphytes were

absent except for lichens.

Areas of Pennisetum pauperum and Acnistus

ellipticus, a more or less closed, evergreen scrub com-

munity, were also found at Midcamp. Common shrub species

were Zanthoxylum fagara, Psidium galapageium and

Tournefortia pubescens. The cover of the herb layer

reached 80 percent after rains and common herb species

were Heliotropium angiospernum, Pennisetum pauperum and

Sida rhombifolia. Epiphytic lichens occurred frequently.

The community of Salvia pseudoserotina and

Polypodium tridens, a more or less closed, evergreen

scrub community occurred on Alcedo's eastern slope around

Midcamp. Common shrubs were Zanthoxylum fagara, Scalesia

microcephala and Psidium galapageium. Common wet season

herbs included Commelina diffusa, Alternanthera

halimifolia and Sida glutinosa. Ferns, bryophytes and

lichens were common.

North Plateau Study Site

The community of Ophioglossum reticulatum and

Tournefortia pubescens, an open or closed evergreen

scrub, was found on the northeastern, northern and west-

ern sections of Alcedo's rim. Trees were absent in this

community and common shrubs, growing to three meters,

were Zanthoxylum fagara, Tournefortia pubescens, Scalesia

microcephala and Darwiniothamnus tenuifolius var.

glabriusculus. Herbs covered 80 to 100 percent of the

ground after the rains and included several ephemeral

species. Some of the most common were Bidens riparia,

Blainvillea dichotoma, Chrysanthellum pusillum,

Ophioglossum reticulatum and Trichoneura lindleyana.

Epiphytic plants were rare.



In the cool dry season when, on most of Alcedo,

puddles and ephemeral plants were desiccating and dying,

the area around Rim Camp remained green and moist, a con-

sequence of almost daily garua. During this time of the

year I suspected that both burros and tortoises converged

on the southeastern rim. There they could feed on damp

grasses and suck what little water was available from the

muddy puddles that formed under dripping trees after a

night of heavy garua. To study the seasonal changes in

the distributions of burros and tortoises, I made regular

bi-monthly around-the-rim censuses. I also made various

burro and tortoise counts in my study sites as frequently

as was feasible.


On all of these censuses and counts I tallied the

numbers, sexes, and ages of all observed burros, and

recorded information on group size and composition when

aggregations of animals were seen. An adult burro, of

course, could not be aged from afar, but a young animal

was classified as either infant (from newborn to five

months), juvenile (from six to ten months), or adoles-

cent (from ten to more than twelve months, but not full

grown). From information kept on all young animals

sighted and from records of all sightings of female

burros in their last months of pregnancy, the peak foal-

ing season for burros was obtained.

Since burro population size could not be calculated

using the traditional methods of mark and recapture,

aerial surveys and the like, I made an estimate of popu-

lation size utilizing data from my counts and censuses.

This estimate was based on the average number of burros

seen in a specific type of habitat on Alcedo and the ex-

tent of that habitat on the volcano. Tortoise population

size was estimated in the same way. A mark and recapture

study might provide a more accurate estimate of the

Alcedo tortoise population, but I was not authorized to

conduct such research.

On the censuses, tortoises were counted and classi-

fied into three arbitrary size categories: small (curved

carapace length of less than 75 centimeters), medium

(curved carapace length between 76 and 105 centimeters),

or large (curved carapace length greater than 106 centi-

meters). I could not reliably sex tortoises, therefore

this information was not recorded.

For the purpose of rim censusing, I mentally

divided the crater rim into four areas based on param-

eters such as local climate, moisture and vegetation

(Figure 2-2). Though these four areas were approximately

equal in length, their widths were variable; the area

which encompassed the North Plateau was by far the lar-

gest. The around-the-rim path I followed on rim censuses

required between six and eight hours to complete, one and

a half to two hours for each area. Because visibility

and rim width varied considerably along my census route,

the amount of land actually surveyed per area also

varied. However, the aim of rim censusing was to monitor

seasonal changes in burro and tortoise distribution.

Therefore, it was important only that I follow the same

path on each census and record all animals seen each

time, so that the resulting identically executed censuses

could later be compared.

Dense garua occasionally reduced the visibility on

censuses. But dense garua rarely occurred on most of

Alcedo, except in the early mornings and along the south-

eastern rim. Fortunately this section of the rim was

sufficiently narrow that animals there were often easy to

count in spite of garua. On garua days, slope counts

could not be made. In order to reduce the loss of data

caused by low visibility due to early morning garua along

the same rim sections, I alternated the direction I went

around the rim on successive counts.

Rim censuses were made as close to the first and

the fifteenth of each month as possible. I began taking

censuses in November 1979 and made a total of 26 through

December 1980. Tortoises were not counted on the 1979

censuses, but they were recorded thereafter. Censuses

were not made in June or September 1980, as I was absent

from Alcedo during those months.

Besides the bi-monthly around-the-rim censuses,

various other censuses and counts were made in my study

areas (Table 3-1). Censuses were made along established

paths and on all of them, excluding the Rim Camp census,

I counted both tortoises and burros. Counts were made

from distant vantage points overlooking or below areas to

be surveyed. From a distance, accurate counts of tor-

toises could not be made, therefore only burros were

recorded on counts. Study site censuses and counts were

made so that I could investigate the local changes in

tortoise and burro abundances and served to reinforce the

distributional patterns elucidated by the around-the-rim


At Rim Camp study site, censuses were made at least

four times each month. These Rim Camp censuses were

taken along the rim, from my camping spot to the fumarole

descent path, at all hours of the day to test whether

burros were observed more often at certain times than at

others. A southeast slope census path was established

and I made the one and a half hour burro and tortoise

census at least twice a month during my year on Alcedo.

Lastly, from a spot on the low outer slopes below the

southeastern rim, burros could be counted on the upper

Table 3-1

Burro and Tortoise Censuses and Counts

Census/ Species Study Site/ Frequency
Count Counted Area Counted of Counts

Rim Census

Rim Camp

SE Slope

Slope Count
from Outer

South Floor

Sulfur Slopes


N. Plateau

burros &


burros &


burros &


burros &


Areas 1-4
rim, and slopes
if visible

Rim Camp/Area 4
rim, and slopes
if visible

Rim Camp/Area
outer slopes

Rim Camp/Area
outer slopes

South Floor
inner caldera

South Floor
inner western


N. Plateau/
Area 2

2 per


4 2-3

4 4 in
a year

2 per

2 per


5 in
a year

slopes. I took the long trip down to the outer south-

eastern slope only four times, but the resulting burro

counts show dramatic differences in numbers of burros on

the southeastern slopes at different times of the year.

South Floor censuses of burros and tortoises at

study site two were made bi-monthly, concurrently with

rim censuses, to study the changing distributions of both

species on the floor. Counts of burros on the sulfur

slopes of the inner western caldera also were made in

conjunction with rim censuses. Midcamp censuses were

made at least twice a month; burros and tortoises were

tallied along a two hour path on the eastern outer slopes

of Alcedo. Finally, five counts of North Plateau burros

were made during the year of study.

