Citation
Genetic improvement of Eucalyptus grandis (Hill) maiden for coppice productivity

Material Information

Title:
Genetic improvement of Eucalyptus grandis (Hill) maiden for coppice productivity
Creator:
Reddy, Kondala Vikram, 1959-
Publication Date:
Language:
English
Physical Description:
xi, 124 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Breeding ( jstor )
Coppice ( jstor )
Frost ( jstor )
Genetic variation ( jstor )
Growth traits ( jstor )
Inbreeding ( jstor )
Inbreeding coefficient ( jstor )
Phenotypic traits ( jstor )
Seed orchards ( jstor )
Seedlings ( jstor )
Biomass energy -- Florida ( lcsh )
Dissertations, Academic -- Forest Resources and Conservation -- UF
Energy crops -- Florida ( lcsh )
Eucalyptus -- Florida ( lcsh )
Forest Resources and Conservation thesis Ph. D
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1985.
Bibliography:
Bibliography: leaves 118-123.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Kondala Vikram Reddy.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
021911859 ( ALEPH )
13564988 ( OCLC )

Downloads

This item has the following downloads:


Full Text











GENETIC IMPROVEMENT OF EUCALYPTUS GRANDIS
(HILL) MAIDEN FOR COPPICE PRODUCTIVITY










BY

KONDALA VIKRAM REDDY
















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA 1985














TABLE OF CONTENTS



Page

ACKNOWLEDGMENTS........................ iii

LIST OF TABLES ........... ...................... iv

ABSTRACT .............................. ......... vii

INTRODUCTION ............ .......... ......... 1

OBJECTIVES............................ ......... 5

LITERATURE REVIEW ......... .................... . 6

Genetic Improvement through
Recurrent Selection ......................... 6
Inbreeding................................... 9
Diallel Analysis ............................. 14
Selection Indices ............................ 15

MATERIALS AND METHODS........................... 20

Study Area for Diallel Analysis ............. 22
Statistical Analysis ......................... 25

RESULTS AND DISCUSSION ....................... 31

Genetic Improvement through
Recurrent Selection ....................... . 31
Variation in Frost and Coppice Scores........ 42 Inbreeding and Seed Orchard Development...... 44
Diallel Analysis ............................. 50
Selection Indices ............................ 69

CONCLUSIONS........................... ................. 73

APPENDIX 1. PEDIGREES AND ORIGINS OF 529 FAMILIES
USED IN GPOP77 .......... ......... .... 76

APPENDIX 2. DESCRIPTION OF MEASUREMENTS OBTAINED
AT 64 MONTH COPPICE AGE ...................... 99

APPENDIX 3.ANALYSIS OF VARIANCE TABLES FOR GROWTH,
COPPICE AND FROST SCORES .................... 102


ii









Page
APPENDIX 4 IDENTIFICATION OF 300 SELECTED TREES USING SELECTION INDEX 3 ...................... 111

LITERATURE CITED............................... 118

BIOGRAPHICAL SKETCH............................ 124

















































iii















ACKNOWLEDGMENTS



I would like to thank members of my supervisory

committee, Drs. Don Rockwood, Ray Goddard, Charles Wilcox,

Frank Martin, Susan Kossuth, and Randy Chow, for their helpful comments and suggestions. Special thanks go to the committee chairman, Dr. Don Rockwood. He allowed me the

space and time to learn on my own, and to make my own mistakes, from which I learned many valuable lessons. His guidance, patience and support throughout this phase of my studies is deeply appreciated. Special thanks are also extended to Drs. Charles Wilcox and Michael DeLorenzo, Dairy Science Department, for their invaluable help in statistical analysis. I would also like to thank George Meskimen and the U. S. Forest Service who maintained the study population through September 1984.

I also wish to thank Laurel Judd, Craig Reed and Karen Bondurant. Their friendship can never be repaid and will always be remembered.


This study was supported in part under Subcontract no. 19X-09050C with Oak Ridge National Laboratory under Martin Marietta Energy Systems, Inc., contract DE-ACO5-840R21400 with the U. S. Department of Energy and by a cooperative

iv








program between the Institute of Food and Agricultural Sciences of the University of Florida and the Gas Research Institute, entitled "Methane from Biomass and Waste."












































v














LIST OF TABLES


Table Page

1 Eucalyptus grandis progenies in
GPOP77 by generation and origin ..................21

2 Diallel analyses of variance for four
Eucalyptus grandis traits in GPOP73 .............. 28

3 Relative economic weights used in the ten
selection index equations for five
Eucalyptus grandis traits........................29

4 Eucalyptus grandis generation means
for individual tree performances in GPOP77
through 64 months after harvest ..................32

5 Percent gains for Eucalyptus grandis
generations compared to the preceding
generation for different traits ................. .35

6 Comparison of origin of Eucalyptus
grandis progenies in GPOP77 ...... ........... .....37

7 Heritabilities of Eucalyptus grandis
traits by and over generation ....................39

8 Genetic correlations and their standard
errors (in parenthesis) among growth traits,
frost and coppice scores of Eucalyptus
grandis in GPOP77. .............. .......... 41
9 Mean frost and coppice scores of Eucalyptus
grandis by origin in GPOP77 ......................43

10 Frequency of different degrees of relationship
in a 50-family subset of Eucalyptus grandis
seed orchard in GPOP77.......................... . 46

11 Mean inbreeding coefficients (F) for five
selection strategies of Eucalyptus grandis
:in GPOP77 ........................................47

12 Predicted genetic gains in 64-month coppice
volume for alternative improvement strategies
of Eucalyptus grandis in GPOP77..................49

vi









Table Page

13 Mean seedling height (dm) at 6 months for 45
crosses of Eucalyptus grandis in GPOP73.......... 51

14 General combining abilities and associated
variances of 10 Eucalyptus grandis
families for seedling height at 6 months and
coppice height at five months in GPOP73..........53

15 Specific combining abilities of 45 Eucalyptus
grandis crosses for seedling height at 6
months in GPOP73 ................................54

16 Mean seedling height (dm) of 45 Eucalyptus
grandis crosses at 2.4 years in GPOP73...........56

17 GCA and associated variances of 10 Eucalyptus
grandis parents for seedling height and
DBH at age 2.4 years in GPOP73..................57

18 Specific combining abilities for seedling height
at 2.4 years for 45 Eucalyptus grandis
crosses in GPOP73............ ............. ..... 58

19 Mean seedling DBH (cm) of 45 Eucalyptus
grandis crosses in GPOP73 at age 2.4 years.......59

20 Specific combining ability of 45 Eucalyptus
grandis crosses for 2.4 year seedling DBH
in GPOP73 ....................................... 60

21 Mean coppice height (dm) of 45 Eucalyptus
grandis crosses at 5 months in GPOP73 ........... 62

22 Specific combining ability of 45 Eucalyptus
grandis crosses for 5 month coppice height in
GPOP73........................ ... ....... ..... 63

23 Estimates of variance components and genetic
variances for four growth traits in Eucalyptus
grandis ..........................................65

24 Rankings of general combining ability of 10 Eucalyptus grandis parents from GPOP73............66

25 Spearman rank correlations for rankings of 10
Eucalyptus grandis parents for four growth
traits in GPOP73 ....... .......................... 68




vii








Table Page

26 Phenotypic variances (diagonal ) and
covariances (off-diagonal) for Eucalyptus
grandis traits used in selection index
equations .......................................70

27 Genotypic variances (diagonal) and
covariances (off-diagonal) for Eucalyptus
grandis traits used in selection index
equations in GPOP77 .............................71

28 Selection index coefficients for five
Eucalyptus grandis traits fot ten
equations in GPUP77 .... ...... ..... .... 72






































viii














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


GENETIC IMPROVEMENT OF EUCALYPTUS GRANDIS
(HILL) MAIDEN FOR BIOMASS PRODUCTION By


Kondala Vikram Reddy


December 1985


Chairman: Dr. Donald L. Rockwood
Major Department: Forest Resources and Conservation


A genetic base population (GPOP77) of Eucalyptus

grandis (Hill) ex. Maiden planted in July 1977, with 529 families representing four generations of selection, was partially harvested in August 1978. Regrowth through December 1983 was evaluated to assess genetic improvement potential for coppice productivity.


Four generations of selection have produced impressive

genetic gains. At 64 months after harvest, first, second, third and fourth generation families averaged 3.9, 13.34, 16.18 and 23.16 dry mt per hectare, respectively. Fourth

generation families had the best frost resilience and quality of coppice. In volume production the best trees


ix









were more than three times larger than fourth generation trees at 64 months after harvest. An obvious potential thus

exists for improvement in biomass production through clonal propagation of the best individuals.


The potential inbreeding depression resulting through mating of related families was examined. The mean inbreeding coefficient in the offspring of all possible matings of selected individuals for six different selection strategies ranged from 0 to 1%. The predicted genetic gains adjusted for inbreeding were high. The highest gain of 90% was predicted by the selection of the top 100 families (three trees per family).


A 10 X 10 half diallel analysis revealed significant amounts of non-additive genetic variation in E. grandis. For seedling height at 6 months and height and DBH at 2.4

years, the dominance variance was higher than additive variance. This suggests potential improvement of seedling growth is higher through the utilization of non-additive genetic variation through clonal propagation. However, for coppice height at five months, the additive variance was higher than dominance variance, suggesting that utilization of additive genetic variance through selection can result in high genetic gains.


Selection index equations were developed using growth

traits and frost and coppice scores. The economic weights used in developing these equations were found not to affect

x









the rankings of individuals. This was evident from the high rank correlations among the ten ranks based on ten selection index equations.


With continued breeding and new introductions, coppice productivity of E. grandis can be improved for biomass

production in southern Florida.








































xi















INTRODUCTION


Eucalypts are evergreen hardwood trees and shrubs

belonging to the genus Eucalyptus in the family Myrtaceae. The genus includes more than 500 species nearly all native to Australia. Several are native to Indonesia, New Guinea

and the Phillippines (FAO, 1979). Eucalypts are noted for their extremely fast growth rates and under favorable environmental conditions grow year round. Although they

occupy perhaps one-fifth of the world's man-made forests (Logan, 1967), only a small part of the world's tree

breeding effort has been devoted to this genus. This is in spite of the fact that, when compared with conifers, the prospects of quicker financial returns on a breeding

investment are high due to earlier and abundant production of seed and shorter plantation rotations (Eldridge, 1975).

In Florida, eucalypts were first planted in 1878, but

industrial plantations were not established until 1972 (Geary et al., 1983). Through cooperative research involving government, private land owners, and industry, approximately 6,000 hectares of Eucalyptus plantations existed in southern Florida by 1981. Eucalyptus grandis (Hill) ex. Maiden, one of the prime species, is recommended



1






2


for the palmetto prairie (poor fertility) and acid flatwoods (Geary et al .,1983).

In Florida there are three advantages of using E. grandis over other hardwood species for biomass production. First, E. grandis has an exceptional growth rate. For six progenies of E. grandis at 2.5 years of age on organic soil, Rockwood and Geary (1982) reported a mean height of 10.2 m and a diameter at breast height (DBH) of 10.6 cm. This surpasses the impressive growth of E. saligna reported by Walters (1980) in Hawaii. Second, E. grandis performs better in low quality soils when compared to other native hardwoods. A mean height of 4.7 m and DBH of 3.3 cms was achieved at 17 months of age on a sandy palmetto prairie soil (Rockwood and Geary, 1982). Although this growth is significantly lower than those

achieved on organic sites, it is impressive when compared with the growth of other hardwoods on similar sites. Third, the ability of E. grandis to coppice makes it an attractive species for short rotation biomass production. In much of the world, E. grandis sprouts prolifically, and in short rotation plantations two to four coppice crops can be

achieved before replanting is necessary (FAO, 1979). Both for the extent to which it has been planted around the world and for its excellent performance E. grandis is one of the most important exotic eucalypts.

Depending on the natural distribution, variation within

Eucalyptus species may be either clinal or ecotypic






3


(FAO, 1979). Eucalyptus grandis has a limited distribution,

which sometimes consists of isolated provenances adapted to specific environmental conditions. Natural occurrence of E. grandis is in northern New South Wales and southern Queensland.

Eucalyptus grandis in Florida constitutes a landrace

developed through three generations of selection (Geary et. al., 1983) and progeny testing in the local environment (Figure 1). The initial introductions along with local selections were planted in Immokalee, Florida, and later converted into a seed orchard. Selections from this orchard plus new introductions were tested in Silver Lake Prairie and constitute the second generation population. Subsequently, this plantation was converted to a seed orchard. Selections from Silver Lake Prairie seed orchard and additional new introductions constituted the third generation population (GPOP73) established in 1973 near LaBelle. GPOP73 was later rogued to form a seed orchard. Seed from GPOP73 seed orchard and new introductions comprise

the current genetic base population (GPOP77).







4


INTRODUCTIONS 'LOCAL SELECTIONS



IMMOKALEEj, IMMOKALEE SEED ORCHARDS INTRODUCTIONS



SILVER LAKE PRAIRIE SILVER LAKE PRAIRI SEED ORCHARD INTRODUCTIONS



GPOP73 GPOP73 SEED ORCHARD INTRODUCTIONS



GPOP77







Figure 1. Development of the current Eucalyptus grandis
breeding population (GPOP77) in southern Florida.
















OBJECTIVES


This study had the following objectives:


(1) To calculate the realized gains achieved through

four generations of selection,

(2) To study the effect of inbreeding through mating

of related trees in a future seed orchard,

(3) To predict the genetic gains, adjusted for

inbreeding, in the next generation for different

selection strategies,

(4) To estimate the amount of dominance variation

in growth traits,

(5) To develop a selection index to aid in

selection of individuals based on a combined

performance for seedling height, coppice height,

coppice volume, frost resilience and coppice

quality.











5















LITERATURE REVIEW


Provenance Variation: Several trials of E. grandis

established around the world have revealed significant provenance variation. Trials in Brazil have shown that for both height and DBH, the Australian origin showed a larger variability than both South African and Brazilian origins (Assis and Brune, 1983). In Australia there were significant differences in productivity between provenances, with those from low altitude sites in southeastern Queensland and the Coff's Harbour areas performing well (Ades and Burgess, 1983). Provenance trials in South Africa

have shown superior growth at four years of age at all sites for seedlots from New South Wales and Queensland (Darrow and Roeder, 1983). In comparison, South African seed orchard stock showed relatively poor growth but superior form. King (1983a, 1983b) reported the establishment of E. grandis provenance trials in Hawaii and California. In California,

sources from higher elevations in Australia had the best growth and survival.








6






7

Genetic Improvement Through Recurrent Selection


Appreciable genetic variation for growth has been observed in E. grandis, and successful selection programs for growth rate have been reported. Selection of the best genotypes is important in increasing E. grandis productivity. Differences among families were observed to be highly significant, as the best families produced as much as 50 percent more volume than the average of the families studied (Rockwood and Meskimen, 1981). In Florida, at 2.5 years of

age, a gain of 163 percent in individual tree volume has been achieved in four generations of selection (Meskimen, 1983). Because of recurrent selection for local adaptation, the trees perform better than do progenies of outstanding trees selected in Australia, South Africa or elsewhere.

Progeny tests in Brazil using 124 open-pollinated

mother trees of E. grandis have shown high heritabilites for growth traits justifying improvement programs based on selection (Borges and Brune, 1983). Similar results were reported from Zimbabwe where in a combined analysis

between-family differences at 18 months indicated considerable diversity for height (Mullin and Gough, 1983).

Research reported on the genetic variation for E.

grandis coppicing ability is limited. In southern Florida, evaluation of four progenies at two different sites showed no significant differences in coppicing ability (Rockwood and Geary, 1982). However, Geary et al. (1983) reported a






8


progeny test that revealed significant genetic variation in coppicing ability.

Resistance to Frost: Species of Eucalyptus do not have annual resting buds. Indeterminate shoots grow continuously as long as favorable conditions exist and there is no requirement for rest or overwintering (FAO, 1979). Of the many Eucalyptus species evaluated in Florida for frost tolerance, E. grandis was notably frost sensitive (Hunt and Zobel, 1978). Limited information is available on genetic variation in frost resilience in E. grandis.

In their native Australia Eucalyptus species grow under a wide range of climatic conditions. Some species grow in regions of freezing temperature and frequent snow (Hunt and Zobel, 1978). In Australia the temperatures gradually decline to mid-winter. A severe frost is always preceded by a series of cold nights which harden the species (Burgess, 1983). In southeastern United States, on the other hand, the temperature drops suddenly from well above freezing to well below (Hunt and Zobel, 1978). Florida's winter freezes have a devastating effect on the survival of E. grandis, where swift moving cold fronts turn balmy afternoons into freezing mornings, severely damaging the tender eucalypts (Geary et al., 1983). There seems to be a positive correlation between growth rate and freeze tolerance. Hunt and Zobel (1978) found the fast growing progenies of E. viminalis not damaged by severe freeze in 1976-77. However, Jahromi (1983) tested several E. viminalis provenances in






9

southern Georgia and northern Florida and found them not suitable for commercial planting. Jahromi concluded that a

genetic improvement program can produce new strains of eucalypts which would be more frost hardy. Geary et al., (1983) report that the bigger the tree at the time of the freeze, the more resistant it is to frost.