Rim Census and Count Results

The total number of animals seen on each rim census

varied greatly (Figure 3-1). The maximum number of

burros counted was 176 animals on July 5, 1980. The

smallest number of burros was observed on

January 30, 1980, when only 21 were counted on the entire

around-the-rim census. Many more tortoises were seen per

census than burros; their numbers also fluctuated from

census to census. The maximum number of tortoises

counted was 640 on July 5, 1980. The minimum number was

169 on October 24, 1980.






. C

- o

- ir

l <

S 0 0 0 0 0 0
0 0 0 0 0 0 0
to .l I IS N
S1IV INV do0 3aIranN

One apparent cause of the observed fluctuations in

the total numbers of burros and tortoises seen on the

censuses was the daily weather. For example, on hot

sunny days, both tortoises and burros spent much of their

time under shade trees and were consequently more diffi-

cult to observe and count. The low counts of burros and

tortoises made in October and November 1980 were the

result of hot weather. On cool, cloudy days, like

July 5, burros and tortoises were especially active and

easy to count. On some mornings thick garua reduced vis-

ibility. On others, however, particularly on an ex-

tremely wet garua morning following a dry spell on the

southeastern rim, burros and tortoises would be out in

large numbers, searching to quench their thirsts in the

drip-puddles forming under trees. These animals were

often so intent on their quest that they were unaware of

my approach, making them especially easy to observe. The

lowest total burro count, made on January 30, 1980, was

the only entire census made in the rain. Visibility was

poor and no doubt animals were less active than normal,

due to the weather conditions.

I averaged census data over months and tabulated

the sightings by area (Tables 3-2 and 3-3). Chi-square

tests on burro and tortoise data showed significant dif-

ferences between the numbers of animals seen in any given

area in different months for all except area one burros

(X2 values ranged from 33.5 to 384.3 with P<.001).

Table 3-2

Average Number of Burros on Rim Censuses
by Month and Area, 1979/80

Jan Mar Oct
Area Nov Dec Feb Apr May Jul Aug Nov Dec

1 6.3 9.0 2.5 3.8 65.0 10.0 14.5 11.0 8.5

2 5.0 9.8 22.5 84.0 66.5 124.0 65.5 11.3 23.0

3 33.7 45.8 14.5 24.0 18.5 36.0 24.5 34.3 47.0

4 43.3 20.0 2.0 2.8 4.5 30.5 21.5 9.7 31.0

Table 3-3

Average Number of Tortoises on Rim Censuses
by Month and Area, 1980

Jan Mar Oct
Area Feb Apr May Jul Aug Nov Dec

1 22.5 35.5 32.0 29.0 3.0 5.3 14.5

2 79.5 111.8 97.0 78.0 24.5 12.3 43.5

3 102.0 119.8 152.5 131.0 67.5 51.3 151.0

4 76.0 70.0 91.5 342.5 266.0 184.3 260.0

~i~"~p -~s~------s~glUIIU1P-P- -~- --~--- I~

A few burros were found on area one consistently through-

out the year. A summary of the changing burro and tor-

toise distributions on the rest of the volcano follows.

Burros were more numerous than would be expected in

a uniform distribution on area two between March and

August, and less common than expected from October to

February (Figure 3-2). On area three, burro numbers were

less than expected or near the expected value between

January and November; more burros were seen than expected

only in December 1979 and 1980. Between January and

July, burros were rarely seen on area four; in the months

of August and December, burros were quite common there.

The changes in tortoise distribution followed a

similar pattern (Figure 3-3). Between March and July

tortoises were common on area one. In December, January

and February, they were seen at the expected rate on cen-

suses. Between August and November there were very few

tortoises along the rim section of area one. During

March, April, and May, there were many more tortoises

than expected on area two; in August through December,

there were fewer. On area three, tortoises were abundant

in May, July, and December. There were relatively fewer

tortoises on area three between August and November. And

finally, tortoises were most numerous on area four be-

tween July and December, while fewer tortoises were

recorded on area four censuses during January to May.


I 0


I I--
I 1 J
I *

x Jx -

Su UJ U J U uJ

N IEn.-I W 1
C a |. 0 CrL
- i i '* 'I i o

0 ( w 0 x w 0 C 0
w UL 2 2 ^ o M
c c 5
S < 1 5
C ~ "" I"(D
i i t ^
l i i u-

. <




0 0
w w
I- tE I- t
w U w Lw u w
0 uj 0 x UJ
x u. 2 uw u.




,0 x

* u



0 x w
w U U.

0 x w
2 UJ u-



#3 r-
S 02
0 0


.4 0

4-1 +
> ^3

To determine whether burros and tortoises were

utilizing the same sections of the volcano at the same

times of the year, I applied separate chi-square tests to

monthly burro and tortoise around-the-rim census data.

All tests showed significant differences between the

burro and tortoise data (January/February atX 2=15.4,

P<.01 and for all others, X2 = 39.0 to 263.1, P<.001);

burros and tortoises were not distributed in the same

manner on Alcedo.

Briefly, the results show that in the early wet

season months of January to April, there were compara-

tively more burros than tortoises on area two and com-

paratively more tortoises than burros on area four. In

May and July, the situation had changed, with tortoises

now comparatively more abundant on area two and burros

more abundant on area four. From August to December,

burros were not utilizing area four to the extent that

tortoises were; however, burros were common relative to

tortoises on areas one, two and three.

The various additional burro and tortoise counts I

made at Rim Camp study site included a southeast slope

census and a slope count from the outer floor. I made a

total of 80 Rim Camp censuses at different hours of the

day on which only burros were tallied. A summary of the

Rim Camp census data, grouped by month and time of day,

is presented in Table 3-4. There was a significant dif-

ference in the number of burros seen at different times

of the day (chi-square test, X 2 = 18.4, P<.01). More

burros were seen than expected in a uniform distribution

from 10:00 a.m. to 2 p.m. in November and December 1979,

and from dawn to 10:00 a.m. between July and December

1980. Fewer burros were seen than expected in October

through December 1980 between 10:00 a.m. and 2:00 p.m. and

in November and December 1979 between 2 p.m. and sunset.

Generally the Rim Camp census data show the same pattern

of burro abundance along the southeast rim that was evi-

dent from the around-the-rim census data for area four.

Few burros were recorded from January to July, with a

substantial increase between August and December.

Fifteen southeast slope censuses were made at

intervals throughout 1980 (Table 3-5). Once again these

data provide evidence that burros and tortoises migrate

into and out of the area along Alcedo's southeastern rim.

Between January and April very few burros and tortoises

were recorded on slope censuses. However, from August to

December many more animals were present on this section

of the volcano. Chi-square tests indicated that statis-

tically significant differences existed between the num-

bers of animals seen on slope censuses during the various

months (X2 =103.7 for burros and 77.5 for tortoises,

P<.001 for both).

Burros on the outer slope also could easily be

counted from down below on the outer floor. I made four

long hikes down to the outer floor to count slope burros

Table 3-4

Rim Camp Censuses

Average Average
Time No. of No. Burros No. Burros
Month Year of Day* Counts** on Rim on Slope

November da-10 6 21.8 33.5
and 79 10- 2 5 30.2 56.6
December 2-du 3 11.3 24.0

January da-10 5 2.2 0
to 80 10- 2 9 1.0 0.2
March 2-du 7 1.9 0.8

April da-10 9 0.8 0
and 80 10- 2 6 0.3 1.0
May 2-du 2 0 0.5

July da-10 8 3.4 3.8
and 80 10- 2 3 1.7 0
August 2-du 6 7.3 0

October da-10 7 25.7
to 80 10- 2 3 10.3 20.5
December 2-du 1 11.0 7.0

*da-10 is from dawn to 10 a.m., 10- 2 is from
10 a.m. to 2 p.m., 2-du is from 2 p.m. to dusk.
**On 37 of the 80 counts, burros on the slope
could not be counted because of thick garua.


January and








Table 3-5

Southeast Slope Censuses

No. of Average
Year Counts No. Burros









Average No.







and a chi-square test showed statistically significant

differences between the total numbers of burros I tallied

on each count. In early November 1979, I counted 161

burros on the slope; in late December 1979, I counted 26;

in early March 1980, only three were counted; and in

mid-October 1980, I counted 118 (X2 = 218.3, P<.001).

Twice each month South Floor censuses of burros and

tortoises, and counts of burros on the inner western sul-

fur slopes were made (Tables 3-6 and 3-7). I found that

both burros and tortoises moved into and out of the cal-

dera; their movements were apparently correlated with

changing water and food availability. The greatest num-

bers of burros on the South Floor were observed between

December and February (Figure 3-2). Tortoises likewise

showed a definite peak of abundance on the floor, but it

occurred a few months later than the burro peak; they

were most numerous between February and May (Figure 3-3).

On the sulfur slopes (the inner caldera slopes of area

three), burros were abundant from March to May and there-

after dwindled to almost zero in December. Chi-square

tests showed significant differences in the number of

burros and tortoises on South Floor censuses, and of

burros on sulfur counts in the different months of 1980

(X2=16.2 for floor burros, P<.02; X2=142.0 for floor

tortoises, P<.001; X2=181.8 for sulfur slope burros,


Table 3-6

South Floor Censuses, 1980



March and




October and


No. of Ave rage

No. of


No. Burros








Average No.








Table 3-7

Sulfur Slope Counts, 1980



March and




October and


No. of Average

No. of


No. Burros









To determine whether tortoises of the three differ-

ent size classes were uniformly distributed on Volcan

Alcedo, I applied chi-square tests to tortoise size data

from my around-the-rim and South Floor censuses. Figures

3-4, 3-5 and 3-6 graphically illustrate the changing dis-

tributions of the tortoise size classes during the dif-

ferent months of the year. Three chi-square tests, each

utilizing size and distribution data for a four month

period, were used; all showed significant differences

among the distributions of different sized tortoises

(X2=112.8 for January to April, X2=88.7 for May to

August, X2=206.5 for October to December; P<.001 for


The tortoise size class distribution results are

complex and difficult to interpret. These data show that

small and medium tortoises seemed to prefer areas one and

four during the early wet season months of January to

April; small tortoises were also common on the South

Floor. Large tortoises preferred area two and were less

common than would be expected if their distribution were

uniform, on areas one and four.

During May to August, small tortoises were more

abundant than predicted by a uniform distribution at the

South Floor study site. Medium tortoises were more

abundant than predicted on area four and large tortoises

were still common on area two. In the months of October

to December, small tortoises were more numerous on areas




Q 0


0 0





U.l U.l
ou ,u U u LUJ
0 x wu 0 x L
2 lu U. 2 u u.


LJ .
o aL&



,- 0

- 0




L) w w
Sw I- I C

o a. 0 0 a. 0 xL

Si1 O L w ALL1


r 2



< S
$ s


one and two in addition to the South Floor. Medium tor-

toises were more abundant than expected at uniform on the

South Floor and large tortoises were common on area four.

Counts made at the Midcamp and North Plateau study

sites were taken at infrequent intervals, whenever I

could travel to these areas (Tables 3-8 and 3-9). Burro

numbers were significantly different between months at

both Midcamp and North Plateau (X2=50.2 for Midcamp

burros, X2=49.5 for North Plateau burros, P<.001 for

both). A larger number of burros than expected was

tallied at Midcamp in February and July and fewer burros

than expected were counted there in November and December

(Figure 3-2). Burros on the North Plateau were most num-

erous in July, and least abundant in November. Tortoise

Midcamp census totals did not vary significantly during

the months of 1980.

Using data from the various censuses and counts

reported above, and knowledge of Volcan Alcedo following

a year of extensive hiking and camping there, I estimated

the Alcedo feral burro population to be between 500 and

700 animals. I estimated the tortoise population to be

around 3,000 animals. MacFarland et al. (1974a) esti-

mated 3,000-5,000 tortoises on Alcedo but I have been

unable to find out on what procedures their estimate was

based. As mentioned earlier, I based my estimates on the

number of burros or tortoises generally observed in a

Table 3-8

Midcamp Censuses, 1980




and May





No. of Average

No. of








No. Burros








Average No.








Table 3-9

North Plateau Counts, 1980






No. of Average

No. of





No. Burros





particular habitat on Alcedo and the extent of that

habitat on the volcano.

Burro Foaling Season and Group Size/Composition Results

Very young burros were recorded on Alcedo during

all months of 1980. Yet there was a definite peak in the

number of births during the rainy season (Figure 3-7).

Between March and July, burros from newborn to five

months old were observed more often than young animals of

other age classes. However, very few newborn to five

month olds were seen near the end of the year, in

November and December. Graphs of the frequency of sight-

ings of juvenile burros (six to nine months old) and

adolescents (ten months to over a year old, but not

fullgrown) follow the pattern established in the graph of

infant sightings. By August through December, the rainy

season infant cohort was classified as juvenile; there

was a corresponding increase in sightings of six to nine

month old burros during these months.

Further evidence demonstrating that foals were born

between February and April or May comes from my record of

sightings of pregnant females in their last months before

giving birth. I noted the vast majority of these females

between January and April.

The distributions of pregnant burros and burros

with young were not uniform among areas on Alcedo

(Table 3-10). Chi-square tests show significant




S 4 .. .

.J -

U -










Figure 3-7 Reproductive Periodicity of Burros





differences in the distribution of young burros and

pregnant females on Alcedo in the different months

(X2 = 321, young burros and X2 = 27.3, pregnant

burros; P<.001 for both). Between January and April,

females with young were concentrated in the areas of the

South Floor and on the eastern outer slopes at Midcamp.

Most late term pregnant females were seen at area two, on

the South Floor and at Midcamp.

In May, July, and August, burros less than ten

months old were abundant at area two, on the North

Plateau. Older juveniles were common at Midcamp. Few

late term pregnant females were seen during these months.

Area four, Rim Camp, supported a majority of the females

with offspring from October to December. The only

infants recorded during these months were around Rim

Camp, where most of the late term pregnant females were

also seen. In addition, a fair number of juvenile burros

were observed at area three.