Inbreeding


Forest tree seed orchards are managed and designed with the main objective of providing a source of genetically high. quality seed. The normal breeding system for trees is

cross-fertilization as a result of insect or wind pollination. Consequently, seed orchards are designed to

promote the opportunities to maximize outcrossing (van Buijtenen, 1971). The essential consequence of two individuals having a common ancestor is that they may both carry replicates of one of the genes present in the

ancestor, and by mating they pass on these replicates to the offspring (Falconer, 1981). Thus, inbreeding increases the

proportion of homozygous genes in a population and decreases the proportion of heterozygous genes.

In a large random mating population, the probabilty

that two allelles in an individual are identical by descent approaches zero and the population is fully outbred (F= 0). Whenever there is a tendency to assortative mating between relatives (e.g., sib-crossing or at the extreme self-pollination), the inbreeding coefficient will depart






10


from zero. Computation of inbreeding coefficient therefore requires no more than tracing the pedigree back to the common ancestor and computing the probabilities at each

segregation or level (Li, 1976). The i nbreedi ng coeffi ci ent after one generation is 0.50, 0.25 and 0.125 for selfing,

full-sib and half-sib mating, respectively (Falconer, 1981).

Among normally outcrossing species any degree of

relatedness between individuals that mate leads to inbreeding depression. As pointed out earlier, selfing is the most severe form of inbreeding and is of interest to the.

tree geneticists because it often results in the inbreeding depression of economically important traits. Selfing is of special interest where clonal orchards are used as a means of mass producing seed of superior quality. In such orchards the potential for selfing is magnified because selfs can be produced not only from selfing of a ramet but also from mating of ramets of the same clone (Squillace and Kraus, 1963).

In recent years research has been done on pollination

biology, amount of selfing and inbreeding depressi on in E. grandis. Eldridge (1976) suggested that if inbreeding is prevalent in eucalypts, reduced improvement will follow the usual breeding methods. Working on the breeding system with several species of eucalypts, Pryor (1961) demonstrated a system of preferential outcrossing. His results showed that selfing does occur but less readily than outcrossing to other individuals. Other studies also found similar results






12


crossed families and germination rate was slightly lower among selfed families. In P. elliottii, self-pollination produced 10% as many germinated seedlings per cone as wind pollination (Snyder, 1968). Fowler (1964) working with P. resinosa Ait. found that seedlings resulting from

self-pollination exhibit little or no inbreeding depression.

Squillace and Kraus (1962) compared the effects of

several degrees of inbreeding (selfing, backcrossing, full-sib crossing and half-sib crossing) against outcrossing, including polycrossing for several characteristics in P. elliottii. Cone yield relative to number of flowers pollinated was not affected. Sound seed yields per cone, on the average, were lower for inbred trees than for trees that were outbred. By regression Squillace and Kraus (1962) also estimated that an increase of 0.1 in inbreeding coefficient resulted in a decrease of about 5 sound seeds per cone. The same increase in F was associated

with 8% decrease in germination and 7% decrease in rate of germination. Growth was decreased by approximately 4% over the mean of the outcrossed individuals. These results were in close agreement with the results obtained by Gansel

(1971) who concluded that inbreeding depression was evident in several traits including diameter, volume and oleoresin yield and was proportional to the degree of inbreeding. Snyder (1972) found that the offspring of selfed








other individuals. Other studies also found similar results of predominant outcrossing for the genus Eucalyptus as a whole. These studies confirm the occurrence of a wide range of self-fertility within species, including two extremes of obligate selfing caused by cleistogamy (Venkatesh et al., 1973) and obligate outcrossing by self incompatability (Pryor, 1957) or male sterility (Davis, 1969), or female sterility (Carr and Carr, 1972).
The amount of natural selfing that occurs in eucalypts is higher than that reported for pines. An average of 7% selfing is reported for most pines (Wright, 1976). Published estimates for the degree of natural selfing in eucalypts vary with species: 24% in E. obliqua (Brown et al., 1975), 37% in E. pauciflora (Phillips and Brown, 1977), 23% in E. delegatensis (Moran and Brown, 1980) and 18% in E. stoatei (Hopper and Moran, 1981). In E. grandis Eldridge (1978) reported 20% to 40% selfing to occur and Van Wyk (1981) estimated this average to be about 30%.

The major consequences of inbreeding depression
resulting from natural self-pollination in seed orchards of P. taeda L. is reduced seed yield (Franklin, 1969). Franklin found the total seed yields to be slightly lower,

yields of filled seed to be drastically lower and average weight of filled seed to be slightly lower after selfing. Percent of filled seed was 74.2 from crossed families and
14.3 from selfed families, a five-fold reduction.






13


P. elliottii were 21% shorter than wind-pollinated offspring at one year and 34% shorter after 5 years.

In E. grandis Hodgson (1974, 1976, 1977) studied the extent of inbreeding for inbred individuals. The height reached by selfed progeny of nearly all clones of E. grandis at 11-18 months of age was 8% to 49% less than that of outcrossed progeny. In addition, selfed progeny was usually more crooked and included about 30% abnormal seedlings. The

study also found self-pollination and self-fertilization treatments adversely affected seed yield and stem form leading to the conclusion that products from a seed orchard in E. grandis may be subject to an average of 30% degrade as a result of selfing. Thus there is an opportunity to

identify and control at least some of the worst affects of selfing in E. grandis seed orchards by culling poorer seedlings in the nursery and by testing and rejecting the more self-fertile clones (Eldridge, 1978). However, in E. regnans the selfed seedlings did not differ in appearance

and were not consistently smaller than outcrossed or open-pollinated seedlings up to one year of age (Eldridge, 1970). In this study inbreeding depression was

not obvious until six years after planting and even then many individuals within selfed families were vigorous and of good form. Eldridge (1978) concluded that in E. grandis many times more seeds were set after crossing than after selfing or open-pollination, whereas in E. regnans selfing,






14


crossing and open-pollination yielded only a small range in numbers of seed per capsule.

In spite of using large numbers of clones and advanced planting designs, which theoretically ensure that trees of

the same clone or related families are planted as far apart as possible in a seed orchard, the risk of self-pollination

can only be reduced but not eliminated. Thus any predicted genetic gains should take into account the losses due to inbreeding depression resulting from selfing and mating of related individuals.


Diallel Analysis


The making of all possible crosses among a set of

parents has been a popular mating scheme for over 40 years. The concept of combining ability is becoming increasingly

important in tree improvement. It is especially useful in testing procedures in which it is desired to study and compare the performances of lines. According to Sprague and Tatum (1942), the term general combining ability (GCA) is

used to designate the average performance of a line when crossed to different lines. The term specific combining ability (SCA) is used to designate those cases in which certain combinations do relatively better or worse than

would be expected on the basis of the average performance of the lines involved.

In the analysis of data from a diallel crosses, the average performance of each progeny is divided into






15


(interaction). If the SCA mean square is not significant, one would accept the hypothesis that the performance of a single cross progeny can adequately be predicted on the basis of GCA. The best performing progeny may be produced by crossing the two parents having the highest GCA. If, on the other hand, the SCA mean square is significant one needs to ask how important the interactions are in determining the performance of single cross progeny. On the basis of the above discussion it can then be generalized that GCA depends on additive genetic variance. A good general combiner consistently produces progeny which do well when crossed with the other parents in the group. On the other hand specific combining ability depends on dominance variance. A

good specific combiner produces good progeny only when crossed to certain parents in the group.

In E. grandis Van Wyk (1976, 1977) found considerable additive as well as non-additive genetic variance. His estimates of GCA for most growth traits were significant,

suggesting a selective breeding program utilizing additive genetic variance will be effective in improving the population mean.



Selection Indices


A tree breeder wishes to improve more than one

character or trait in his breeding work. Tandem selection appears impractical because of the long time span involved.

Independent culling levels entail selecting for all the






16


Independent culling levels entail selecting for all the characters at the same time but independently, rejecting all individuals that fail to come up with a certain standard for

each character regardless of their values for other traits of interest. The main problem, however, is that if one

wishes to improve a large number of traits, a very large population must be screened.

The best method, from the theoretical point of view, to deal with this problem is the use of the selection index method. A selection index (I) is a linear function of the phenotypic value of trait (Xi) weighted by index weights

(bi) in such a way that maximum correlation with the economic breeding value is obtained. In other words, the index is the best linear prediction of an individual's

breeding value and it takes the form of a multiple regression of the breeding values on all the sources of information (Falconer, 1981).

Use of selection index is expected to give the most

rapid improvement in breeding value for economic merit. It involves the application of selection simultaneously to all the component traits, appropriate weights being given to each trait based on its relative importance, its heritability and the genotypic and phenotypic correlations between the different traits (van Vleck, 1981). So the

multiple trait selection takes into account a combination of traits according to their relative importance. Each






17


individual in the population to be selected from will have an index value (score). It is possible that the highest yielding individual will not get the highest score.

Although the use of selection indices in tree breeding has been limited, plant breeders have successfully utilized this procedure as an important tool in selection programs. In most selection procedures breeders use an intuitive selection index and the index value is based on the

breeder's decision.

The final index value of each individual is of the form

I = b1P1 + b2 P2 + .... + bP


where the big's are the unknown factors by which each measurement- of a trait (Pi) is to be weighted. The problem in constructing the index is to find the best value for each weighting factor.

The optimum selection index was first used by Smith (1936) in plants and later by Hazel and Lush (1942) in animals. They concluded that the superiority of an index increases with increasing number of traits under selection, but decreases with increasing differences in relative importance (i.e., the superiority of the selection index as a selection procedure is at a maximum when the traits of interest are considered equally important). Gain from selection for any given trait is expected to decrease as additional traits are included in the index, so the choice

of traits to be included in the model must be made






18


objectively (Hallauer and Miranda Fo, 1981). In addition, precision of the variance and covariance estimates is generally low, thus limiting greater use of the selection index. The genotypic value of an individual considering several traits is is of the following form


Hij = alGlj + a2G2j + .... + aiGij

where ai is the relative weight and Gij is the genotypic value of the j individual for trait i. The objective is to make the selection so that the above combination of characters (H.) is the most desirable in terms of the economic value. Because Gij are not known but are estimated from P..j (phenotypic values), Hij must be evaluated by an index I, based on the phenotypi c values such that the correlation between H and I is as great as possible. The maximization of this correlation leads to the following relationship (Turner and Young, 1969)


GA = PB

where G is the matrix of additive genotypic variances and covariances, A is the vector of relative economic weights, P is the matrix of phenotypic variances and covariances and B is a vector of unknown values of b. 's. The solution of the set of equations leads to the estimation of values of bi that give the highest correlation with Hij when applied to index I (Hallauer and Miranda Fo, 1981).






19


The genotypic and phenotypic variances and covariances of all the traits are needed for the estimation of these coefficients. Harris (1964) used approximate formulae and computer simulation to show that the predicted gains from selection for an index based on variance and covariance estimates tends to exceed the expected value of the gain. Williams (1962) calculated the probability of correct selection based on an index involving sample estimates of variance and covariance. He concluded that in cases in which variance and covariance estimates are poor, the index may not be highly correlated with true genetic values. So the less reliable the estimates of variances and covariances, the less reliable are the indices.

The deduction leading to the superiority of indices

over other methods of selection and the correctness of the estimated superiority of one index over another are dependent on having correct values for the genotypic and phenotypic variances and covariances and the relative economic weights. Since advances accomplished through

selection are generally small, the breeder is forced to select towards a predicted goal. This problem of goal is not peculiar to index selection but to breeding programs in general.















MATERIALS AND METHODS



The E. grandis base population was planted near LaBelle, Florida, in July 1977. The U.S.D.A. Forest

Service, Southeastern Forest Experiment Station, installed and maintained the experimental materials through September 1984. GPOP77 contains 529 open-pollinated families representing four generations of selection (Table 1). One hundred and forty-four families (27%) are first-generation, 211 (40%) are second-generation, 126 (24%) are third-generation and 48 (9%) are fourth-generation. Geographically 69% of the families directly trace their origin to Australia (37% from New South Wales and 32% from Queensland). Twenty-one % came from South Africa, 4% from other nations and 6% of the families cannot be traced beyond Florida (Table 1). The pedigrees and origins of the 529 families in GPOP77 is given in Appendix 1.

Each family typically is represented 60 times in a completely randomized single tree plot design on 17.3 hectares for a total of 31,725 trees. Stocking of the plantation is 1,916 trees per hectare (775 trees per acre) laid out in a precise 141 rows X 225 columns grid mapped to preserve the location and pedigree of each tree. Planting

beds are paired, with 2.3 m spacing within a pair 20






21


Table 1. Eucalyptus grandis progenies in GPOP77
by generation and origin.





Generation


1 2 3 4 Total


Progeny

Number 144 211 126 48 529 percent 27 40 24 9 100

Origin


NSW QLD SA Other Unknown Total


Progeny

Number 196 169 111 21 32 529 Percent 37 32 21 4 6 100


* NSW = New South Wales, Australia
QLD = Queensland, Australia
SA = South Africa
Unknown = Mainly south Florida
Other = Other nations






22


and 3.5 m (11.5 ft.) between pairs. The seedlings were

planted 1.8 m (6 ft.) apart along the beds.

The southern half of the plantation (141 rows and

columns 116 through 225) were harvested in August 1978 to

obtain coppice information. The results reported in this study pertain to the southern half of GPOP77 which includes all 529 families with an average of 30 individuals per family for a total of 15,510 trees. A summary of activities

in GPOP77 is given in Figure 2 and the details of the measurements obtained are summarized in Appendix 2.


Study Area for Diallel Analyses


The study area for the diallel analyses is located on a adjacent site near LaBelle, Florida. This E. grandis population (GPOP73) was planted in the summer of 1973. The completely randomized single tree plot design contained 181 families or provenance collections either as imported seed lots from Africa and Australia or as seed lots from selected trees from Silver Lake Prairie orchard. In addition, 116 controlled-pollinated families in a 15-tree diallel were

established. However, due to mortality after planting and failure to coppice after harvest, a 10 X 10 half-diallel was used including only the lines that are common to GPOP77.

The diallel analysis were performed on four traits:






23


Date Activity 7/1977 GPOP77 Planted 2/1978 7-Month Seedling Height 8/1978 South-half Felled 11/1978 3-Month Coppice Height 3/1981 30-Month Coppice Height, DBH, No. of Stems


1/1982 Severe Freeze 12/1983 64-Month Coppice Height, DBH, No. Stems, Frost Response, Coppice Quality





Figure 2. Summary of activites and measurements
for Eucalyptus grandis in GPOP77.






24


height at 6 months, height and DBH at 2.4 years and coppice

height at 5 months after harvest at age 3.3 years.



Statistical Analyses


The data were analyzed using the Statistical Analysis System (SAS). Since the data were unbalanced, the variance components were estimated using the general linear model (GLM) procedure. The analysis computes the Type 1 sums of

squares for each effect, equates each mean square involving only random effects to its expected value, and solves the resulting system of equations (Gaylor et al., 1970). The following random effects model was used in the analysis:

Y = U + Gi + F(G) + Eij
ijk 1 j(i) ijk where


Yijk = individual observation

U = overall mean

Gi = effect of the ith generation

F(G)j(i)= effect of the jth family within

the i generation

Eijk = random error associated with each

observation

Individual and family tree heritabilities were

estimated using the procedures and formulae outlined by Wright (1976). The selection intensities used to estimate the predicted genetic gains were obtained from Namkoong and






25

Snyder (1968). Since the coppice and frost codes were non-continuous tests were performed to see if their

distribution approximates a normal distribution.

Some families in GPOP77 have common ancestors, thereby

leaving a potential for the occurrence of inbreeding due to mating of related individuals. Genetic gains through alternative selection strategies, were adjusted for inbreeding depression that may occur due to the presence of

relatives. For each selection strategy, the inbreeding coefficient (F) was calculated for all possible matings by tracing the pedigree of each mating to its common ancestor

and computing the probabilities at each level (Li, 1976). Representative pedigrees and inbreeding coefficients are shown in Figure 3.