Burros were often seen alone or in pairs, and

groups of all sizes up to 23 animals were recorded on

Alcedo. Three larger groups, each consisting of 29 bur-

ros, were also observed. Seventy percent of all burros

were found solitary or in groups of two to five animals

(Tables 3-11 and 3-12). Twenty percent of the Alcedo

burros were observed in groups of two animals. The most

frequently observed "group" size was one; solitary burros

were sighted on 668 occasions and 34.6 percent of all

Table 3-10

Seasonal Distribution of Young and Pregnant Burros

Month/Area Infants Juveniles Adolescents Females

Jan-Apr/1 5 3 1 0
2 7 8 5 17
3 18 9 5 1
4 11 9 3 2
South Floor 35 26 11 18
Midcamp 10 21 15 12

May-Aug/1 4 8 3 0
2 52 42 9 2
3 16 10 5 3
4 3 14 2 0
South Floor 12 8 2 2
Midcamp 6 11 16 3

Oct-Dec/1 0 4 3 4
2 0 5 2 3
3 0 27 14 1
4 7 70 45 19
South Floor 0 9 10 5
Midcamp 0 3 5 0






5 a a



o a a
n n






I -








I (N N


q. N *D

inc -
W; o r:

4 10 -

N N -

WI ffI -

N U9
0 a

a a

a a

m a

d d
- o
d W

* n

- W

*o N

iU 1
U'I 1


0 0




s -

B a
s cc

I I I I I n

N -
N IN ;

N a
0 1 I I I


-N a a

0 0 1

Q 2 I I a '
N U N m

a a e

A 0

a a a a I *

0 I IO
0- o w o
P 9 in
A A 14 C
0 0 a


0 d di 0. 1 -

n I I i
SN in 0. 2

i i -- i

0, 4 -
112 2 13w

"groups" seen were of single animals. Kolmogorov-Smirnov

tests (Siegel, 1956) indicated that burro group sizes did

not vary significantly either between areas or with the

months of the year.

Complete burro group composition data were diffi-

cult to obtain. On approximately half of the occasions

when I sighted burro groups I could sex all the adult

animals so that complete group composition information

could be recorded. In the total of 5,345 burros observed

during my year of research on Alcedo, 2,683 were sexed.

Five types of burro groups were recognized: 1) all-male;

2) all-female; 3) male and female; 4) female and young;

5) male, female and young. Chi-square tests were used to

determine whether the frequency of sightings of the dif-

ferent group types varied between the months and from

area to area. Both tests showed highly significant

differences between the expected and observed frequencies

of group types (X2=65.1 for groups compared in differ-

ent months, X2=65.1 for groups compared on different

areas, P<.001 for both).

All-male groups were seen at approximately equal

frequencies in all months of the year (Figure 3-8). More

all-female groups than would be expected in a uniform

distribution were found during the wettest months of

January through April. Fewer all-female groups were

observed in October to December.

I cn


I +


I I + u
I Io o

0 0 o

+ s
o 5

I- 1-
I 0 a
0 m

I 0+e0
I -I
-J 1 ..

0 0
C 0 U,

I i II I I
I I'-- .J

oo O a 0X
Qo o ox r

W U. Z I I. : i U. .-r

SdnOld0 JO AON3no03h. 3A1VYI3U

Few male and female groups and male, female and

young groups were recorded between October and December.

Females with young were seen less frequently than a uni-

form distribution would predict in the months of January

to April and more often than expected in October to


Male groups were most frequently recorded on area

two (North Plateau) and at Midcamp (Figure 3-9). Fewer

all-male groups were seen on the rest of Alcedo. All-

female groups apparently preferred area four, around Rim

Camp. Male and female groups were most common on the

South Floor and least common in the Rim Camp and Midcamp

study site areas. Females with young preferred areas

three and four and were notably less frequent on area

two, the South Floor and at Midcamp. Burro groups con-

taining males, females and young were more common than

expected in a uniform distribution on areas one and

three, and infrequently recorded on area four at Rim


All-male groups of up to 16 animals were observed;

however, over 50 percent of the males in all-male groups

were either solitary or in pairs. Late term pregnant

females seemed to have a tendency to be in all-female

groups. Twenty-four percent of the female burros in sin-

gles, pairs or trios of all-female animals were pregnant.

For comparison, only seven percent of the females in male

and female groups were noticeably pregnant. No all-female







o Q 0 a
U. ww 1 U ,
C C~ ~ W ~ II-
LLI UA L w jw w wLLJ
0) xJx OxL w x ,W
SwLL w. 2 L J UL w UJL w IJIL

SdnOulO d0 AON3nO3ut: 3AI1V131U




groups of more than three animals were recorded. Female

groups of up to four or five and even more may exist on

Alcedo; it was more difficult to quickly sex females than

males from afar.

The largest mixed sex and age burro groups observed

were of 15 animals, including three young burros. The

largest male and female groups were of six animals; one,

made up of one male and five females and another of three

males and three females. Other larger mixed sex groups

no doubt existed but I could never rapidly sex all the

animals in the largest groups before they fled. Sixty-

four percent of the male and female groups contained

either one male and one female or one male and two

females. I never saw more than five young burros in a

group (a group of one male, five females and five young)

and groups of up to seven females with three young were

observed. Seventy-five percent of the female with young

groups were of a single mother and offspring. Eleven

percent were made up of a mother and her offspring plus

another female, six percent were of two females with two

young and four percent were of three females with one

young. All other female and young groups were seen less

than one percent of the time.


Burro and Tortoise Distribution

Studies of feral burros in North America have shown

that burros use a wide range of habitats and frequently

have seasonally distinct habitat preferences (Moehlman,

1974; Woodward, 1976; USDI, 1977; O'Farrell, 1978;

Seegmiller and Ohmart, 1981). Summer burro distributions

in the southwestern United States were found to be

strongly influenced by water availability. Burros were

generally concentrated within three or four kilometers of

water sources during the hottest months of the year. In

summer, Moehlman (1974) observed that adult burros

watered every 24 hours in Death Valley, while females

with young foals watered several times a day. During the

cooler winter months, burros ranged up to 10 or 13 kilo-

meters from water. In winter in the Chemehuevi Mountains

of California, burros watered every three days (Woodward,


Water availability apparently also influences the

distribution of burros on Alcedo. During the early half

of 1980, the puddles which formed after heavy rains pro-

vided ample water for both burros and tortoises on all

sections of the volcano. At this time of the year,

burros were most common on the South Floor, the North

Plateau and Midcamp areas. They avoided the Rim Camp

area which was often cool and foggy, particularly in the

mornings. On Alcedo, unlike the southwestern U.S., the

hottest months of the year are also the wettest. Hence,

when Alcedo's weather was hot, burros usually had plenty

of water available to them. There was abundant forage on

all of the volcano during the warm wet season.

In mid-August, burros began leaving the wide North

Plateau and Midcamp areas where grasses and ephemeral

plant species were desiccating. By late in the year,

large numbers of burros were concentrated on the moist

Rim Camp section of Alcedo. This was particularly true

in November and December of 1979. In 1980, light rains

had begun to fall on Alcedo by mid-November but 1979 was

dry until late December; this explains the greater number

of burros at Rim Camp at the end of that year. There

burros could obtain moisture from garua dampened grasses

and occasionally find water in garua drip-pools under


Burros are known to travel for several kilometers

to reach water (Woodward, 1976). Alcedo burros appar-

ently traveled regularly from other areas in search of

water near Rim Camp during the dry season. On wet garua

nights, particularly on wet nights following several con-

secutive dry days on the southeast rim, burros were

active and noisy during the night. All night long,

caravans of burros would pass my camp. I could hear

burros just ten meters from my tent, stomping and slurp-

ing in the mud under a tree from which garua dripped and

formed a favorite drinking hole. In the mornings, I

could hide and watch as single file burro trains passed

me, going from one muddy puddle to the next in search of

moisture. I once followed a lone male burro for an hour

as he investigated over a dozen garua drip-holes, wading

and putting his muzzle in the mud. Alcedo burros behave

differently around water than do the feral burros of

California; Woodward (1976) reports that she never saw a

burro so much as put its hoof in the mud or water of the

Colorado River. Twice I saw burros licking water

droplets off mosses and leaves.

The shortage of water in the dry months of the year

on Alcedo also seemingly influenced the distribution of

pregnant female burros and females with offspring. By

the end of the dry season, in October to December,

pregnant females and females with young were concentrated

on area four at Rim Camp. Since lactating females

provide fluids for themselves and their young, their

water requirements are greater than those of the average

adult female burro.

The lack of a permanent source of water on Volcan

Alcedo is a relatively new situation on that volcano.

Until the late 1960s, a pool of water surrounding the

southeastern inner caldera wall fumarole provided year

round water for burros and tortoises. When the pool mys-

teriously dried up, several hundred burros died. The

current burro population on Alcedo may be much smaller

than it once was (MacFarland, pers. comm.)

Burros are well adapted to life in arid environ-

ments and are able to withstand a water loss of up to 30

percent of their body weight (Maloiy, 1970). On Alcedo,

however, they must be under water stress during the dry

season when they may go for several months without the

chance to drink fully. Both their behavior at times of

some garua-water availability and the changes in their

distribution as the dry season progressed provide evi-

dence that water shortages do exist on Alcedo and may

strongly influence the feral burro population.

Tortoises can go for months without food or water.

This fact led to the over-exploitation of the Galapagos

tortoises for oil and meat by pirates, sealers, whalers

and other seamen in the seventeenth and eighteenth cen-

turies. Tortoises were stored alive aboard ships for

several months. One tortoise, lost aboard the ship Niger

out of New Bedford, was found alive in the lower holds

after two years (Townsend, 1925).

In spite of their ability to withstand long periods

without water, tortoises too seemed to be distributed on

Alcedo in response to water, and possibly to food avail-

ability. And, like the burros, tortoises were very

intent on searching for water under dripping trees fol-

lowing a wet garua night on the southeastern rim. On

mornings in the dry season when there was only a very

small amount of water in garua drip-puddles, tortoises

would spend hours going from puddle to puddle nosing in

the mud (see Chapter Seven for details).

Tortoise distribution was actually more straight-

forwardly related to moisture availability than was burro

distribution. This was probably because tortoises,

unlike burros, are not capable of rapid mobility. While

burros could easily travel several kilometers to reach

the southeast rim during a wet night, drink and leave,

all in the span of a few hours, tortoises travel too

slowly to do this. They apparently had more regular

migrations to and from the various sections of Volcan


During the rainy months, many tortoises were con-

centrated on the South Floor where the largest pools for

drinking and wallowing formed. Tortoises were also com-

mon on areas one and two along the rim, where food and

water were plentiful during the wet season. Rim area

four, Rim Camp, with its many rainy season puddles and

abundant forage, had few tortoises, again perhaps because

of the cool foggy days that were common there yet rare on

other sections of Alcedo.

Data on both burro and tortoise use of the Rim Camp

area confirm that tortoises moved into that section of

the volcano in June or July after the rains had ended and

remained there until the end of the dry season. Burros,

however, were moving in and out of the Rim Camp area.

The two species were not identically distributed on

Alcedo, although their patterns of distribution are

similar. Both burro and tortoise distributions may be

the result of the combined influences of water and food

availability. Food availability on Alcedo will be dis-

cussed in Chapter Six.

Burro Foaling Season and Group Size/Composition

A peak in burro natality occurs on Alcedo during

the rainy months, when ample food and water resources are

available. Infant burros were seen during all months of

the year on Alcedo, however. Moehlman (1974) likewise

found year round reproduction in the burros of Death

Valley, with a peak in births occurring when forage was

abundant, between May and July. Foaling in the burro

herds of the Grand Canyon was restricted to the months of

March to July (USDI, 1977). Woodward (1976) reported no

peak foaling season for the Colorado River feral burro

herds and proposed that the mild winters there may not

exert selective pressures towards the development of a

distinct breeding season. Water is available to these

burros all year long. Indeed, it seems likely that peak

birth seasons are selected for only in populations where

foals produced out of season have a decreased chance of

survival. This may well be the case on Alcedo; giving

birth in the dry season may be selected against since

females with foals would be burdened with an added fluid


Studies of the feral Equus asinus populations in

North America have revealed that feral burros exhibit a

range of social organizations and behaviors. In the

southwestern United States, adult male burros are typi-

cally solitary and some are territorial. Temporary

groups of mixed ages and sexes are common and the basic

stable unit is the mother and offspring pair (Woodward,

1979). Koehler (1974) is alone in reporting the occur-

rence of stable groups in the Southwest of four to six

animals existing for several months on the periphery of

the feral burro range at Bandelier, New Mexico. However,

on the lush humid island of Ossabaw, Georgia, stable

burro groups are the rule, rather than the exception

(Moehlman, 1979 and McCort, 1979). On Ossabaw, the

occurrence of stable harem groups and of high sociability

(greeting, mutual grooming, social play in foals, etc.)

among these groups may be in response to an environment

with near optimal conditions (Moehlman, 1979).

The social organization of the wild African ass

(Equus asinus) has been described by Klingel (1972, 1977)

as a form of territoriality where the only stable groups

are mother and offspring pairs. Fowler et al., (in

prep.) found that during the wet season on Alcedo, stable

harem burro groups with specific home ranges occurred.

During the dry season, when food and water resources are

not abundant, these groups may disband and the individ-

uals then disperse. Evidently burro social organization

is extremely plastic and is strongly influenced by

various environmental factors.

In the Alcedo population, 34.6 percent of all burro

sightings were of single animals, mostly males. This

falls into the range of percentages of solitary animals

recorded in other feral burro studies; 23.9 percent soli-

tary males reported by Moehlman in Death Valley and

50 percent by O'Farrell (1973) in one of his study herds

in Nevada-Arizona. For Death Valley burros, 60 percent

of all groups contained two to four individuals; 51 per-

cent of the groups were of two to four animals on Alcedo.

Larger groups of up to 20-29 burros were reported in most

of the studies from the southwestern United States.

These large herds were usually associated with a scarce

resource such as water or an estrous female. I saw simi-

lar herds on Volcan Alcedo; many of the largest were

gathered around shade trees and favorite dust-bathing

localities. Other large, predominantly male groups were

undoubtedly temporarily attracted together by an estrous

female. On Alcedo, only four percent of all burro groups

contained eight or more animals; similarly in Death

Valley, three percent of the burro groups sighted were

of 8-21 individuals (Moehlman, 1979). Solitary female

burros and all-female groups appear to be more common on

Alcedo than in other study areas; other researchers have

seldom noted females alone.