The mathematical model used in the diallel analysis for computing the general and specific combining abilities is as follows:


Z = U + Gi + G. + Si + E.



where

Zij= mean performance of the progeny of ith

parental line mated with jth parental line
U = overall mean

G(i) G(j)= general combining ability of the
ith (jth) parental line






26

S..ij = interaction of the ith and jth

parental lines (specific combining abilities)

E.. = random error

The diallel analyses were performed according to the

procedures outlined by Becker (1984). The means of crosses

were used in the analysis of variance calculations. The error variance for the ANOVA based on cross means (V ) was obtained by dividing the error variance based on individual observations (Ve) by the harmonic mean number of individuals per cross. The analysis of variance table and the expected mean squares are given in Table 2.

Estimation of Selection Indices: A subset of GPOP77 including only the trees that were alive at coppice age of 64 months was used to estimate variances and covariances of

the five traits. The selection index derived was of the following form:



I = blShgt + b2Chgt + b3Vol + b4FR + b5CQ

where,

bi = unknown weights

Shgt = Seedling height at 7 months

Chgt = Coppice height at 3 months Vol = Coppice volume at 64 months

FR = Frost resilience score at 64 months

CQ = Coppice quality score at 64 months






27












INBREEDING COEFFICIENT
GENERATION
0.25 0.125 0.0625 0.03125 0.0156 0.0078


1 295 295 295 294 295 295



2 836 836 836 117 126 36 838 5 35 36 3 101 101 88 90 26 943 101 98 95 90 4 1011 1010 1011 1002 X 1028 991 011



5 x x x x X










Figure 3. Representative pedigrees and inbreeding
coefficients of different degrees of
relationships among top 50 families of
Eucalyptus grandis in GPOP77.






28


Table 2. Diallel analysis of variance for four Eucalyptus
grandis traits in GPOP73




Mean Squares

Source DF H6 D24 H24 C5 Expected Mean Squares


GCA 9 2.2 0.9 225.6 4.3 V +V ca+8V gca SCA 35 1.4 0.4 135.3 0.3 V +V
w sca
Error 942 1.1 0.2 43.9 0.2 V

H6 = Seedling height at 6 months D24 = Seedling DBH at 2.4 years H24 = Seedling height at 2.4 years C5 = Coppice height at 5 months






29

Table 3. Relative economic weights used in the ten
selection index equations for five
Eucalyptus grandis traits in GPOP77



seedling coppice coppice frost coppice
Index height height volume resilience quality
7-months 3-months 64-months


1 0.35 0.75 0.75 0.75 0.35 2 0.32 0.50 0.50 0.50 1.00 3 0.50 1.00 1.00 1.00 0.50 4 0.50 0.50 0.50 0.50 0.50 5 0.10 0.60 0.60 0.60 0.60 6 0.75 0.35 0.75 0.35 0.75 7 0.50 0.32 0.10 0.50 0.50 8 0.50 0.10 0.50 0.10 0.10 9 0.00 0.00 0.54 0.25 0.25

10 0.60 0.60 0.10 0.60 0.60






30



The variances and covariances were estimated using the the procedures outlined by Van Vleck (1981). The relative economic weights for the five traits used in the selection

index were based on their relative importance for biomass production in Florida (Table 3).















RESULTS AND DISCUSSION


Genetic Improvement Through Recurrent Selection

Four generations of selection have achieved significant genetic improvement for all coppice growth characteristics studied. Comparisons of means of height, DBH and volume showed impressive gains at all ages with generations of improvement. At 7 months after planting the mean height for first-, second-, third- and fourth-generation trees was

1.60, 1.79, 1.81 and 1.83 m, respectively (Table 4). At the same age survival was similar across generations. The same trend continued for later ages with the fourth generation families performing the best. At 64 months of age the average first-generation tree contained 7.04 dm3 of stem wood. Second-generation trees were 206% larger averaging

21.54 dm3 , third-generation trees were an additional 20% larger at 25.91 dm3 and the fourth-generation trees had yet another 55% gain with an average of 40.16 dm3 of wood (Table 5). Mean coppice height at 64-months for the four

generations is given in Figure 4.








31






32


Table 4. Eucalyptus grandis generation means for individual
tree performances in GPOP77 through 64 months
after harvest.


Generati on

200 Best
Rotation/ Overall 1 2 3 4 Trees
Trait/Age


Seedl i ng
-7 months
Height (m) 1.75 1.60 1.79 1.81 1.83 2.93 Coppi ce
-3 months
Height (m) 0.63 0.56 0.63 0.67 0.71 1.69

-30 months
DBH (cm) 3.32 2.46 3.37 3.74 4.19 8.13 Height (m) 4.30 3.43 4.34 4.71 5.28 7.09 Dry Weight 0.87 0.29 1.02 1.24 2.00 4.16
(kg)

Volume (dm3) 2.58 1.56 2.85 3.24 4.58 8.39

-64 months
Height (m) 7.36 5.00 7.75 8.42 9.39 16.79 DBH (cm) 6.66 4.03 7.09 7.75 9.18 17.49 Dry Weight 10.88 3.39 11.60 14.08 20.10 73.13
(kg)

Volume(dm3) 20.16 7.04 21.54 25.91 40.16 130.25 Frost 1.47 1.75 1.39 1.34 1.25 0.00
Resi li ence

Frost Defect 1.43 1.05 1.49 1.64 1.76 0.00 Coppice 0.54 0.77 0.48 0.41 0.40 0.00
Quality

Coppice 0.99 1.28 0.92 0.83 0.79 0.00
Defect





33







10 9 8


7


- 6


5

I 4





2 1

0 I
1 2 3 4 GENERATION a 7-MON + 3-MON * 30-MON A 64-MON



Figure 4. Mean seedling (7-month) and coppice height
(3-, 30- and 64-month) for four generations
of Eucalyptus grandis progenies in GPOP77.






34


Percent gains achieved in each generation over the preceding generation is given In Table 5. The rate of

improvement achieved over the preceding generation decreased with generations of improvement. For the traits studied the

highest gain was achieved in the second generation over the first . This seems to indicate that the amount genetic variation has declined over generations which is reflected in the declining realized gains. For 64-month coppice volume, the second-generation trees had a threefold gain over the first-generation. This gain was considerably lower"

for later generations.

Per hectare biomass productivity differed significantly with generations with the fourth-generation families producing the most. However, the productivity was due to

low survival resulting from harvest in August (Rockwood et al., 1984). Since survival was similar for all generations at 64 months, per hectare biomass yield differences were

essentially the same as individual tree volume and weight differences. At 64 months after harvest the first-generation trees produced 0.7 dry mt of stem and branches , second-generation trees had 2.5 dry mt, third generation trees yielded 3.0 dry mt and the fourth-generation trees produced 4.3 mt per hectare per year. The effective density after accounting mortality at 64 months after harvest was 1190 tree per hectare. A projected yield of 25.7 dry mt/ha/yr for the 200 best trees





35


Table 5. Percent gains for Eucalyptus grandis generations

compared to the preceding generation for different

traits





Generation

200 best

Trait Age 2 3 4 Trees (mos)


Seedling height 7 12 1 1 60 Coppice height 3 13 6 6 138 Coppice height 30 27 9 2 34 Coppice DBH 30 37 11 2 94 Coppice volume 30 83 13 2 83 Coppice height 64 55 9 2 79 Coppice DBH 64 76 9 9 91 Coppice volume 64 206 20 5 224






36


under the assumption of 100% survival is unrealistic, but it

does identify the potential of clonal propagation of trees with excellent coppice growth.

Variation among sources: Geographic origin of

E. grandis may be important in developing good coppicing trees for southern Florida. For early seedling growth sources were generally similar in each generation. Introductions from South Africa (crosses among selected parents) were slightly taller when first grown in Florida (Table 6), but their selected offspring were virtually the same height as other sources in the second generation. For 64-month coppice height, Queensland sources improved considerably with each generation of selection beyond the initial introduction and surpassed third- and fourth-generation means for other sources by more than 1 meter (Figure 5). The results indicate that greater gains could be achieved by concentrating the future introductions to Queensland regions of Australia.

Additive genetic variation: High individual and family heritabilities were observed for all the traits examined. The analysis of variance tables for the growth traits are given in Appendix 3. For 7-month seedling height, individual tree heritability was twice that of estimates reported by van Wyk (1977) for E. grandis (Table 7). High individual tree heritabilities suggest that high genetic gains can be achieved through mass selection of the best individuals. The overall individual tree heritabilities






37


Table 6. Comparison of origin of Eucalyptus grandi s
progenies in GPOP77




Ori gin

Trait Gen. NSW QLD SA Other Unknown


7-mo 1 1.6bB 1.6aB 1.7aA 1.6aB Seedling
Height 2 1.8aA 1.8bA 1.8aA 1.8bA 1.7B
(m)
3 1.8aA 1.7bA - 1.8bA

4 1.8aA 1.8bA - -

64-mo 1 5.2aA 4.9aA 4.9aA 5.4aA Coppice
Height 2 7.8bA 8.2bA 7.5bA 7.7bA 7.6A
(m)
3 8.4cB 9.8cA - 8.4bB

4 7.4dB 9.9cA - -


* NSW = New South Wales, QLD = Queensland, SA = South Africa, Other = Mainly south Florida
** Origins within generations not sharing same upper case
letter are significantly different at 5%.
Generations within an origin not sharing the same lower
case letter are significantly different at 5% level.





38







10


^' 9


1 8
I
z I0

FIgu 5 cpe i N
6Q S


I 5

I 4




0
0 2
z

S1



NSW QLD SA OTHER SOURCE OF ORIGIN I GEN I GEN 2 G GEN3 ~ GEN 4



Figure 5. Mean coppice height at 64 months by generation
for four sources of Eucalyptus grandis
progenies in GPOP77 (NSW=New South Wales,
QLD=Queensland, SA=South Africa,
Other=Mainly from southern Florida)






39


Table 7. Heritabilities of Eucalyptus grandis traits by
and over generations




Individual Heritability Family Heritabilty

Generation Generation

Trait 1 2 3 4 Overall 1 2 3 4 Overall


Seedling Hei ght 0.31 0.71 (7-mon)

Coppice Height 0.30 0.68 (3-mon)

Coppice Height 0.63 0.31 0.29 0.38 0.39 0.75 0.59 0.58 0.64 0.76 (64-mon)

Coppice DBH 0.57 0.33 0.29 0.33 0.39 0.73 0.61 0.58 0.61 0.76 (64-mon)

Coppice Volume 0.37 0.27 0.25 0.25 0.31 0.63 0.56 0.54 0.54 0.59 (64-mon)






40


through 64 month of age after harvest ranged from 0.31 to

0.39 and the family heritabilities ranged from 0.65 to 0.75. When calculated separately for each generation, the first generation trees have higher heritabilities compared to the other three. This explains the high genetic gains that were

realized in the second generation over the first generation. Indications are that enough additive genetic variation exists in E. grandis to yield substantial improvement in biomass productivity in a recurrent selection program.

Genetic correlations among coppice growth traits were high (Table 8). The correlation between seedling height at

7 months and coppice height at 3 months was negative but not significant. This indicates that early coppice height cannot be predicted based on early seedling height. The genetic correlation between coppice height at 3 months and coppice height and DBH at 64 months was high, suggesting that early coppice growth could be a good indicator of coppice growth at later ages.

One of the assumptions for the analysis of variance is that the errors are normally distributed. The frost and coppice scores in this study were tested against a normal distribution with mean and variance equal to the sample mean and variance (SAS, 1985). The normality tests on frost and coppice scores revealed that the distribution approximates a normal distribution and therefore can be analyzed as if they were continuous.






41

Table 8. Genetic correlations and their standard errors (in
parentheses) among growth traits, frost
and coppice scores of Eucalyptus grandis in
GPOP77





Trait


Shgt Chgt H64 D64 FD CQ CD

Shgt -0.27 0.26 0.25
(0.013) (0.012) (0.013)

Chgt 0.56 0.56 (0.007) (0.007)
H64 0.97 (0.001)
FR 0.20 0.81 0.80 (0.016) (0.005) (0.005)
FD -0.40 -0.54 (0.017) (0.013)
CQ 0.80
(0.007)
* Shgt = Seedling height at 7 months
Chgt = Coppice height at 3 months
H64 = Coppice height at 64 months
D64 = Coppice DBH at 64 months
FR = Frost resilience score
FD = Frost defect score
CQ = Coppice quality score
CD = Coppice defect score






42

Variation in Frost and Coppice Scores


Variation in frost scores:Considerable genetic

variation was observed for resilience to frost as well as

defects in coppice attributed to frost. The mean frost resilience score based on individual trees was 1.75, 1.39,

1.34 and 1.25 for first-, second-, third- and fourth-generation trees, respectively. However, for frost defect scores the fourth-generation trees scored the worst and the first generation trees the best (Table 4). The

genetic correlation between frost resilience and frost defect scores was low (Table 8). Thus selecting individuals for frost resilience would not improve the frost defect score. Comparison of trees from different geographical sources showed that trees from New South Wales were the most frost resilient, whereas trees from South Africa scored best for frost defect score (Table 9). Therefore by

concentrating future introductions from New South Wales one would be able to select for higher frost resilient trees.

Individual tree heritabilities for frost resilience and frost defect scores were 0.29 and 0.15, respectively. These

low heritabilites suggest low gains for frost defect scores can be expected from mass selection. The family heritabilities for frost resilience and frost defect scores were 0.71 and 0.51, respectively, suggesting that higher gains can be achieved through combined selection when compared to mass selection. However, highest gains could be achieved through combined selection of the best trees






43


Table 9. Mean frost and coppice scores of Eucalyptus
grandis by origin in GPOP77




Trait Origin

NSW QLD SA Other


Frost resi- 1.33 1.43 1.38 1.42
lience

Frost defects 1.65 1.40 1.36 1.42 Coppice 0.44 0.56 0.53 0.63 quality

Coppice 0.86 0.99 0.99 1.10 defects


* NSW = New South Wales
QLD = Queensland
SA = South Africa
Other = Mainly from southern Florida






44

from the best families. This would utilize both between-and

within-family variation thereby getting a higher response to sel ecti on.

Variation in coppice quality: Fourth-generation trees were also the best performers for coppice quality and coppice defect scores (Table 4). In terms of geographic origin, again the best performers were the sources from New South Wales (Table 9). Individual and family heritabilities for coppice quality and coppice defect scores were 0.12, 0.13 and 0.45, 0.46, respectively. Low individual tree

heritabilities suggest the ineffectiveness of mass selection for coppice quality. Higher family heritabilities indicate

the potential for improving coppice quality through family selection. Genetic correlation between the two coppice traits was 0.9. This high correlation suggests that by selecting and improving one trait, we would simultaneoulsy be improving the other. Genetic correlations between frost

and coppice scores are given in Table 8.


Inbreeding and Seed Orchard Development

GPOP77 could be converted into a seed orchard. Six different selection strategies including individual, combined and family selection with different intensities of selection assumed that initial introductions were not

related. For each selection strategy the inbreeding coefficient (F) was calculated for all possible matings







45


among the selected trees that constitute the selected population.

Predicted Inbreeding:Inbreeding coefficients among all possible crosses ranged from 0 to 0.25. The frequency of F values for a 50-family seed orchard is given in Table 10.

The mean inbreeding coefficient for five selection strategies ranged from 0 to 1% (Table 11). This was attributed to the fact that only a small percentage of

individuals (13% in a 50 family orchard) has an F greater than zero. In addition two or three generations have passedsince direct relationship by common ancestry. The presence

of related families in a seed orchard does not seem to affect the predicted gains through loss from inbreeding depression. However, when predicting the genetic gains, care should be taken to adjust the gains for potential inbreeding depression that may arise from selfing.

Extensive research has been done on the frequency of natural self-fertilization that occurs in pines. Franklin

(1971) working with Pinus taeda reported that an average of

2.4% of self-fertilized seed is produced. He used 25 marker-carrying trees and estimates for individual trees

ranged from 0 to 13.5%. Kraus and Squillace (1964) reported evidence of selfing in clonal orchards of Pinus elliottii. According to their observations total selfing is






46


Table 10. Frequency of different degrees of
relationship in a 50-family subset of
Eucalyptus grandis seed orchard
in GPOP77






Inbreeding coeffi ci ent (F) Frequency


0.250 2 0.125 13 0.063 6 0.031 93 0.016 19 0.008 28 0.000 1064 Mean 0.005 Total 1225






47


Table 11. Mean inbreeding coefficient (F) for five selection
strategies of Eucalyptus grandis in GPOP77




Number of Selected Families Mean F (%)


300 0.00 100 0.00

50 0.12 30 0.90

10 0.97





48

selfing is estimated to be 8% in a 20 clone seed orchard. Wright (1976) reported this figure to average about 7%. Squillace and Bingham (1958) recommended the elimination from seed orchards of individuals with high degree of

preference for selfing.