Young burros (those still with their mothers and

not yet full grown) make up 9.3 percent of the Alcedo

burro population. Seegmiller and Ohmart (1981) found

that the age structure in the Bill Williams Mountains was

64.4 percent adult, 16.7 percent yearlings and 18.9 per-

cent foals. Woodward (1976) has comparable data from the

burro herds in southeastern California. According to

her, the Chemehuevi Mountain burros epitomize a success-

ful colonizing exotic species in a habitat without limi-

tations; they show precocious sexual maturity and have a

high reproductive rate. Twenty-three percent of that

population is made up of young burros. The Alcedo

population, by comparison, may be near the carrying capa-

city of its environment and restricted by some limiting

resource (water or food), hence, the relatively low

reproductive rate.



The feral burros on Volcan Alcedo live completely

free from predators. The endemic and introduced pred-

ators of Alcedo (hawks, owls and cats) are all too small

to prey on burros. Packs of feral dogs may hunt young

and weak burros elsewhere in the Archipelago, but there

are no feral dogs on Alcedo. And, because theirs is an

isolated population which never comes into contact with

other large mammalian species, the burros of Alcedo

rarely encounter disease. Hence I was intrigued, when

I arrived on Alcedo in October 1979, to find fresh car-

casses of numerous adult and juvenile burros.

Mortality in the herds of feral burros of the

southwestern United States has been mentioned by several

researchers. In the populations studied, the observed

natural adult mortality rates were uniformly low, with

juvenile mortality somewhat higher and more variable

(Moehlman, 1974; Norment and Douglas, 1977; USDI, 1977;

Seegmiller and Ohmart, 1981). To investigate the appar-

ently high level of burro mortality on Volcan Alcedo,

I examined burro carcasses and collected teeth from the



Recently dead burros were easily located by smell.

The bleached bones of older skeletons also could be

readily detected, particularly in the dry season when

plant growth was minimal. With the bi-monthly around-

the-rim censuses, frequent study site censuses and

counts, and trips to the landing beach every fifteen days

for food supplies, I had ample opportunity to search much

of Alcedo for dead burros. There were, of course,

sections of the volcano that I did not visit frequently,

and some areas that I never explored. I did not discover

all the burros that died on Alcedo in 1980. But I

certainly found a large percentage of the animals that

died along the crater rim and in my four study site


Beginning in January 1980, I searched for all dead

burros that were detected by scent. In only four in-

stances was I unable to locate or to reach a carcass. In

addition to investigating fresh carcasses, I searched the

crater rim and my four study site areas for older skele-

tons. Dead burros were sexed and aged whenever possible.

Young burros could be aged based on knowledge of Equus

tooth eruption timing and sequence, but adults could not

be aged in the field. Young burro carcasses could be

sexed but skeletons of juvenile animals could not be.

Animals older than four years could always be sexed;

males have large canine teeth, while in females the

canines are absent or rudimentary (Simpson, 1951). The

first incisor, first premolar and first molar of upper

and lower jaws were collected from the skulls of adult

burros. Based on the extent of decay of carcasses, I

estimated the approximate date of death for each animal.

Carcasses were examined for clues to the cause of death.

Older skeletons were classified as to length of time

since death of the animal according to the weathered

appearance of the bones. I was later able to use bone

weathering information from Behrensmeyer (1978) to

translate these classifications into approximate

estimates of years since death.

Equus spp. have traditionally been aged based on

tooth wear. Wear is related to diet, soil conditions,

and in addition, varies from individual to individual.

More recently, mammalogists have been using an aging

technique in which tooth cementum layers are analyzed.

Cementum is produced throughout a mammal's life by

cementoblast cells on the outer surface of the tooth

roots. Dark cementum bands are formed when there is a

change in cementoblast activity. In North American, dark

bands are thought to be formed during winter; in the

tropics, they apparently coincide with the dry season

(Matson, 1981).

Matson's Commercial Microtechniques Lab in

Milltown, Montana specializes in tooth sectioning and

aging by cementum analysis. I sent incisors from dead

burros I had located on Alcedo to Matson's for process-

ing. Decalcified teeth were sectioned longitudinally at

14 microns, stained with Giemsa and permanently mounted

on microscope slides for aging (see Humason, 1972 for a

description of standard paraffin preparation method).

Cementum band patterns vary among species (Matson, 1981).

For aging burro teeth, Matson's assumed that tooth erup-

tion occurred before the age of three years; hence the

first major dark cememtum band on a tooth section marked

the third year of life.

Matson's can only handle burro incisors; premolars

and molars are too large to fit into their trimming and

sectioning equipment. But as skulls deteriorate in the

field, the first teeth to become loose, and therefore

lost, are the incisors. Hence for some of the oldest

skeletons found on Alcedo, I was unable to collect

incisors. Ages of animals for which only premolars and

molars were collected were based on wear criteria.

Molars were measured according to Joubert (1972) and

then, using linear regression, correlated to the molar

measurements of animals aged by cementum analysis

(correlation coefficient of lower molar length to age =

.90 for females, .75 for males and .80 combined).


Thirty recently dead burros were found on Alcedo

in 1980. For 22 of these, I made notes in the field

concerning the stage of tooth eruption or collected teeth

for laboratory aging. Four carcasses of adult burros

were too fresh to extract teeth when I first discovered

them and I could not locate them later. I could smell

but was unable to find or reach another four dead


Over half (56.7%) of the dead burros I discovered

had died in January 1980, at the very end of the dry sea-

son (Figure 4-1). All the others, except for one which

died during the rainy month of March, died between

September and December of 1979 or of 1980, which were

dry months on Alcedo. Twelve of the 30 burro carcasses

were of adults and fourteen were of animals under two

years of age.

Table 4-1 summarizes the sexes and ages of the

126 burro skeletons and carcasses examined. Aging was

done by the three methods previously described. The

adult male to female ratio was 1:.95. Other researchers

also have reported sex ratios approaching one-to-one in

some of the feral burro populations of the southwestern

United States (Moehlman, 1974; Woodward, 1976; USDI,

1977; Norment and Douglas, 1977). Burros in the age

classes of three to five years, six to nine years and ten

to 14 years were found at almost equal frequencies. Each

class comprised between 17 and 23 percent of the total

number of dead burros. Fewer dead animals (7.1%) that

were older than 15 years were found. More burros of

0 18


0 6-

1 I I I i I 1 4 2
O'ND J F 'M A'M' J J A' ON D
1979 1980

Figure 4-1 Months in Which Burros Died







Table 4-1

Burro Sex and Age at Death

Sex and Age in Years
Not 3-5 6-9
:ed ? ? ? ?

L -



10-14 15+
? d ?

- 6

10 0

4 2

Total 42

Percent 33.3

*Based on teeth

29 24 22 9

23.0 19.0 17.5 7.1

from 104 skeletons and 22 carcasses.





newborn to two years (33.3%) were found than any other

age class.

The 42 skeletons and carcasses of young burros

examined could be aged to within a few months

(Table 4-2). More than half the dead young animals were

less than nine months old, with seven to nine month old

individuals the most numerous.

The 30 burro carcasses and 104 skeletons that I

located were grouped into categories based on the approx-

imate number of years since death of the animals

(Table 4-3). The carcasses, and skeletons to which bits

of flesh and hide still clung, belonged to animals that

had died within a year. Of these dead animals, 46.2 per-

cent were adults and slightly more, 53.8 percent, were

the remains of burros less than two years of age. The

skeletons of animals that had died one to three years

previously included a similar percentage of adults

(47.6%) and juveniles (52.4%). But the older skeletons

were predominantly of adult animals (70.6% and 85.7%);

fewer skeletons of young burros that had lain exposed to

weathering for more than four years were found.