Percent selfing in E. grandis occurs at a higher rate than in most pines. In adjusting the predicted gains for selfing, Hodgson's (1976) and Eldridge's (1978) estimates of

inbreeding depression were used. They predicted a height loss of 8% to 49% for selfed individuals with an inbreeding coefficient of 0.50. Using the estimated 30% selfing that occurs in E. grandis, it can be deduced that 30% of the offspring from the seed orchard will have an F of 0.50 and

the remaining 70% will have the estimated F for a given strategy. *Thus for each selection strategy the weighted mean inbreeding was calculated for both sel fi ng and mating of relatives. Using the conservative approach of the range of inbreeding depression reported for E. grandis, a 50% loss in height growth for selfed progenies (F=0.5) amounts to 10% loss for every F of 0.10. So for a given selection strategy the predicted genetic gains were reduced by a proportion

equal to the mean inbreeding coefficient.

Predicted genetic gains: The predicted genetic gains (adjusted for inbreeding) that could be achieved through different selection strategies are given in Table 12.





49


Table 12. Predicted genetic gains in 64-month coppice volume
for alternative improvement strategies of
Eucalyptus grandis in GPOP77



Selection Strategy Genetic gain (%)


Mass Selection

200 best trees 80 300 best trees 69

Combined Selection

10 top families 41
(30 trees per family)

30 top families 61
(10 trees per family)

100 top families 90
(3 trees per family)

300 top families 86
(1 tree per family)

Vegetative Propagation

clonal propagation of 441
200 best trees

* Gain over the population mean






50


The six strategies give an effective orchard size of approximately 300 trees for an area of approximately 8.5 hectares. At 64 month coppice age, the highest gain (90%)

in volume over the population mean is predicted from a combined selection of the top 100 families with 3 trees per family. Comparable gains also are predicted for other

selection strategies: 86% for family selection of the top 300 families; 80% for mass selection of the top 200 trees. Prediction of higher gains through combined and family

selection compared to mass selection agrees with the results" reported for E. robusta in Florida (Dvorak et al., 1981).



Diallel Analyses


Results from the 10 X 10 half-diallel cross identified some of the parental lines with a high potential for genetic improvement. The 10 parents analyzed demonstrated considerable genetic variation for all the four traits studied. For seedling growth significant non-additive

genetic variance was found . On the other hand, for coppice height the additive variance was significant. Genetic improvement through the use of non-additive genetic variance seems to be advantageous for early seedling growth.

Height at 6 months: At 6 months after planting the

best cross (1188X1189) had a mean height of 18.4 dm, while the poorest performance was demonstrated by cross 1178 X 1180 (Table 13). Individual tree heritability was 0.02,






51


Table 13. Mean seedling height (dm) at 6 months for
45 crosses of Eucalyptus grandis in
GPOP73






Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 Total 198 14.2 15.6 14.6 14.4 14.8 16.9 14.9 12.8 14.6 132.8 1178 16.7 12.5 15.2 15.7 15.9 17.0 15.3 14.9 137.5 1179 13.4 14.9 14.9 13.8 17.2 12.7 14.7 133.9 1180 13.9 15.7 14.4 13.8 13.6 14.6 126.5 1182 13.9 15.0 14.1 14.7 14.2 130.1 1183 15.1 14.9 17.4 13.8 135.9 1185 16.1 14.8 15.1 133.8 1188 18.4 14.4 140.7 1189 14.1 133.8 1191 130.3






52


lower than that estimated by van Wyk (1976) for the same E. grandis at the same age. The variance due to GCA (0.099) was lower than due to SCA variance (1.387). This suggests that greater improvement could be achieved through the use of non-additive genetic variation like clonal propagation than from the traditional method of improving through mass selection. Considerable variation also was observed in the estimates of GCA among all parents. The GCA estimates and their standard deviations for the 10 parents are given Table 14. Family 1188 has the highest (0.8510) and family 1180 had the lowest (-0.9165). The 45 crosses from the

diallel exhibited considerable variation in SCA (Table 15). Cross 1179 X 1189 has the lowest SCA while cross 1188 X 1189

had the highest. The above results seem to indicate that parent 1188 is a very good general combiner and can be used as one of the parents in crosses to produce superior progeny. Cross 1188 X 1189 is the best cross among all the

45 crosses and can be used to produce better producing offspring for improving biomass production.

Height at 2.4 years: When compared with the performances of families for height at 6 months, considerable changes for the performances of parents as well as for crosses were observed for the same trait at age 2.4 years. Parents that ranked low for height at 6 months did relatively better and hence were ranked considerably higher for height at 2.4 years and vice- versa. This suggests that





53


Table 14. General combining ability (GCA) and associated
variances of 10 Eucalyptus grandis families
for seedling height at 6 months and coppice
height at five months in GPOP73



Seedling height Coppice height Family at 6 months at 5 months

GCA Variance GCA Variance

198 -0.132 -0.100 -0.23 0.03 1178 0.449 0.083 0.87 0.74 1179 0.000 -0.118 0.28 0.06 1180 -0.917 0.722 1.28 1.62 1182 -0.473 0.106 -0.25 0.04 1183 0.256 -0.052 -0.57 0.31 1185 0.420 0.059 -0.49 0.22 1188 0.851 0.607 -1.24 1.52 1189 0.003 -0.118 0.31 0.08 1191 -0.452 0.086 0.02 0.02






54


Table 15. Specific combining ability of 45 Eucalyptus
grandis crosses for seedling height at 6
months in GPOP73






Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191



198 -0.96 0.81 0.80 0.17 -0.25 0.77 -0.68 -1.90 0.27

1178 1.35 -1.93 0.30 0.13 0.19 0.81 -0.02 0.09 1179 -0.53 0.48 -0.17 -1.50 1.40 -2.20 0.26

1180 0.45 1.50 0.02 -1.05 -0.34 1.05 1182 -0.98 0.19 -1.17 0.34 0.20 1183 -0.48 -1.08 -0.30 -0.95 1185 -0.04 -0.46 0.27 1188 2.62 -0.89 1189 -0.30







55


height performances of families at early ages are not good indicators of height at later ages. Thus selection for

height at early ages may lead to elimination of better parents. Cross 198 X 1179 demonstrated the highest mean height at 2.4 years at 86.87 dm, and the lowest mean of 41.15 dm was observed for cross 1179 X 1183 (Table 16). Heritability was slightly higher (0.06) than at age 6 months

but the additive variance was not high enough to have a higher GCA variance than the SCA variance. The GCA variance was 11.29 compared to the SCA variance of 91.45. Family 1178 had the highest GCA (7.72) and parent 1182 had the lowest (-7.10) (Table 17). Table 18 summarizes the considerable variation observed in the SCA variances for the 45 crosses.

DBH at 2.4 years: The amount of variation in DBH at

2.4 years observed among the 45 crosses was high (Table 19).

The mean DBH among the crosses ranged from 3.74 cm for cross 1179 X 1183 to 6.56 cm for cross 1183 X 1189. As observed for height at early age the variance of SCA (0.26) was higher than the variance of GCA (0.06), suggesting again the use of non-additve genetic variation in the genetic improvement for DBH at age 2.4 years is more productive. The estimates of GCA and their variances for the 10 parents is given in Table 17. Line 1185 had the lowest GCA (-0.39) and line 1189 has the highest (0.47). The GCA rankings of parents for height at 6 months and DBH at 2.4 years changed considerably. Individual tree heritability was low which is






56


Table 16. Mean seedling height (dm) of 45 Eucalyptus
grandis crosses at 2.4 years in GPOP73




Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 Total 198 64.8 86.9 61.9 59.7 80.3 64.5 68.2 57.7 71.3 615.3 1178 83.3 67.3 75.0 70.8 71.9 87.0 68.0 77.6 665.6 1179 76.9 65.0 41.2 44.8 66.6 63.9 70.0 597.8 1180 67.3 83.7 67.7 73.2 62.1 70.0 629.9 1182 47.8 59.1 47.3 72.1 53.8 547.0 1183 49.2 47.6 84.9 54.6 559.9 1185 51.1 77.6 63.5 549.2 1188 92.9 67.1 600.9 1189 83.0 662.1 1191 610.2






57


Table 17. GCA and its variances of 10 Eucalyptus
grandis parents for seedling height and DBH at age 2.4 years in GPOP73







Height DBH



Parent GCA Variance GCA Variance

198 1.44 -2.86 -0.14 0.00 1178 7.72 54.66 0.43 0.17 1179 -0.75 -4.37 0.20 0.02 1180 3.26 5.69 0.22 0.03 1182 -7.10 45.48 -0.38 0.13 1183 -5.49 25.16 -0.23 0.04 1185 -6.82 41.58 -0.39 0.14 1188 -0.36 -4.81 -0.05 -0.01 1189 7.29 48.21 0.47 0.20 1191 0.80 -4.29 -0.03 -0.02






58


Table 18. Specific combining ability for seedling height
at 2.4 years for 45 Eucalyptus grandis
crosses in GPOP73






Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191



198 -11.5 19.6 -9.9 -1.7 17.2 2.8 0.1 -18.1 2.0

1178 9.2 -10.8 7.3 1.4 3.9 12.6 -7.9 8.3 1179 7.3 5.7 -19.7 -14.8 0.6 -9.8 2.4 1180 4.1 18.8 4.1 3.2 -15.5 -1.2 1182 -6.7 5.9 -12.3 4.9 -7.0 1183 -5.6 -13.6 16.0 -7.8 1185 -8.8 10.0 2.5 1188 18.8 -0.5 1189 7.7






59


Table 19. Mean seedling DBH (cm) of 45 Eucalyptus
grandis crosses in GPOP73 at age 2.4 years






Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 Total



-198 -4.87 6.25 4.84 4.30 5.54-4.75 5.04 4.56 5.04 45.21

1178 6.39 5.01 5.22 5.57 5.62 6.10 5.50 5.48 49.76 1179 5.94 5.39 3.74 4.32 5.34 5.09 5.42 47.88 1180 5.31 6.25 4.91 5.44 5.11 5.24 48.05 1182 4.07 4.55 3.96 5.27 4.31 42.38 1183 4.14 4.05 6.56 4.52 44.44 1185 4.40 5.57 4.91 43.17 1188 6.40 5.19 45.92 1189 5.96 50.02 1191 46.09







60

Table 20. Specific combining ability of 45 Eucalyptus
grandis crosses for 2.4-year seedling DBH
in GPOP73





Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 198 -0.57 1.04 -0.39 -0.22 0.76 0.13 0.08 -0.91 0.08

1178 0.61 -0.79 0.70 0.79 1.00 1.14 0.03 0.50 1179 0.38 0.54 -1.37 -0.63 0.04 -0.72 0.10 1180 0.44 1.12 -0.06 0.12 -0.72 -0.09 1182 -0.35 0.29 -0.65 0.15 -0.32 1183 -0.38 -0.82 1.18 -0.37 1185 -0.31 0.35 0.18 1188 0.84 0.12 1189 0.38






61


also indicated by the magnitude of the GCA variance. Thus

we cannot expect much improvement through selection as the amount of additive genetic variation present for this trait is limited. The SCA for the 45 crosses given in Table 20 shows that cross 1183 X 1189 has the highest SCA (1.18),

with cross 1179 X 1183 having the lowest (-1.37).

Coppice height at 5 months: Mean coppice height of the 45 crosses 5 months after roguing at 3.3 years ranged from

2.05 dm for cross 1185 X 1188 to 6.13 dm demonstrated by cross 1179 X 1180 (Table 21). The three parents that rankedhighest at 2.4-year seedling height were still the top three performers for coppice height at 5 months. Unlike the other three traits studied, the variance in coppice height at 5 months seems to be predominantly additive in nature. This is demonstrated by the high heritability (0.45) as well as

by the higher GCA variance (0.497) than the SCA variance (0.131). This indicates that if improvement in coppice growth is to be achieved, the use of additive genetic variance is more efficient, since the non-additve genetic variance for this trait is less. The GCA variance of the 10

parents ranged from -1.24 for parent 1188 to 1.28 for parent 1180 (Table 14). Cross 1178 X 1179 exhibited the highest SCA variance (0.91) and cross 1178 X 1188 had the lowest at

-1.01 (Table 22).

The results from the analyses of seedling traits

suggest that non-additive variance is a very important






62


Table 21. Mean coppice height (dm) of 45 Eucalyptus
grandis crosses at 5 months in GPOP73





Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 Total 198 4.02 4.41 4.56 3.56 2.73 3.28 2.80 3.45 3.56 32.37 1178 5.86 5.34 4.56 4.86 4.55 2.42 4.96 4.62 41.19 1179 6.13 3.28 2.78 3.40 2.13 3.56 4.87 36.42 1180 5.54 4.70 4.04 4.18 4.73 5.25 44.47 1182 2.60 3.02 2.43 4.08 3.15 32.22 1183 2.70 2.15 4.08 3.06 29.66 1185 2.05 4.06 3.18 30.28 1188 3.64 2.48 24.28

1189 4.15 36.71 1191 34.32






63


Table 22. Specific combining ability of 45 Eucalyptus
grandis crosses for 5-month coppice height
in GPUP73





Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191



198 -0.43 0.56 -0.30 0.24 -0.27 0.20 0.47 -0.44 -0.03

1178 0.91 -0.62 0.13 0.75 0.37 -1.01 -0.03 -0.07

1179 0.77 -0.55 -0.73 -0.19 -0.71 -0.83 0.78

1180 0.70 0.18 -0.55 0.34 -0.67 0.15 1182 -0.39 -0.04 0.12 0.19 -0.42 1183 -0.04 0.16 0.53 -0.19 1185 -0.02 0.44 -0.15 1188 0.77 -0.10 1189 0.02






64


source of variation that could be exploited for improving

seedling growth of E. grandis. Low heritabilities suggest the relatively low gains possible through mass selection.

This however is not the case for the early coppice height-; High heritabilities for this trait suggest good gains can be achieved through selective breeding. In a selection program that-utilizes additive genetic-variance, the variance component attributed to GCA (V gca) is an indicator of the amount of additive genetic variance. Table 23 shows the variance of SCA (Vsca ), variance of GCA (Vgca), the additive genetic variance (VA) and the dominance variance (VD). In this study Vgca was significant for DBH at 2.4 years and coppice height at 5 months and the Vsca was found to be significant for both height and DBH at 2.4 years. For all the traits except coppice height, Vgca was lower than Vsca. Consequently, the additive genetic variance was less than the dominance variance for these traits, suggesting that breeding methods that utilizes non-additive genetic variation are more efficient than selective breeding programs that utilize additive genetic variation. A good example of such a program utilizing non-additive genetic variation is clonal propagation of selected individuals. However, for early coppice height, significant Vgca, in addition to high heritability, suggests that additive genetic variation can be successfully utilized through the use of traditional






65

Table 23. Estimates of variance components and genetic
variances for four growth traits in
Eucalyptus grandis



Variance component

Trait GCA SCA Additive Dominance

Seedl ing
Height at 0.06 0.22 0.24 0.88
6 months

Seedling *
Height at 14.71 82.51 58.84 330.04
2.4 years

Seedl i ng , DBH at 0.07 0.24 0.29 0.94
2.4 years

Coppice ** height 0.50 0.11 1.99 0.44 at 5 months


* and ** Siginificant at 5% and 1%, respectively






66


Table 24. Rankings of general combining ability of 10
Eucalyptus grandis parents from GPOP73




Parent


Rank Seedling Seedling Seedling DBH Coppice
height at height at at 2.4 years height at 6 months 2.4 years 5 months


1 1188 1189 1178 1180 2 1185 1178 1189 1178 3 1178 1180 1180 1189 4 1183 1179 198 1179 5 1189 1191 1191 1191 6 1179 1188 1188 198 7 198 198 1179 1182 8 1191 1183 1183 1185 9 1182 1182 1185 1183 10 1180 1185 1182 1188






67


breeding procedures, higher genetic improvement can be achieved through the use of additive genetic variance.

The GCA rankings for the four traits of the 10 parents is given in Table 24. The rankings of parents change considerably through age as well as among traits. Spearman's Rank correlations of the 10 parents based on their GCA rankings is given in Table 25. These correlations are an indication of how well the rankings of parents for the 4 traits are correlated. Significant correlations were observed for coppice height at 5 months with DBH at 2.4 years (0.79) and with seedling height at 2.4 years (0.72). The correlation was also significant between height and DBH at 2.4 years. Seedling height at 6 months was not correlated with any other trait. This suggests that 6 months is too early an age for the prediction of growth at

later ages. The highly significant correlation between coppice height and seedling height at 2.4 years suggest that by selecting parents based on seedling height at 2.4 years, we would also be selecting for coppice performance.