Foal mortality rates are variable in the feral

burro herds of the southwestern United States.

Seegmiller and Ohmart (1981) and Ohmart et al. (1975)

found no evidence of foal mortality in Arizona.

Table 4-2

Young Burro Age at Death

Less Than 7-9 10-12 13-24
6 Months Months Months Months

Number 8 17 7 10

Percent 19.0 40.5 16.7 23.8

*Aged by tooth eruption in 28 old skulls and
14 fresh skulls.

Table 4-3

Years Since Death Based on Weathering of Burro Bones

Approximate Years No. Carcasses Percent Percent
Since Death Or Skeletons Adults Young**

1 30 46.2 53.8

1-3 42 47.6 52.4

4-6 34 70.6 29.4

7-15 28 85.7 14.3

*Based on teeth from 104 skeletons and 30 carcasses.
**Less than two years of age.

Moehlman (1974), working in the Panamint Range of Death

Valley, found that foal mortality rates were moderate.

She observed one dead foal and three pregnant females who

either aborted or lost their foals at an early age. By

counting the foals and yearlings in her population

between 1970 and 1972, she estimated first year mortality

at 20-30 percent. In Grand Canyon, Arizona, USDI (1977)

found few burros less than one year of age, and proposed

that an even higher level of mortality of the young may

occur there.

Natural mortality among burros older than one year

is consistently low. Seegmiller and Ohmart (1981) report

very low mortality in the Bill Williams Mountains of

Arizona; during their year-long study only one case of

natural mortality was documented. Other investigators

have likewise observed few burro deaths due to disease or

predation and state that most yearling to adult mortality

is caused by man (Moehlman, 1974; Ohmart et al., 1975;

USDI, 1977).

Norment and Douglas (1977), during a 16-month study

of approximately 160 burros in the Panamint Mountains,

found nine dead burros. At least three of the seven dead

adults had been illegally shot. The two foals had appa-

rently died of natural causes. In all of the studies

mentioned above, natural mortality among burros older

than one year varied from about one to five percent.


The estimated mortality rate on Volcan Alcedo for

yearling to adult burros approaches five to seven per-

cent. This calculation assumes an adult burro

population of approximately 500 animals and takes into

consideration that 13 dead burros, two years of age or

older, were found during 1980 on the crater rim and in

the four study sites alone (less than 1/3 of the total

area of the Volcano). Furthermore, using data on years

since death based on bone weathering criteria, the 1980

mortality rate was not unusual. Although yearly adult

burro mortality rates can only be approximated from the

death tally data, mortality has apparently been consist-

ently high over at least the past ten years.

The 28 skeletons that had been exposed to weather-

ing for seven to 15 years represent only a portion of the

animals that died during that time period. Behrensmeyer

(1978) reported that bone weathering rates depend on tem-

perature,' humidity and soil. Under most conditions,

bones will be completely broken down after ten to

15 years of exposure to the elements. In equable envi-

ronments (swamps and dense woodlands for instance) bone

weathering is slow. But the conditions on Alcedo (fluc-

tuating wet and dry periods, hot tropical sun) probably

result in the complete disintegration of skeletons well

before the 15th year of exposure.

In general, stable mammalian populations (constant

size and age structure) in which no more than one

offspring is produced per female per year, have mortality

profiles that are roughly U-shaped; highest mortality

occurs in the very young and the very old (Caughley,

1966; Klein, 1982). In such populations, young, sexually

mature adults have a substantially lower mortality rate.

Spinage (1972) found that mortality rates for Equus

burchelli were similar for the sexes. The mortality

profile he obtained closely approximated the U-shaped

curve shared by many large mammals.

Not knowing whether the feral burro population is

stable makes it impossible to construct a lifetable from

the Alcedo burro death tally data. However, it can be

tentatively noted that, on Volcan Alcedo, an unusually

large number of young sexually mature adults and animals

in their prime are dying.

Based on the data from carcasses and skeletons that

had lain exposed to weathering for less than three years,

Alcedo foal mortality is higher than adult mortality.

But foal mortality was certainly underestimated. Spinage

(1972) in his studies of African ungulates found that the

youngest age classes are usually underrepresented in a

death tally. The skeletons of juvenile animals deterio-

rate more rapidly than do the bones of adults, hence the

samples of skeletons that were four to fifteen years old

contained increasingly fewer skeletons of young burros.

Behrensmeyer (1978) found that the bones of juvenile

animals rarely survive more than six to eight years of


Only one carcass exhibited clues to the cause of

death; this individual was found on a steep slope section

of the crater rim with a broken neck. Cause of death

could not be determined for any of the other 29 burros.

I suggest that the high level of burro mortality on

Alcedo is the result of a severe water shortage. Almost

all (97%) of the observed burro deaths occurred during

the latter part of Alcedo's dry season. In particular,

56 percent occurred in January 1980 at the very end of

the 1979 dry season.

That the feral burros were under water stress could

be deduced from both behavioral and distributional obser-

vations. During the dry season when wet garua covered

the southeastern rim, burros would travel from other sec-

tions of the volcano and spend hours searching for mois-

ture in the muddy puddles that formed beneath Zanthoxylum

trees. Burros were even seen to lick water droplets from

mosses. During the later months of the dry season,

females with very young foals were observed exclusively

along the moist southeastern rim and slopes. (See

Chapter Three for details of dry season burro distribu-

tion and behavior).

Burros are well-adapted to dry habitats and can

withstand heat stress and aridity because they are able

to tolerate extreme desiccation of the body (water losses

of up to 30% total body weight), have an ability to

reduce losses of evaporative and fecal water when dehy-

drated, and can continue to eat for several days when

deprived of drinking water (Maloiy, 1970). In laboratory

experiments (Maloiy, 1970), burros deprived of water and

exposed to temperatures of 22 + 2 degrees C had depressed

their food intake by 80-83 percent by the end of an eight

to twelve day period. Daily losses in body weight of

.7 to 1.5 percent body weight (1-3 kg) per 24 hours were

recorded. The longest period of experimental water

deprivation was 12 days. During this period one burro

lost 34 kg and another 28 kg body weight (total burro

body weight approximately 150 kg). Food intake was

reduced from 2.76 and 2.82 to .28 and .36 kg food per

100 kg body weight, respectively.

Adolph and Dill (1938) in a study of water metabo-

lism in the desert, found that a walking burro lost

7.8 percent body weight in one day and required six

liters of drinking water per day. Dr. Robert Ohmart,

Arizona State University (pers. comm.), who studied

burros in the southwestern United States, suspects that

burros, even in a cool environment with high relative

humidity, can not survive for one month without free

water. In a dry season on Alcedo there might not be rain

for up to six months. Occasional wet garua nights may

not provide enough water to sustain the entire feral

burro population between rains.