Selection Indices


Future options for E. grandis in Florida include

conversion of GPOP77 into seedling seed orchard. To aid in the selection process, a selection index equations were

developed to select individuals based on seedling height at

7 months, coppice height at 3 months, coppice volume, coppice quality and frost resilience scores at 64 months.





68


Table 25. Spearman rank correlations for rankings of 10
Eucalyptus grandis parents for four
growth traits in GPOP73




Height at DBH at Coppice height
2.4 years 2.4 years at 5 months


Height at 0.05 -0.02 -0.42
6 months

Height at 0.87** 0.72*
2.4 years

DBH at 0.79**
2.4 years


* and ** Significant at 5% at 1%, respectively






69

The phenotypic variances and covariances, genotypic variances and covariances and the relative weights (bi 's) for the ten index equations are given in Tables 26, 27 and 28 respectively-. Considerable range in -the values of indices found between the individuals agrees with the significant -genetic variation found in the traits. To test the significance of the relative economic weights assigned for each index equation, the ten index values obtained for each individual were subjected to Spearman's rank correlations. The results showed highly significant correlations between the ten index values within an individual suggesting that the economic weights assigned did not effect the rankings of the individuals.

There are however some limitations recognized in the

design of GPOP77. Slow growing trees at early stages may be suppressed by the fast growing ones. This could explain the higher growth differential between earlier and later generation trees at 64 months. Another limitation is the

effect of microsite variation that is not accounted for the design. This may inflate the variance components and consequently the predicted gains. Such problems can be avoided by using a randomized block design. The experiment could be conducted on different sites thereby getting an estimate of genotype X environment interaction which seems to be important. Appendix 4 gives a listing of the 300

selected trees based on index 3 values.






70


Table 26. Phenotypic variances (diagonal) and
covariances (off diagonal) for Eucalyptus
grandis traits used in selection index
equations.



* ** *** Frost
Shgt Chgt Cvol Resi 1 - Coppice ience quality



Shgt 18.10 12.15 618.11 -0.19 -0.02 Chgt 1127.14 15085.96 -2.40 -4.21 Cvol 817054.98 120.37 139.03 Frost resili- 0.28 0.24 nce

Coppice quality 0.58



* Shgt = Seedling height at 7 months
** Chgt = Coppice height at 3 months
*** Cvol = Coppice volume at 64 months






71


Table 27. Genotypic variances (diagonal) and
covariances (off diagonal) for Eucalyptus
grandis traits used in selection index
equations in GPOP77.





* ** *** Frost resi- Coppice
Shgt Chgt Cvol li ence Quality


Shgt 5.69 -11.56 173.01 -0.17 -0.08 Chgt 332.32 5763.93 -1.16 -1.14 Cvol 256449.50 -495.00 -351.40


Frost resili- 0.08 0.06 nce

Coppice 0.07
quality


* = Seedling height at 7 months
** = Coppice height at 3 months
*** = Coppice volume at 64 months






72


Table 28. Selection index coefficients for five
Eucalyptus grandis growth traits
for ten equations in GPOP77.






Trait.
Index Shgt Chgt Vol FR CO


1 -29.03 -8.00 0.66 -1781.50 72.96 2 -19.36 -5.33 0.44 -1187.66 48.64 3 -38.69 -10.66 0.88 -2375.34 97.29 4 -19.27 -5.33 0.44 -1187.72 48.65 5 -23.27 -6.39 0.53 -1424.91 58.45 6 -28.51 -8.10 0.66 -1779.70 72.47 7 - 3.93 -1.01 0.09 - 238.51 10.04 8 -18.88 -5.44 0.44 -1186.10 48.08 9 -20.45 -5.89 0.47 -1279.80 51.92 10 - 4.18 -0.94 0.09 - 239.96 10.38


* Shgt = Seedling height at 7 months Chgt = Coppice height at 3 months Vol = Coppice Volume at 64 months FR = Frost resilience CQ = Coppice quality















CONCLUSIONS


Considerable gains have been achieved through four

generations of selection of E. grandis for increased coppice stem size, frost resilience, coppice quality and form. Genetic variation in the base population still seems to be

high enough to expect further gains in biomass productivity through continued recurrent selection.

In coppice volume production at 64 months after

harvest, fourth-generation families produced nearly 100 % more wood than the overall population mean and nearly 47 % more when compared to the third generation families. When

the best trees were rated against fourth-generation families, nearly two-fold differences in traits such as height and DBH were common. On a coppice volume basis, the best trees were more than three times larger at 64 months. This suggests an obvious potential for improvement in wood

productivity through clonal propagation of selected individuals to capture the full genetic potential.

Differences in biomass productivity between sources

seems to be an important factor that should be considered in future introductions. For coppice growth at 64 months sources from Queensland performed the best. However, for



73






74


frost resilience and coppice quality sources from New South Wales performed the best.

Individual and family heritabilities were higher than

those reported for E. grandis. High heritabilities indicate the potential of genetic improvement through mass and family selection. Inbreeding depression resulting from the mating of related individuals appears negligible with no significant effect on predicted genetic gains. The inbreeding coefficient (F) of all possible crosses in future seed orchard for GPOP77 ranged from 0 to 0.25. The mean F for all the matings was considerably low (0 to 1%). This

was attributed to the low percentage of related individuals and the passing of two-three generations since direct relationship due to common ancestry. However, percent

selfing occurring in E. grandis is significantly higher than pines. Based on published estimates of height loss of 50 % for selfed progeny, the predicted gains in volume were reduced by a factor proportional to the estimated mean F for the next generation. A 90 % gain in coppice volume at 64

months is predicted for a combined selection of the top 100 families with three trees per family. Other selection

strategies also had high predicted genetic gains.

The amount of non-additive genetic variation in

seedling growth is found to be high in 10 parents of E. grandis studied. Parental lines 1178, 1180 and 1189 consistently produced progeny that ranked high for seedling as well as coppice growth at later ages. The variance of






75


SCA (V sca) was higher than variance due to general combining ability (V gca) for the seedling growth traits suggesting that genetic improvement can be achieved through techniques that utilize non-additive genetic variance. The

genetic superiority exhibited by the top 200 trees in GPOP77 for 64-month coppice height provide a good source for clonal propagation. However, for early coppice height the V gca
was higher than Vsca . Significant Vgca along with high heritabilities suggest that improvement for coppice height can be better achieved through the use traditional selectionprocedures. Spearman's rank correlations of the 10 parents showed high correlation between height at 2.4 years and early coppice height. This suggests that coppice height can be improved by selecting the parents based on the seedling height growth at 2.4 years.

Selection indices were developed based on ten different economic weights being assigned to: seedling height at 7 months, coppice height at 3 months, coppice volume, frost resilience and coppice qualityat 64 months. Spearman's rank

correlations of the ten index values of an individual was highly significant indicating no significant effect due to relative economic weights assigned. Based on these selection index values, individuals from GPOP77 can be

selected and converted to a seedling seed orchard to provide a source for the genetic material for next generation. The best strategy for orchard development would be to retain three trees of the top 100 families.






76


Predicted genetic gains for this strategy was the highest (90%).

With continued introductions to broaden the genetic base and selection for growth and frost resilience, the

future looks promising for E. grandis to be a prime hardwood species in Florida for woody biomass production.










APPENDIX 1. PEDIGREES AND ORIGINS OF 529 FAMILIES USED IN GPOP77 Table 1.1 Pedigrees and origin* of 529 Eucalyptus grandis families
in GPOP77.

Pedigree Origin Generation

Vacant UNPLANTED SPOT 2

88) 836) 295 SLP)IMOK)GYMPIE Q 3 90) 836) 295 SLP)IMOK)GYMPIE Q 3 97) 293 SLP)URUNGA N 2 101) 836) 295 SLP)IMOK)GYMPIE Q 3 123) 294 BPP)COFFS HRBR N 2 125) 294 BPP)COFFS HRBR N 2 186) 293 SLP)URUNGA N 2 293 URUNGA N 1 294 COFFS HRBR N 1 295 GYMPIE Q 1 296 MID-NORTH COAST N 1 298 MISIONES, ARGENTINA 1 586 OSCEOLA CO 1 587 OSCEOLA CO 1 635 INDIAN RIVER CO 1 674 PLANT CITY 1 675 PLANT CITY 1 675 TAMPA 1 819 CONCORDIA ARGENTINA 1 829 LAKE HELL'N BLAZES 1 834 BAINBRIDGE GA 1 842) 5) 293 G73)FFF)URUNGA N 3 844) 18) 293 G73)FFF)URUNGA N 3

77






78


Table 1.1 - continued.

Pedigree Origin Generation 845) 24) 293 G73)FFF)URUNGA N 3 846) 24) 293 G73)FFF)URUNGA N 3 847) 24) 293 G73)FFF)URUNGA N 3 848) 24) 293 G73)FFF)URUNGA N 3 849) 26) 293 G73)FFF)URUNGA N 3 850) 26) 293 G73)FFF)URUNGA N 3 853) 72) 293 G73)FFF)URUNGA N 3 854) 72) 293 G73)FFF)URUNGA N 3 856) 73) 293 G73)FFF)URUNGA N 3 857) 73) 293 G73)FFF)URUNGA N 3 858) 73) 293 G73)FFF)URUNGA N 3 859) 75) 293 G73)FFF)URUNGA N 3 860) 75) 293 G73)FFF)URUNGA N 3 861) 76) 293 G73)FFF)URUNGA N 3 864) 77) 293 G73)FFF)URUNGA N 3 865) 77) 293 G73)FFF)URUNGA N 3 867) 78) 293 G73)FFF)URUNGA N 3 868) 78) 293 G73)FFF)URUNGA N 3 869) 79) 293 G73)FFF)URUNGA N 3 870) 79) 293 G73)FFF)URUNGA N 3 871) 79) 293 G73)FFF)URUNGA N 3 873) 80) 293 G73)FFF)URUNGA N 3 874) 80) 293 G73)FFF)URUNGA N 3 875) 80) 293 G73)FFF)URUNGA N 3






79

Table 1.1 - continued.

Pedigree Origin Generati on 876) 97) 293 G73)SLP)URUNGA N 3 877) 97) 293 G73)SLP)URUNGA N 3 879) 97) 293 G73)SLP)URUNGA N 3 880) 97) 293 G73)SLP)URUNGA N 3 881) 97) 293 G73)SLP)URUNGA N 3 883) 107) 293 G73)SLP)URUNGA N 3 884) 107) 293 G73)SLP)URUNGA N 3 885) 109) 293 G73)SLP)URUNGA N 3 887) 109) 293 G73)SLP)URUNGA N 3 888) 109) 293 G73)SLP)URUNGA N 3 889) 118) 293 G73)SLP)URUNGA N 3 890) 119) 293 G73)SLP)URUNGA N 3 892) 119) 293 G73)SLP)URUNGA N 3 893) 273 G73)PINECREEK N 2 894) 293 G73)URUNGAN 2 895) 293 G73)URUNGA N 2 896) 81) 294 G73)BIG)COFFS HRBR N 3 897) 81) 294 G73)BIG)COFFS HRBR N 3 898) 84) 294 G73)BIG)COFFS HRBR N 3 899) 84) 294 G73)BIG)COFFS HRBR N 3 900) 84) 294 G73)BIG)COFFS HRBR N 3 901) 85) 294 G73)BIG)COFFS HRBR N 3 902) 85) 294 G73)BIG)COFFS HRBR N 3 903) 85) 294 G73)BIG)COFFS HRBR N 3





80

Table 1.1 - continued.

Pedigree Origin Generation 904) 112) 294 G73)SLP)COFFS HRBR N 3 905) 112) 294 G73)SLP)COFFS HRBR N 3 906) 113) 294 G73)SLP)COFFS HRBR N 3 907) 113) 294 G73)SLP)COFFS HRBR N 3 908) 114) 294 G73)SLP)COFFS HRBR N 3 909) 114) 294 G73)SLP)COFFS HRBR N 3 910) 114) 294 G73)SLP)COFFS HRBR N 3 911) 114) 294 G73)SLP)COFFS HRBR N 3 912) 114) 294 G73)SLP)COFFS HRBR N 3 913) 114) 294 G73)SLP)COFFS HRBR N 3 914) 115) 294 G73)SLP)COFFS HRBR N 3 915) 116) 294 G73)SLP)COFFS HRBR N 3 916) 116) 294 G73)SLP)COFFS HRBR N 3 917) 116) 294 G73)SLP)COFFS HRBR N 3 918) 116) 294 G73)SLP)COFFS HRBR N 3 919) 116) 294 G73)SLP)COFFS HRBR N 3 920) 117) 294 G73)SLP)COFFS HRBR N 3 921) 117) 294 G73)SLP)COFFS HRBR N 3 923) 117) 294 G73)SLP)COFFS HRBR N 3 924) 117) 294 G73)SLP)COFFS HRBR N 3 925) 117) 294 G73)SLP)COFFS HRBR N 3 926) 117) 294 G73)SLP)COFFS HRBR N 3 927) 117) 294 G73)SLP)COFFS HRBR N 3 928) 120) 294 G73)SLP)COFFS HRBR N 3





81

Table 1.1 - continued

Pedigree Origin Generation 931) 120) 294 G73)SLP)COFFS HRBR N 3 933) 121) 294 G73)BPP)COFFS HRBR N 3 934) 121) 294 G73)BPP)COFFS HRBR N 3 935) 121) 294 G73)BPP)COFFS HRBR N 3 936) 121) 294 G73)BPP)COFFS HRBR N 3 937) 121) 294 G73)BPP)COFFS HRBR N 3 938) 123) 294 G73)BPP)COFFS HRBR N 3 941) 125) 294 G73)BPP)COFFS HRBR N 3 943) 126) 294 G73)BPP)COFFS HRBR N 3 944) 127) 294 G73)BPP)COFFS HRBR N 3 946) 128) 294 G73)BPP)COFFS HRBR N 3

947) 128) 294 G73)BPP)COFFS HRBR N 3 948) 129) 294 G73)BPP)COFFS HRBR N 3 949) 129) 294 G73)BPP)COFFS HRBR N 3 950) 129) 294 G73)BPP)COFFS HRBR N 3 952) 132) 294 G73)SFDP)COFFS HRBR N 3 953) 132) 294 G73)SFDP)COFFS HRBR N 3 954) 132) 294 G73)SFDP)COFFS HRBR N 3 955) 132) 294 G73)SFDP)COFFS HRBR N 3 956) 132) 294 G73)SFDP)COFFS HRBR N 3 958) 134) 294 G73)SFDP)COFFS HRBR N 3 959) 134) 294 G73)SFDP)COFFS HRBR N 3 960) 134) 294 G73)SFDP)COFFS HRBR N 3 961) 134) 294 G73)SFDP)COFFS HRBR N 3





82


Table 1.1 - continued.

Pedigree Ori gin Generation 962) 134) 294 G73)SFDP)COFFS HRBR N 3 963) 136) 294 G73)SFSP)COFFS HRBR N 3 964) 137) 294 G73)SFSP)COFFS HRBR N 3 965) 138) 294 G73)SFSP)COFFS HRBR N 3 966) 138) 294 G73)SFSP)COFFS HRBR N 3 967) 138) 294 G73)SFSP)COFFS HRBR N 3 968) 138) 294 G73)SFSP)COFFS HRBR N 3 969) 138) 294 G73)SFSP)COFFS HRBR N 3 971) 228) 294 G73)SLHQ)COFFS HRBR N 3 972) 228) 294 G73)SLHQ)COFFS HRBR N 3 973) 228) 294 G73)SLHQ)COFFS HRBR N 3 974) 234)1736) 294 G73)F69)BIG)COFFS HRBR N 4 975) 234)1736) 294 G73)F69)BIG)COFFS HRBR 4 976) 234)1736) 294 G73)F69)BIG)COFFS HRBR 4 977) 235)1736) 294 G73)F69)BIG)COFFS HRBR 4 978) 235)1736) 294 G73)F69)BIG)COFFS HRBR 4 979) 235)1736) 294 G73)F69)BIG)COFFS HRBR 4 981) 236)1736) 294 G73)F69)BIG)COFFS HRBR 4 982) 236)1736) 294 G73)F69)BIG)COFFS HRBR 4 983) 237)1736) 294 G73)F69)BIG)COFFS HRBR 4 984) 237)1736) 294 G73)F69)BIG)COFFS HRBR 4 985) 237)1736) 294 G73)F69)BIG)COFFS HRBR 4 986) 86) 835) 295 G73)SLP)IMOK)GYMPIE Q 4





83

Table 1.1 - continued.

Pedigree Ori gi n Generation 987) 86) 835) 295 G73)SLP)IMOK)GYMPIE Q 4 988) 86) 835) 295 G73)SLP)IMOK)GYMPIE Q 4 989) 96) 835) 295 G73)SLP)IMOK)GYMPIE Q 4 990) 96) 835) 295 G73)SLP)IMOK)GYMPIE Q 4 991) 96) 835) 295 G73)SLP)IMOK)GYMPIE Q 4 992) 96) 835) 295 G73)SLP)IMOK)GYMPIE Q 4 993) 96) 835) 295 G73)SLP)IMOK)GYMPIE Q 4 994) 88) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 996) 88) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 997) 88) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 998) 88) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1000) 88) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1001) 88) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1002) 90) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1003) 90) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1004) 90) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1005) 91) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1006) 91) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1007) 91) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1009) 91) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1010) 101) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1011) 101) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1012) 101) 836) 295 G73)SLP)IMOK)GYMPIE Q 4 1016) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4






84

Table 1.1 - continued.

Pedigree Ori gi n Generation 1017) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1018) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1019) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1020) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1020) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1022) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1023) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1024) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1025) 92) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1026) 103) 837) 295 G73)SLP)IMOK)GYMPIE Q 4 1027) 98) 838) 295 G73)SLP)IMOK)GYMPIE Q 4 1028) 98) 838) 295 G73)SLP)IMOK)GYMPIE Q 4 1029) 95) 295 G73)SLP)GYMPIE Q 3 1030) 95) 295 G73)SLP)GYMPIE Q 3 1031) 95) 295 G73)SLP)GYMPIE Q 3 1033) 295 G73)GYMPIE Q 2 1034) 295 G73)GYMPIE Q 2 1035) 100) 296 G73)SLP)NORTH COAST N 3 1036) 100) 296 G73)SLP)NORTH COAST N 3 1038) 100) 296 G73)SLP)NORTH COAST N 3 1039) 100) 296 G73)SLP)NORTH COAST N 3 1040) 296 G73)NORTH COAST N 2 1041) 296 G73)NORTH COAST N 2 1043) 104) 298 G73)SLP)ARGENTINA 3






85

Table 1.1 -- continued.

Pedigree Ori gin Generation 1045) 104) 298 G73)SLP)ARGENTINA 3 1046) 104) 298 G73)SLP)ARGENTINA 3 1047) 185 G73)ORLANDO 2 1049) 189 G73)0RLANDO 2 1052) 212 G73)MULBERRY 2 1053) 212 G73)MULBERRY 2 1054) 217 G73)PALMDALE 2 1055) 217 G73)PALMDALE 2 1056) 219 G73)CABBAGE VAT 2 1057) 220 G73)CABBAGE VAT 2 1058) 220 G73)CABBAGE VAT 2 1059) 222 G73)CABBAGE VAT 2 1060) 222 G73)CABBAGE VAT 2 1063) 224 G73)CABBAGE VAT 2 1065) 226 G73)TICE 2 1067) 226 G73)TICE 2 1068) 227 G73)TICE 2 1069) 227 G73)TICE 2 1070) 227 G73)TICE 2 1071) 227 G73)TICE 2 1072) 227 G73)TICE 2 1077) 248 G73)ITALY 2 1078) 248 G73)ITALY 2 1079) 250 G73)PINE CREEK N 2






86

Table 1.1 continued.

Pedigree Origin Generation


1080) 250 G73)PINE CREEK N 2 1081) 250 G73)PINE CREEK N 2 1082) 250 G73)PINE CREEK N 2 1083) 251 G73)PINE CREEK N 2 1085) 251 G73)PINE CREEK N 2 1086) 252 G73)PINE CREEK N 2 1087) 252 G73)PINE CREEK N 2 1088) 252 G73)PINE CREEK N 2 1091) 253 G73)PINE CREEK N 2 1092) 254 G73)PINE CREEK N 2 1093) 256 G73)PINE CREEK N 2 1094) 258 G73)PINE CREEK N 2 1095) 259 G73)PINE CREEK N 2 1096) 260 G73)PINE CREEK N 2 1097) 262 G73)PINE CREEK N 2 1098) 267 G73)PINE CREEK N 2 1099) 268 G73)PINE CREEK N 2 1100) 273 G73)PINE CREEK N 2 1101) 273 G73)PINE CREEK N 2 1102) 277 G73)PINE CREEK N 2 1103) 277 G73)PINE CREEK N 2 1104) 278 G73)PINE CREEK N 2 1105) 278 G73)PINE CREEK N 2 1107) 257 G73)N 2






87

Table 1.1 - continued.

Pedigree -Ori gin Generation 1108) 257 G73)N 2 1109) 263 G73)NEWRY N 2 1110) 264 G73)NEWRY N 2 1111) 265 G73)NEWRY N 2 1112) 274 G73)NEWRY N 2 1113) 274 G73)NEWRY N 2 1114) 275 G73)NEWRY N 2 1115) 276 G73)NEWRY N 2 1117) 269 G73)BOAMBEE N 2 1118) 269 G73)BOAMBEE N 2 1119) 271 G73)CONGLOMERATE N 2 1120) 271 G73)CONGLOMERATE N 2 1122) 272 G73)CONGLOMERATE N 2 1123) 281 G73)CONGLOMERATE N 2 1124) 283 G73)CONGLOMERATE N 2 1125) 283 G73)CONGLOMERATE N 2 1127) 285 G73)CONGLOMERATE N 2 1128) 285 G73)CONGLOMERATE N 2 1129) 285 G73)CONGLOMERATE N 2 1130) 287 G73)CONGLOMERATE N 2 1131) 287 G73)CONGLOMERATE N 2 1132) 288 G73)CONGLOMERATE N 2 1133) 288 G73)CONGLOMERATE N 2 1134) 288 G73)CONGLOMERATE N 2






88

Table 1.1 - continued.
Pedi gree Origin Generation


1135) 290 G73)CONGLOMERATE N 2 1136) 290 G73)CONGLOMERATE N 2 1137) 290 G73)CONGLOMERATE N 2 1138) 291 G73)CONGLOMERATE N 2 1139) 292 G73)CONGLOMERATE N 2 1140) 292 G73)CONGLOMERATE N 2 1141) 292 G73)CONGLOMERATE N 2 1142) 307 G73)WOONDUM Q 2 1143) 307 G73)WOONDUM Q 2 1144) 309 G73)WOONDUM Q 2 1145) 309 G73)WOONDUM Q 2 1146) 309 G73)WOONDUM Q 2 1147) 309 G73)WOONDUM Q 2 1148) 310 G73)WOONDUM Q 2 1149) 310 G73)WOONDUM Q 2 1150) 311 G73)WOONDUM Q 2 1151) 311 G73)WOONDUM Q 2 1152) 312 G73)WOONDUM Q 2 1153) 312 G73)WOONDUM Q 2 1154) 312 G73)WOONDUM Q 2 1155) 313 G73)GYMPIE Q 2 1156) 313 G73)GYMPIE Q 2 1158) 317 G73)GYMPIE Q 2 1160) 319 G73)KENILWORTH Q 2





89


Table 1.1 - continued.

Pedigree Ori gi n Generation

1161) 320 G73)BELLTHORPE Q 2 1162) 321 G73)BELLTHORPE Q 2 1163) 324 G73)POMONA Q 2 1164) 325 G73)POMONA Q 2 1165) 326 G73)POMONA Q 2 1166) 326 G73)POMONA Q 2 1168) 327 G73)POMONA Q 2 1169) 328 G73)POMONA Q 2 1170) 328 G73)POMONA Q 2 1171) 329 G73)POMONA Q 2 1172) 330 G73)POMONA Q 2 1173) 331 G73)POMONA Q 2 1174) 331 G73)POMONA Q 2 1175) 331 G73)POMONA Q 2 1176) 331 G73)POMONA Q 2 1177) 331 G73)POMONA Q 2 1192) 195 G73)RSA 2 1193)1013 G73)RSA 2 1194)1013X1187 G73)RSA 2 1195)1013X198 G73)RSA 2 1196)1013X1178 G73)RSA 2 1197)1178X1186 G73)RSA 2 1198)1178X1186 G73)RSA 2 1199)1179X1178 G73)RSA 2




Full Text

PAGE 1

GENETIC IMPROVEMENT OF EUCALYPTUS 6RANDIS (HILL) MAIDEN FOR COPPICE PRODUCTIVITY BY KONDALA VIKRAM REDDY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1985

PAGE 2

I TABLE OF CONTENTS Page ACKNOWLEDGMENTS ii, LIST OF TABLES iv ABSTRACT ^ INTRODUCTION 1 OBJECTIVES 5 LITERATURE REVIEW 6 Genetic Improvement through Recurrent Selection 6 Inbreeding g Di al 1 el Anal ysi s , 14 Sel ecti on Indi ces 15 MATERIALS AND METHODS 20 Study Area for Diallel Analysis 22 Statistical Analysis 25 RESULTS AND DISCUSSION 31 Genetic Improvement through Recurrent Selection 31 Variation in Frost and Coppice Scores 42 Inbreeding and Seed Orchard Development 44 Diallel Analysis 50 Selection Indices 69 CONCLUSIONS 73 APPENDIX 1. PEDIGREES AND ORIGINS OF 529 FAMILIES USED IN GPOP77 7g APPENDIX 2. DESCRIPTION OF MEASUREMENTS OBTAINED AT 64 MONTH COPPICE AGE 99 APPENDIX 3. ANALYSIS OF VARIANCE TABLES FOR GROWTH COPPICE AND FROST SCORES ' 102 ii

PAGE 3

Page APPENDIX 4 IDENTIFICATION OF 300 SELECTED TREES USING SELECTION INDEX 3 Ill LITERATURE CITED 118 BIOGRAPHICAL SKETCH 124 i i i

PAGE 4

1 ACKNOWLEDGMENTS I would like to thank members of my supervisory committee, Drs. Don Rockwood, Ray Goddard, Charles Wilcox, Frank Martin, Susan Kossuth, and Randy Chow, for their helpful comments and suggestions. Special thanks go to the committee chairman. Dr. Don Rockwood. He allowed me the space and time to learn on my own, and to make my own mistakes, from which I learned many valuable lessons. His guidance, patience and support throughout this phase of my studies is deeply appreciated. Special thanks are also extended to Drs. Charles Wilcox and Michael DeLorenzo, Dairy Science Department, for their invaluable help in statistical analysis. I would also like to thank George Meskimen and the U. S. Forest Service who maintained the study population through September 1984. I also wish to thank Laurel Judd, Craig Reed and Karen Bondurant. Their friendship can never be repaid and will always be remembered. This study was supported in part under Subcontract no. 19X-09050C with Oak Ridge National Laboratory under Martin Marietta Energy Systems, Inc., contract DE-AC05-840R21400 with the U. S. Department of Energy and by a cooperative IV

PAGE 5

program between the Institute of Food and Agricultural Sciences of the University of Florida and the Gas Research Institute, entitled "Methane from Biomass and Waste." V

PAGE 6

! LIST OF TABLES Tabi e Page 1 Eucal y pt us grandi s prog eni es in GP0P77 by generation and origin 21 2 Diallel analyses of variance for four Eucalyptus grandi s traits in GP0P73 28 3 Relative economic weights used in the ten selection index equations for five Eucal yptus grandi s traits 29 4 Eucal yptus grandi s generation means for individual tree performances in GP0P77 through 64 months after harvest 32 5 Percent gains for Eucalyptus grandi s generations compared to the preceding generation for different traits 35 6 Comparison of origin of Eucal ypt us grandi s progenies in GP0P77 37 7 Heri tabi 1 i ti es of Eucal yptus grandi s traits by and over generation 39 8 Genetic correlations and their standard errors (in parenthesis) among growth traits, frost and coppice scores of Eucal ypt us grandi s in GP0P77 41 9 Mean frost and coppice scores of Eucal yptus grandi s by origin in GP0P77 43 10 Frequency of different degrees of relationship in a 50family subset of Eucalyptus grandi s seed orchard in GP0P77 45 11 Mean inbreeding coefficients (F) for five selection strategies of Eucalyptus grandis in GP0P77 47 12 Predicted genetic gains in 64-month coppice volume for alternative improvement strategies of Eucalyptus grandi s in 6P0P77 49 vi

PAGE 7

Mean seedling height (dm) at 6 months for 45 crosses of Eucalyptus grandi s in GP0P73 51 General combining abilities and associated variances of 10 Eucalyptus grandi s families for seedling height at 6 months and coppice height at five months in GP0P73 53 Specific combining abilities of 45 Eucalyptus grandis crosses for seedling height at 6 months in GP0P73 54 Mean seedling height (dm) of 45 Eucal yptus grandi s crosses at 2.4 years in GPOP73 . 56 GCA and associated variances of 10 Eucal yptus grandis parents for seedling height and DBH at age 2.4 years in GP0P73 57 Specific combining abilities for seedling height at 2.4 years for 45 Eucal yptus grandis crosses in GP0P73 58 Mean seedling DBH (cm) of 45 Eucalyptus grandi s crosses in GP0P73 at age 2.4 years 1.59 Specific combining ability of 45 Eucal yptus g randi s crosses for 2.4 year seedl i nq DBH in GP0P73 60 Mean coppice height (dm) of 45 Eucalyptus grandi s crosses at 5 months in GP0P73 62 Specific combining ability of 45 Eucal yptus grandi s crosses for 5 month coppice height in 6P0P73. ,63 Estimates of variance components and genetic variances for four growth traits in Eucalyptu s grandi s . .65 Rankings of general combining ability of 10 Eucalyptus grandi s parents from GP0P73 66 Spearman rank correlations for rankings of 10 Eucal yptus grandi s parents for four growth traits in GP0P73 68 vn

PAGE 8

Tabi e Page 26 Phenotypic variances (diagonal) and covariances ( of f -di ag onal ) for Eucalyptus grandi s trai ts used in selection index equati ons 70 27 Genotypic variances (diagonal) and covariances (off-diagonal) for Eucal yptus grandis traits used in selection index equations in GPOP77 71 28 Selection index coefficients for five Eucal yptus grandi s traits fot ten equations in GP0P77 72 viii

PAGE 9

I Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy GENETIC IMPROVEMENT OF EUCALYPTUS GRANDIS (HILL) MAIDEN FOR BIOMASS PRODUCTTUN By Kondala Vikram Reddy December 1985 Chairman: Dr. Donald L. Rockwood Major Department: Forest Resources and Conservation A genetic base population {GP0P77) of Eucalyptus grandi s (Hill) ex^. Maiden planted in July 1977, with 529 families representing four generations of selection, was partially harvested in August 1978. Regrowth through December 1983 was evaluated to assess genetic improvement potential for coppice productivity. Four generations of selection have produced impressive genetic gains. At 64 months after harvest, first, second, third and fourth generation families averaged 3.9, 13.34, 16.18 and 23.16 dry mt per hectare, respectively. Fourth generation families had the best frost resilience and quality of coppice. In volume production the best trees ix

PAGE 10

were more than three times larger than fourth generation trees at 64 months after harvest. An obvious potential thus exists for improvement in biomass production through clonal propagation of the best individuals. The potential inbreeding depression resulting through mating of related families was examined. The mean inbreeding coefficient in the offspring of all possible matings of selected individuals for six different selection strategies ranged from 0 to 1%. The predicted genetic gains adjusted for inbreeding were high. The highest gain of 90% was predicted by the selection of the top 100 families (three trees per family). A 10 X 10 half diallel analysis revealed significant amounts of non-additive genetic variation in grandi s . For seedling height at 6 months and height and DBH at 2.4 years, the dominance variance was higher than additive variance. This suggests potential improvement of seedling growth is higher through the utilization of non-additive genetic variation through clonal propagation. However, for coppice height at five months, the additive variance was higher than dominance variance, suggesting that utilization of additive genetic variance through selection can result in high geneti c gains. Selection index equations were developed using growth traits and frost and coppice scores. The economic weights used in developing these equations were found not to affect X

PAGE 11

the rankings of individuals. This was evident from the high rank correlations among the ten ranks based on ten selection i ndex equati ons . With continued breeding and new introductions, coppice productivity of grandi s can be improved for bi omass production in southern Florida. xi

PAGE 12

INTRODUCTION Eucalypts are evergreen hardwood trees and shrubs belonging to the genus Eucalyptus in the family Myrtaceae. The genus includes more than 500 species nearly all native to Australia. Several are native to Indonesia, New Guinea and the Phillippines (FAO, 1979). Eucalypts are noted for their extremely fast growth rates and under favorable environmental conditions grow year round. Although they occupy perhaps one-fifth of the world's man-made forests (Logan, 1967), only a small part of the world's tree breeding effort has been devoted to this genus. This is in spite of the fact that, when compared with conifers, the prospects of quicker financial returns on a breeding investment are high due to earlier and abundant production of seed and shorter plantation rotations (Eldridge, 1975). In Florida, eucalypts were first planted in 1878, but industrial plantations were not established until 1972 (Geary et al., 1983). Through cooperative research involving government, private land owners, and industry, approximately 6,000 hectares of Eucalyptus plantations existed in southern Florida by 1981. Eucalyptus grandis (Hill) ex. Maiden, one of the prime species, is recommended 1

PAGE 13

for the palmetto prairie (poor fertility) and acid flatwoods (Geary et al 1983) . In Florida there are three advantages of using E. grandi s over other hardwood species for biomass production. First. E. grandi s has an exceptional growth rate. For six progenies of E. grandi s at 2.5 years of age on organic soil. Rockwood and Geary (1982) reported a mean height of 10.2 m and a diameter at breast height (DBH) of 10.6 cm. This surpasses the impressive growth of £. saligna reported by Walters ( 1980) in Hawaii. Second. grandi s performs better in low quality soils when compared to other native hardwoods. A mean height of 4.7 m and DBH of 3.3 cms was achieved at 17 months of age on a sandy palmetto prairie soil (Rockwood and Geary, 1982). Although this growth is significantly lower than those achieved on organic sites, it is impressive when compared with the growth of other hardwoods on similar sites. Third, the ability of E. grandi s to coppice makes it an attractive species for short rotation biomass production. In much of the world. E. grandi s sprouts pr ol i f i cal 1 y , and in short rotation plantations two to four coppice crops can be achieved before replanting is necessary (FAQ, 1979). Both for the extent to which it has been planted around the world and for its excellent performance E. grandis is one of the most important exotic eucalypts. Depending on the natural distribution, variation within Eucalyptus species may be either cl i nal or ecotypi c

PAGE 14

3 (FAO. 1979). Eucalyptus grandis has a limited distribution, which sometimes consists of isolated provenances adapted to specific environmental conditions. Natural occurrence of E^. grandi s is in northern New South Wales and southern Queens! and . Eucalyptus grandis in Florida constitutes a landrace developed through three generations of selection (Geary et . al., 1983) and progeny testing in the local environment (Figure 1). The initial introductions along with local selections were planted in Immokalee, Florida, and later converted into a seed orchard. Selections from this orchard plus new introductions were tested in Silver Lake Prairie and constitute the second generation population. Subsequently, this plantation was converted to a seed orchard. Selections from Silver Lake Prairie seed orchard and additional new introductions constituted the third generation population (GP0P73) established in 1973 near LaBelle. GP0P73 was later rogued to form a seed orchard. Seed from GP0P73 seed orchard and new introductions comprise the current genetic base population (GP0P77).