The female tortoises of Alcedo begin nesting at

the end of the rainy season. Many tortoises dig their

egg chambers in one of two caldera floor sites. Others

nest along the crater rim, on the outer crater slopes and

in spots on the inner floor where the dirt is dense and

deep, and of the appropriate consistency for nest build-

ing (see MacFarland et al., 1974b for'details on nest

site characteristics). Prior to this study, the natural

hatching success of Geochelone elephantopus vandenburghi

was not known. However, MacFarland et al. (1974a) have

obtained natural hatching success data for two other

races of Geochelone in the Galapagos. Studies of the

giant land tortoises of Aldabra Atoll by Swingland and

Coe (1978, 1979) provide comparable hatching success

values for Geochelone gigantea.

Because the two major nesting sites on Alcedo are

regularly frequented by feral burros, Charles Darwin

Research Station and National Park personnel have long

suspected that burros damage incubating nests. In this

phase of my study, I recorded the natural hatching and

emergence success of 1979/1980 tortoise nests. Then, in

July 1980, I began monitoring newly-laid clutches to

observe the effect of trampling by burros on nest



In February 1980, I visited the north caldera floor

nesting area and located nests which had been constructed

during the 1979 nesting season, and from which young had

emerged. These nests were easily found by spotting the

exit hole in the hard sun-baked mud cap which had once

protected the incubating eggs. Several additional un-

opened nests were discovered in which the young had

hatched, moved up the nest column, and were ready to

emerge. These nests were located via their rounded nest

cap. Because most nests were already empty by February,

I broke into them.

The north floor nests were excavated and the number

of egg shells, dead embryos, dead and live young, and

undeveloped eggs in each was recorded. Although the egg

shells were torn, it was still possible to estimate the

original clutch size from these tattered shells (see

Fowler, 1979 for egg shell counting method and accuracy).

In late December 1979, and again in early 1980,

I located vacated nests on the south floor nesting area.

These nests also were excavated and their contents

examined and recorded.

For the above 1979/1980 nests, no estimation of

burro damage to nests could be made because nest research

was begun after most of the incubation period was over.

However, in July 1980, I began marking newly established

nests, and these were monitored at irregular intervals

until December. Each time the nests were checked, the

caps were inspected for any sign of disturbance. If a

nest had been disturbed or damaged, or the young had

emerged, I dug until eggs or shells were reached. I did

not remove or tamper with unopened eggs but any broken

ones were removed, counted and discarded. The contents

of vacated nests were recorded and disturbed nests were

resealed so that remaining eggs might go on incubating.

No attempt was made to build a new mud cap for disturbed

nests, the eggs were simply reburied. But many of these

nests had one-half to three-fourths of their mud caps

still intact.

Many of my marked nests were situated in active

burro paths or tortoise sleeping forms. A few of these

nests were impossible to relocate, and were deleted from

my analysis. Due to the endangered status of the

Galapagos tortoise, I did not disturb active nests to

count eggs, check nest progress, and the like. When I

left Alcedo in December 1980, many of the marked nests

were still incubating; hence my data for these nests are



Emergence Success of 1979-1980 Nests

By mid-December 1979, hatchling tortoises had begun

to emerge from nests in both the north and south nesting

areas. Fourteen nests were excavated and examined on

four separate visits to the south nesting area

(Table 5-1). Clutch sizes ranged from six to 14 eggs,

with an average of 11 eggs + 2.3 S.D. per nest. Undevel-

oped eggs, eggs that were infertile and those in which

the embryo died before it attained a visible size, aver-

aged 2.8 eggs + 2.8 S.D. per nest and were found in all

but three (78%) of the 14 nests. Eleven dead embryos,

and one dead twin embryo, were found. From the total of

153 eggs, 99 hatchlings emerged; thus the overall emer-

gence success for the south caldera floor nests was

64.7 percent.

In February 1980, I located 28 nests that had been

constructed in the north nesting area during the previous

year's nesting season (Table 5-2). The clutch sizes of

these nests ranged from seven to 26 eggs, with an average

of 14.5 eggs + 5.0 S.D. per nest. Again, most nests

(75%) contained one or more undeveloped eggs. Stranded

live young were found in the columns of three nests, and

there were 59 dead embryos in a total of 406 eggs and

shells examined. The emergence success of these nests,

including the young in four unopened nests that were

hatched and ready to emerge, was 65 percent.

Table 5-1

Success of South Caldera Floor Nests

Nest Date
Number Examined

5 Dec 79

6 Dec 79

7 Dec 79

8 Dec 79

9 Dec 79

0 Dec 79

1 Dec 79

2 Dec 79

3 Jan 80

4 Jan 80

1 Feb 80

2 Feb 80

3 Mar 80

4 Mar 80















Undev. Dead Live Dead Emerged

Total: 14 nests 153 40 10 1 1 99

per nest: 10.9 2.8 .7 7.1

Deviation: 2.3 2.8 3.2


*This included two broken eggs which may have
broken by burros or by the nesting female.

Undev. Dead Live Dead Emerged
Eggs Embryos Young Young Young

6 3

2 1 8

2 1 10

1 5

twin 11

3 8

1 3 4

1 8

2 8

7 4 3


3 7

6 5

8 5

Table 5-2

Success of North Caldera Floor 1979/1980 Nests
(Examined February 1980)

Nest Total Undev. Dead Live Dead Emerged
Number Eggs Eggs Embryos Young Young Young

1 7 2 1 4
2 23 12 9 2
3 8 4 4
4 11 1 3 7
5* 10 1 1 2 (6)
6 20 5 6 9
7 14 10 3 1
8 8 2 6
9 15 15
10 15 1 14
11* 18 2 1 (15)
12* 16 2 1 (13)
13 16 2 14
14 8 1 7
15 19 2 5 1 1 10
16 10 10
17 11 2 3 2 4
18 14 2 3 9
19 26 6 20
20* 10 2 (8)
21 12 1 1 10
22 16 6 2 1 7
23 12 1 11
24 22 1 2 19
25 10 1 9
26 19 11 2 6
27 16 1 1 14
28 20 8 2 10

Total: 406 69 59 6 8 264
28 nests

per nest: 14.5 2.5 2.1 0.2 0.3 9.4

Deviation: 5.0 3.3 2.6 4.8

*Nests in which young

had hatched but not yet

1980 Nesting Results

The 1980 nesting season began in May or June, while

I was not on Alcedo, and continued at least into August.

The latest freshly established nest, still damp, was dis-

covered on August 4. A few nests were established by

tortoises after this date, however, because four new ones

were found when the nesting areas were next checked in

October. The earliest hatchlings emerged a few days

before November 4. Because I monitored nests at irreg-

ular intervals and many of my marked nests still had not

emerged by mid-December when I last checked them, I was

unable to get accurate measurements of incubation

periods. However, for the eight nests that produced

hatchlings naturally after an undisturbed incubation

period, the minimum possible length of incubation was

90 days and the possible maximum was '150 days. Similar

incubation periods of 110-250, 85-120 and 98-148 days

have been reported by MacFarland et al. (1974b) for

Geochelone elephantopus porteri and G. e. ephippium and

by Swingland and Coe (1979) for G. gigantea, respect-


On the south caldera floor nesting area, 19 nests

were found and monitored regularly. An additional nine

nests were located late in their incubation cycle and

monitored thereafter. Of the total 28 south floor nests

(Table 5-3), nine (32.1%) were broken into by burros

(i.e., the hard protective mud cap was punctured by a