PAGE 15

INTRODUCTIONS LOCAL SELECTIONS IMMOKALEE IMMOKALEE SEED ORCHARDS INTRODUCTIONS SILVER LAKE PRAIRIE|^ SILVER LAKE PRAIRIE SEED ORCHARD INTRODUCTIONS GPOP73 GP0P73 SEED ORCHARD INTRODUCTIONS GP0P77 Figure 1. Development of the current Eucalyptus grandi s breeding population (GP0P77) in southern Florida.

PAGE 16

OBJECTIVES This study had the following objectives: (1) To calculate the realized gains achieved through four generations of selection, (2) To study the effect of inbreeding through mating of related trees in a future seed orchard, (3) To predict the genetic gains, adjusted for inbreeding, in the next generation for different selection strategies, (4) To estimate the amount of dominance variation i n growth trai ts , (5) To develop a selection index to aid in selection of individuals based on a combined performance for seedling height, coppice height, coppice volume, frost resilience and coppice qual i ty . 5

PAGE 17

LITERATURE REVIEW Provenance Variation : Several trials of grandi s established around the world have revealed significant provenance variation. Trials in Brazil have shown that for both height and DBH, the Australian origin showed a larger variability than both South African and Brazilian origins (Assis and Brune, 1983). In Australia there were significant differences in productivity between provenances, with those from low altitude sites in southeastern Queensland and the Coff's Harbour areas performing well (Ades and Burgess, 1983). Provenance trials in South Africa have shown superior growth at four years of age at all sites for seedlots from New South Wales and Queensland (Darrow and Roeder, 1983). In comparison. South African seed orchard stock showed relatively poor growth but superior form. King (1983a, 1983b) reported the establishment of Ej_ grandi s provenance trials in Hawaii and California. In California, sources from higher elevations in Australia had the best growth and survi val . 6

PAGE 18

7 Genetic Improvement Through Recurrent Selection Appreciable genetic variation for growth has been observed in E^. grandi s , and successful selection programs for growth rate have been reported. Selection of the best genotypes is important in increasing E_. grandi s productivity. Differences among families were observed to be highly significant, as the best families produced as much as 50 percent more volume than the average of the families studied (Rockwood and Meskimen, 1981). In Florida, at 2.5 years of age, a gain of 163 percent in individual tree volume has been achieved in four generations of selection (Meskimen, 1983). Because of recurrent selection for local adaptation, the trees perform better than do progenies of outstanding trees selected in Australia, South Africa or elsewhere. Progeny tests in Brazil using 124 open -pol 1 i nat ed mother trees of E^. grandi s have shown high heri tabi 1 i tes for growth traits justifying improvement programs based on selection (Borges and Brune, 1983). Similar results were reported from Zimbabwe where in a combined analysis between -fami 1 y differences at 18 months indicated considerable diversity for height (Mullin and Gough, 1983). Research reported on the genetic variation for £. grandi s coppicing ability is limited. In southern Florida, evaluation of four progenies at two different sites showed no significant differences in coppicing ability (Rockwood and Geary, 1982). However, Geary et al . (1983) reported a

PAGE 19

1 8 progeny test that revealed significant genetic variation in coppicing ability. Resi stance to Frost : Species of Eucalyptus do not have annual resting buds. Indeterminate shoots grow continuously as long as favorable conditions exist and there is no requirement for rest or overwintering (FAO, 1979 ). Of the many Eucalyptus species evaluated in Florida for frost tolerance, £. grandi s was notably frost sensitive (Hunt and Zobel , 1978). Limited information is available on genetic variation in frost resilience in E^. grandi s . In their native Australia Eucalyptus species grow under a wide range of climatic conditions. Some species grow in regions of freezing temperature and frequent snow (Hunt and Zobel, 1978). In Australia the temperatures gradually decline to mid-winter. A severe frost is always preceded by a series of cold nights which harden the species (Burgess, 1983). In southeastern United States, on the other hand, the temperature drops suddenly from well above freezing to well below (Hunt and Zobel, 1978). Florida's winter freezes have a devastating effect on the survival of E^. grandi s , where swift moving cold fronts turn balmy afternoons into freezing mornings, severely damaging the tender eucalypts (Geary et al., 1983). There seems to be a positive correlation between growth rate and freeze tolerance. Hunt and Zobel ( 1978) found the fast growing progenies of E. vi mi nal i s not damaged by severe freeze in 197677. However, Jahromi ( 1983) tested several E_^ vi mi nal i s provenances in

PAGE 20

southern Georgia and northern Florida and found them not suitable for commercial planting. Jahromi concluded that a genetic improvement program can produce new strains of eucalypts which would be more frost hardy. Geary et al., (1983) report that the bigger the tree at the time of the freeze, the more resistant it is to frost. Inbr eedi ng Forest tree seed orchards are managed and designed with the main objective of providing a source of genetically high quality seed. The normal breeding system for trees is cross-fertilization as a result of insect or wind pollination. Consequently, seed orchards are designed to promote the opportunities to maximize outcrossing (van Buijtenen, 1971). The essential consequence of two individuals having a common ancestor is that they may both carry replicates of one of the genes present in the ancestor, and by mating they pass on these replicates to the offspring (Falconer, 1981). Thus, inbreeding increases the proportion of homozygous genes in a population and decreases the proportion of heterozygous genes. In a large random mating population, the probabilty that two allelles in an individual are identical by descent approaches zero and the population is fully outbred (F= 0). Whenever there is a tendency to assortative mating between relatives (e.g., sib-crossing or at the extreme self-pollination), the inbreeding coefficient will depart

PAGE 21

10 from zero. Computation of inbreeding coefficient therefore requires no more than tracing the pedigree back to the common ancestor and computing the probabilities at each segregation or level (Li, 1976). The inbreeding coefficient after one generation is 0.50, 0.25 and 0.125 for selfing, full-sib and half-sib mating, respectively (Falconer, 1981). Among normally outcrossing species any degree of relatedness between individuals that mate leads to inbreeding depression. As pointed out earlier, selfing is the most severe form of inbreeding and is of interest to the tree geneticists because it often results in the inbreeding depression of economically important traits. Selfing is of special interest where clonal orchards are used as a means of mass producing seed of superior quality. In such orchards the potential for selfing is magnified because selfs can be produced not only from selfing of a ramet but also from mating of ramets of the same clone (Squillace and Kraus, 1963). In recent years research has been done on pollination biology, amount of selfing and inbreeding depression in E. grandi s . Eldridge (1976) suggested that if inbreeding is prevalent in eucalypts, reduced improvement will follow the usual breeding methods. Working on the breeding system with several species of eucalypts, Pryor (1961) demonstrated a system of preferential outcrossing. His results showed that selfing does occur but less readily than outcrossing to other individuals. Other studies also found similar results

PAGE 22

! 12 crossed families and germination rate was slightly lower among selfed families. In P^. el 1 i otti i . self-pollination produced 10% as many germinated seedlings per cone as wind pollination (Snyder, 1968). Fowler (1964) working with P_. resi nosa Ait. found that seedlings resulting from self-pollination exhibit little or no inbreeding depression. Squillace and Kraus (1962) compared the effects of several degrees of inbreeding (selfing, backer ossi ng , full-sib crossing and half-sib crossing) against outcrossing, including polycrossing for several characteristics in P^. el 1 i otti i . Cone yield relative to number of flowers pollinated was not affected. Sound seed yields per cone, on the average, were lower for inbred trees than for trees that were outbred. By regression Squillace and Kraus ( 1962 ) also estimated that an increase of 0.1 in inbreeding coefficient resulted in a decrease of about 5 sound seeds per cone. The same increase in F was associated with 8% decrease in germination and 7% decrease in rate of germination. Growth was decreased by approximately 4% over the mean of the outcrossed individuals. These results were in close agreement with the results obtained by Gansel (1971) who concluded that inbreeding depression was evident in several traits including diameter, volume and oleoresin yield and was proportional to the degree of inbreeding. Snyder ( 1972) found that the offspring of selfed

PAGE 23

11 other individuals. Other studies also found similar results of predominant outcrossing for the genus Eucalyptus as a whole. These studies confirm the occurrence of a wide range of sel f -f erti 1 i ty within species, including two extremes of obligate selfing caused by cleistogamy (Venkatesh et al., 1973) and obligate outcrossing by self i ncompatabi 1 i ty (Pryor, 1957) or male sterility (Davis, 1969), or female sterility (Carr and Carr, 1972). The amount of natural selfing that occurs in eucalypts is higher than that reported for pines. An average of 7% selfing is reported for most pines (Wright, 1976). Published estimates for the degree of natural selfing in eucalypts vary with species: 24% in E. obi i qua (Brown et al., 1975), 37% in E. pauci fl ora (Phillips and Brown, 1977), 23% in E^. del egatensi s (Moran and Brown, 1980) and 18% in E^. stoatei (Hopper and Moran, 1981). In E. grandi s Eldridge (1978) reported 20% to 40% selfing to occur and Van Wyk (1981) estimated this average to be about 30%. The major consequences of inbreeding depression resulting from natural self-pollination in seed orchards of P^. taeda L. is reduced seed yield (Franklin, 1969 ). Franklin found the total seed yields to be slightly lower, yields of filled seed to be drastically lower and average weight of filled seed to be slightly lower after selfing. Percent of filled seed was 74.2 from crossed families and 14.3 from selfed families, a five-fold reduction.

PAGE 24

13 K elliottii were 21% shorter than wi nd -pol 1 i nat ed offspring at one year and 34% shorter after 5 years. In E. grandi s Hodgson ( 1974, 1976, 1977) studied the extent of inbreeding for inbred individuals. The height reached by selfed progeny of nearly all clones of E. grandi s at 11-18 months of age was 8% to 49% less than that of outcrossed progeny. In addition, selfed progeny was usually more crooked and included about 30% abnormal seedlings. The study also found self-pollination and self-fertilization treatments adversely affected seed yield and stem form leading to the conclusion that products from a seed orchard in grandi s may be subject to an average of 30% degrade as a result of selfing. Thus there is an opportunity to identify and control at least some of the worst affects of selfing in E. grandi s seed orchards by culling poorer seedlings in the nursery and by testing and rejecting the more self-fertile clones (Eldridge, 1978). However, in regnans the selfed seedlings did not differ in appearance and were not consistently smaller than outcrossed or open -pol 1 1 nat ed seedlings up to one year of age (Eldridge, 1970). In this study inbreeding depression was not obvious until six years after planting and even then many individuals within selfed families were vigorous and of good form. Eldridge ( 1978) concluded that in E. grandi s many times more seeds were set after crossing than after selfing or open -pol 1 i nat i on , whereas in E_^ regnans sel fi nq.

PAGE 25

14 crossing and open-pollination yielded only a small range in numbers of seed per capsule. In spite of using large numbers of clones and advanced planting designs, which theoretically ensure that trees of the same clone or related families are planted as far apart as possible in a seed orchard, the risk of self-pollination can only be reduced but not eliminated. Thus any predicted genetic gains should take into account the losses due to inbreeding depression resulting from selfing and mating of related individuals. Pi all el Analysis The making of all possible crosses among a set of parents has been a popular mating scheme for over 40 years. The concept of combining ability is becoming increasingly important in tree improvement. It is especially useful in testing procedures in which it is desired to study and compare the performances of lines. According to Sprague and Tatum (1942), the term general combining ability (GCA) is used to designate the average performance of a line when crossed to different lines. The term specific combining ability (SCA) is used to designate those cases in which certain combinations do relatively better or worse than would be expected on the basis of the average performance of the 1 i nes i nvol ved . In the analysis of data from a diallel crosses, the average performance of each progeny is divided into

PAGE 26

15 (interaction). If the SCA mean square is not significant, one would accept the hypothesis that the performance of a single cross progeny can adequately be predicted on the basis of GCA. The best performing progeny may be produced by crossing the two parents having the highest GCA. If, on the other hand, the SCA mean square is significant one needs to ask how important the interactions are in determining the performance of single cross progeny. On the basis of the above discussion it can then be generalized that GCA depends on additive genetic variance. A good general combiner consistently produces progeny which do well when crossed with the other parents in the group. On the other hand specific combining ability depends on dominance variance. A good specific combiner produces good progeny only when crossed to certain parents in the group. In E^. grandi s Van Wyk ( 1976, 1977) found considerable additive as well as non-additive genetic variance. His estimates of GCA for most growth traits were significant, suggesting a selective breeding program utilizing additive genetic variance will be effective in improving the population mean. Sel ecti on Indi ces A tree breeder wishes to improve more than one character or trait in his breeding work. Tandem selection appears impractical because of the long time span involved. Independent culling levels entail selecting for all the

PAGE 27

1 16 Independent culling levels entail selecting for all the characters at the same time but independently, rejecting all individuals that fail to come up with a certain standard for each character regardless of their values for other traits of interest. The main problem, however, is that if one wishes to improve a large number of traits, a very large population must be screened. The best method, from the theoretical point of view, to deal with this problem is the use of the selection index method. A selection index (I) is a linear function of the phenotypic value of trait ( ) weighted by index weights (b^. ) in such a way that maximum correlation with the economic breeding value is obtained. In other words, the index is the best linear prediction of an individual's ' breeding value and it takes the form of a multiple regression of the breeding values on all the sources of information (Falconer, 1981). Use of selection index is expected to give the most rapid improvement in breeding value for economic merit. It involves the application of selection simultaneously to all the component traits, appropriate weights being given to each trait based on its relative importance, its heritability and the genotypic and phenotypic correlations between the different traits (van Vleck, 1981). So the multiple trait selection takes into account a combination of traits according to their relative importance. Each

PAGE 28

17 individual in the population to be selected from will have an index value (score). It is possible that the highest yielding individual will not get the highest score. Although the use of selection indices in tree breeding has been limited, plant breeders have successfully utilized this procedure as an important tool in selection programs. In most selection procedures breeders use an intuitive selection index and the index value is based on the breeder ' s d eci si on . The final index value of each individual is of the form I = b,Pi + b,P^ + + b P 112 2 n n where the b. 's are the unknown factors by which each measurement of a trait (P.) is to be weighted. The problem in constructing the index is to find the best value for each weighting factor. The optimum selection index was first used by Smith (1936) in plants and later by Hazel and Lush ( 1942) in animals. They concluded that the superiority of an index increases with increasing number of traits under selection, but decreases with increasing differences in relative importance (i.e., the superiority of the selection index as a selection procedure is at a maximum when the traits of interest are considered equally important). Gain from selection for any given trait is expected to decrease as additional traits are included in the index, so the choice of traits to be included in the model must be made

PAGE 29

18 objectively (Hallauer and Miranda Fo, 1981). In addition, precision of the variance and covariance estimates is generally low, thus limiting greater use of the selection index. The genotypic value of an individual considering several traits is is of the following form H. . = a^G^. + a2G2j + .... + a.G. . where a. is the relative weight and G. ^ is the genotypic value of the j^^ individual for trait i. The objective is to make the selection so that the above combination of characters (H.) is the most desirable in terms of the economic value. Because G^ ^ are not known but are estimated from P.. (phenotypic values), H. . must be 'J I J evaluated by an index I, based on the phenotypic values such that the correlation between H and I is as great as possible. The maximization of this correlation leads to the following relationship (Turner and Young, 1969 ) : G A = P B where G is the matrix of additive genotypic variances and covariances, A is the vector of relative economic weights, P is the matrix of phenotypic variances and covariances and B is a vector of unknown values of b^. 's. The solution of the set of equations leads to the estimation of values of b^ that give the highest correlation with H. . when applied to index I (Hallauer and Miranda Fo, 1981).

PAGE 30

The genotypic and phenotypic variances and covariances of an the traits are needed for the estimation of these coefficients. Harris (1964) used approximate formulae and computer simulation to show that the predicted gains from selection for an index based on variance and covariance estimates tends to exceed the expected value of the gain. Williams (1962 ) calculated the probability of correct selection based on an index involving sample estimates of variance and covariance. He concluded that in cases in which variance and covariance estimates are poor, the index may not be highly correlated with true genetic values. So the less reliable the estimates of variances and covariances, the less reliable are the indices. The deduction leading to the superiority of indices over other methods of selection and the correctness of the estimated superiority of one index over another are dependent on having correct values for the genotypic and phenotypic variances and covariances and the relative economic weights. Since advances accomplished through selection are generally small, the breeder is forced to select towards a predicted goal. This problem of goal is not peculiar to index selection but to breeding programs in general .

PAGE 31

] MATERIALS AND METHODS The E^. grandi s base population was planted near LaBelle, Florida, in July 1977. The U.S.D.A. Forest Service, Southeastern Forest Experiment Station, installed and maintained the experimental materials through September 1984. GP0P77 contains 529 open -pol 1 i nat ed families representing four generations of selection (Table 1). One hundred and forty-four families (27%) are first -generati on, 211 (40%) are second-generation, 126 (24%) are t hi rd -generati on and 48 (9%) are fourth-generation. Geographically 69% of the families directly trace their origin to Australia (37% from New South Wales and 32% from Queensland). Twenty-one % came from South Africa, 4% from other nations and 6% of the families cannot be traced beyond Florida (Table 1). The pedigrees and origins of the 529 families in GP0P77 is given in Appendix 1. Each family typically is represented 60 times in a completely randomized single tree plot design on 17.3 hectares for a total of 31, 725 trees. Stocking of the plantation is 1,916 trees per hectare (775 trees per acre) laid out in a precise 141 rows X 225 columns grid mapped to preserve the location and pedigree of each tree. Planting beds are paired, with 2.3 m spacing within a pair 20

PAGE 32

1 21 Table 1. Eucal yptus grandi s progenies in GP0P77 by generation and origin. Generati on 1 2 3 4 Total Progeny Number 144 211 126 48 529 percent 27 40 24 9 100 * Origin NSW QLD SA Other Unknown Total Progeny Number 196 169 111 21 32 529 Percent 37 32 21 4 6 100 * NSW = New South Wales, Australia QLD = Queensland, Australia SA = Sout h Af ri ca Unknown = Mainly south Florida Other = Other nations

PAGE 33

22 and 3.5 m (11.5 ft.) between pairs. The seedlings were planted 1.8 m (6 ft.) apart along the beds. The southern half of the plantation (141 rows and columns 116 through 225) were harvested in August 1978 to obtain coppice information. The results reported in this study pertain to the southern half of GP0P77 which includes all 529 families with an average of 30 individuals per family for a total of 15,510 trees. A summary of activities in GP0P77 is given in Figure 2 and the details of the measurements obtained are summarized in Appendix 2. Study Area for Diallel Analyses The study area for the diallel analyses is located on a adjacent site near LaBelle, Florida. This E^. g r a n d i s population (GP0P73) was planted in the summer of 1973. The completely randomized single tree plot design contained 181 families or provenance collections either as imported seed lots from Africa and Australia or as seed lots from selected trees from Silver Lake Prairie orchard. In addition, 116 control 1 ed-pol 1 i nat ed families in a 15-tree diallel were established. However, due to mortality after planting and failure to coppice after harvest, a 10 X 10 half-diallel was used including only the lines that are common to GP0P77. The diallel analysis were performed on four traits:

PAGE 34

23 Date 7/1977 2/1978 8/1978 11/1978 3/1981 1/1982 12/1983 Acti vi ty GP0P77 Planted 7-Month Seedling Height South-hal f Fel 1 ed 3-Month Coppice Height 30-Month Coppice Height, DBH, No. of Stems Severe Freeze 64-Month Coppice Height, DBH, No. Stems, Frost Response, Coppi ce Qua! i ty Figure 2. Summary of activites and measurements for Eucal yptus grandi s in GP0P77.

PAGE 35

24 height at 6 months, height and DBH at 2.4 years and coppice height at 5 months after harvest at age 3.3 years. Statistical Analyses The data were analyzed using the Statistical Analysis System (SAS). Since the data were unbalanced, the variance components were estimated using the general linear model (GLM) procedure. The analysis computes the Type 1 sums of squares for each effect, equates each mean square involving only random effects to its expected value, and solves the resulting system of equations (Gaylor et al., 1970). The following random effects model was used in the analysis: Y. .. = U + G, + F(G) . V + E. .. wh er e ^ijk ~ individual observation U = overall mean G^ = effect of the i^^ generation F(G)j^.^= effect of the j^^ family within the i generati on , ^-fji^ random error associated with each observati on Individual and family tree heri tabi 1 i ti es were estimated using the procedures and formulae outlined by Wright (1976). The selection intensities used to estimate the predicted genetic gains were obtained from Namkoong and

PAGE 36

25 Snyder ( 1968). Since the coppice and frost codes were non-continuous tests were performed to see if their distribution approximates a normal distribution. Some families in GP0P77 have common ancestors, thereby leaving a potential for the occurrence of inbreeding due to mating of related individuals. Genetic gains through alternative selection strategies, were adjusted for inbreeding depression that may occur due to the presence of relatives. For each selection strategy, the inbreeding coefficient (F) was calculated for all possible matings by tracing the pedigree of each mating to its common ancestor and computing the probabilities at each level (Li, 1976). Representative pedigrees and inbreeding coefficients are shown i n Fi gur e 3 . The mathematical model used in the diallel analysis for computing the general and specific combining abilities is as f ol 1 ows : Z. . = U + G, + G. + S. . + E. . where Z^j = mean performance of the progeny of i^*^ parental line mated with j^'^ parental line U = overall mean '^(i)'^(j)^ general combining ability of the • th/.thi ^ I , . 1 (j ) parental line

PAGE 37

S. j = interaction of the i and j parental lines (specific combining abilities) E. . = random error The diallel analyses were performed according to the procedures outlined by Becker (1984). The means of crosses were used in the analysis of variance calculations. The error variance for the ANOVA based on cross means (V ) was w obtained by dividing the error variance based on individual observations (V^) by the harmonic mean number of individuals per cross. The analysis of variance table and the expected mean squares are given in Table 2. Estimation of Selection Indices : A subset of GPOP77 including only the trees that were alive at coppice age of 64 months was used to estimate variances and covariances of the five traits. The selection index derived was of the f ol 1 owi ng f orm : I = bjShgt + b2Chgt + b3Vol + b^FR + b^CQ where, b. = unknown weights Shgt = Seedling height at 7 months Chgt = Coppice height at 3 months Vol = Coppice volume at 64 months FR = Frost resilience score at 64 months CQ = Coppice quality score at 64 months

PAGE 38

27 INBREEDING COEFFICIENT GENERATION 0725 0.125 0.0625 0.03125 0.0156 0.0078 Figure 3. Representative pedigrees and inbreeding coefficients of different degrees of relationships among top 50 families of Eucal yptus grandi s in GP0P77.

PAGE 39

28 Table 2. Diallel analysis of variance for four Eucalyptus grandi s traits in GP0P73 Source Mean Squares DF H6 D24 H24 C5 Expected Mean Squares GCA 9 2.2 0.9 225 .6 4.3 V +V +8V w sea gca SCA 35 1.4 0.4 135.3 0.3 V +V w sea Error 942 1.1 0.2 43.9 0.2 V w * H6 = Seedling height at 6 months D24 = Seedling DBH at 2.4 years H24 = Seedling height at 2.4 years C5 = Coppice height at 5 months

PAGE 40

29 Table 3. Relative economic weights used in the ten selection index equations for five Eucalyptus grandis traits in GP0P77 seedling coppice coppice frost coppice Index height height volume resilience quality 7-months 3-months 64-months 1 0.35 0.75 0.75 0.75 0.35 2 0,32 0.50 0.50 0.50 1.00 3 0.50 1 .00 1.00 1.00 0.50 4 0.50 0.50 0.50 0.50 0.50 5 0.10 0.60 0.60 0.60 0.60 6 0. 75 0.35 0. 75 0.35 0.75 7 0.50 0.32 0.10 0.50 0.50 8 0.50 0.10 0.50 0.10 0.10 9 0.00 0.00 0.54 0.25 0.25 10 0.60 0.60 0.10 0.60 0.60

PAGE 41

! 1 The variances and covariances were estimated using the the procedures outlined by Van VI eck (1981). The relative economic weights for the five traits used in the selection index were based on their relative importance for bi omass production in Florida (Table 3).

PAGE 42

RESULTS AND DISCUSSION Genetic Improvement Through Recurrent Selection Four generations of selection have achieved significant geneti c i mprovement for all coppice growth characteristics studied. Comparisons of means of height, DBH and volume showed impressive gains at all ages with generations of improvement. At 7 months after planting the mean height for first-, second-, thirdand f ou rt h -g en erat i on trees was 1 .60, 1 . 79, 1 . 81 and 1.83 m, respectively (Table 4). At the same age survival was similar across generations. The same trend continued for later ages with the fourth generation families performing the best. At 64 months of age the average f i r st -g enerati on tree contained 7.04 dm^ of stem wood. Second -g enerati on trees were 206% larger averaging 3 21.54 dm , t hi rd -g enerati on trees were an additional 20% 3 larger at 25.91 dm and the fourth-generation trees had yet another 55% gain with an average of 40.16 dm'^ of wood (Table 5). Mean coppice height at 64-months for the four generations is given in Figure 4. 31

PAGE 43

32 Table 4. Eucal y ptus grandi s generation means for individual tree performances in GP0P77 through 64 months aft er harvest . Generati on Rotati on/ Trai t/Age Overall 200 Best Trees Seedl i ng -7 months Height (m) 1 .75 1 .60 1 .79 1.81 1. 83 2.93 Coppi ce -3 months Height (m) 0 .63 0 .56 0 .63 0.67 0. 71 1.69 -30 months DBH (cm) 3 .32 2 .46 3 .37 3. 74 4. 19 8.13 Hei ght (m ) 4 .30 3 .43 4 .34 4.71 5. 28 7.09 Dry Wei ght (kg) 0 .87 0 .29 1 .02 1.24 2. 00 4.16 3 Vol ume (dm ) 2 .58 1 .56 2 .85 3.24 4. 58 8.39 -64 months Height (m) 7 .36 5 .00 7 .75 8.42 9. 39 16.79 Uo n ^ cm / o . o A . U 0 7 7 7^ q 1 ft 1 o 1 7 dQ Dry Wei ght (kg) 10 .88 3 .39 11 .60 14.08 20. 10 73.13 3 Volume(dm ) 20 .16 7 .04 21 .54 25.91 40. 16 130.25 Frost Resi 1 i ence 1 .47 1 .75 1 .39 1.34 1. 25 0.00 Frost Defect 1 .43 1 .05 1 .49 1.64 1. 76 0.00 Coppi ce Qual i ty 0 .54 0 .77 0 .48 0.41 0. 40 0.00 Coppi ce Defect 0 .99 1 .28 0 .92 0.83 0. 79 0.00

PAGE 44

1 33 Figure 4. Mean seedling (7-month) and coppice height (3-, 30and 64-month) for four generations of Eucalyptus grandi s progeni es in GP0P77.

PAGE 45

34 Percent gains achieved in each generation over the preceding generation is given In Table 5. The rate of improvement achieved over the preceding generation decreased with generations of improvement. For the traits studied the highest gain was achieved in the second generation over the first. This seems to indicate that the amount genetic variation has declined over generations which is reflected in the declining realized gains. For 64-month coppice volume, the s econd -g enerati on trees had a threefold gain over the fi rst -generati on . This gain was considerably lower* for later generations. Per hectare biomass productivity differed significantly with generations with the fourt h -g enerati on families producing the most. However, the productivity was due to low survival resulting from harvest in August (Rockwood et al., 1984). Since survival was similar for all generations at 64 months, per hectare biomass yield differences were essentially the same as individual tree volume and weight differences. At 64 months after harvest the fi rst -generati on trees produced 0.7 dry mt of stem and branches , s econd -g en er at i on trees had 2.5 dry mt , third generation trees yielded 3.0 dry mt and the fourth-generation trees produced 4.3 mt per hectare per year. The effective density after accounting mortality at 64 months after harvest was 1190 tree per hectare. A projected yield of 25.7 dry mt/ha/yr for the 200 best trees I

PAGE 46

35 Table 5. Percent gains for Eucalyptus grandi s generations compared to the preceding generation for different t rai ts Generati on 200 best Trait Age 2 3 4 Trees (mos ) Seedl i ng hei ght 7 12 1 1 60 Coppi ce hei ght 3 13 . 6 6 138 Coppi ce hei ght 30 27 9 2 34 Coppi ce DBH 30 37 11 2 94 Coppi ce vol ume 30 83 13 2 83 Coppi ce hei ght 64 55 9 2 79 Coppi ce DBH 64 76 9 9 91 Coppi ce vol ume 64 206 20 5 224

PAGE 47

36 under the assumption of 100% survival is unrealistic, but it does identify the potential of clonal propagation of trees with excellent coppice growth. Variation among sources : Geographic origin of £. g ra ndi s may be important in developing good coppicing trees for southern Florida. For early seedling growth sources were generally similar in each generation. Introductions from South Africa (crosses among selected parents) were slightly taller when first grown in Florida (Table 6), but their selected offspring were virtually the same height as other sources in the second generation. For 64-month coppice height, Queensland sources improved considerably with each generation of selection beyond the initial introduction and surpassed thirdand fourth-generation means for other sources by more than 1 meter (Figure 5). The results indicate that greater gains could be achieved by concentrating the future introductions to Queensland regions of Australia. Additive genetic variation : High individual and family heri tabi 1 i ti es were observed for all the traits examined. The analysis of variance tables for the growth traits are given in Appendix 3. For 7-month seedling height, individual tree heritability was twice that of estimates reported by van Wyk ( 1977) for E^. grandi s (Table 7). High individual tree heri tabi 1 i ti es suggest that high genetic gains can be achieved through mass selection of the best individuals. The overall individual tree heri tabi 1 i ti es

PAGE 48

37 Table 6. Comparison of origin of Eucalyptus grandi s progeni es i n GPOP 77 * ** Origin ' Trait Gen . NSW QLD SA Other Unknown 7-mo 1 1 . 6bB 1 . 6aB 1 . 7aA 1 .6aB Seed! i ng Hei ght 2 1.8aA 1.8bA 1.8aA 1.8bA 1.7B { m ) 3 1.8aA 1.7bA 1.8bA 4 1.8aA 1 .8bA 64-mo 1 5.2aA 4.9aA 4.9aA 5.4aA Coppi ce Hei ght 2 7.8bA 8.2bA 7.5bA 7.7bA 7.6A ( m ) 3 8.4cB 9.8cA 8.4bB 4 7.4dB 9.9cA * NSW = New South Wales, QLD = Queensland, SA = South Africa, Other = Mainly south Florida * Origins within generations not sharing same upper case letter are significantly different at 5%. Generations within an origin not sharing the same lower case letter are significantly different at 5% level.

PAGE 49

SOURCE OF ORIGIN IXS GEN 2 ^ GEN 3 GEN 4 Mean coppice height at 64 months by generation for four sources of Eucal yptus grandi s progenies in GP0P77 (Ni;w = New South Wal es, QLD=Queens1 and , SA=South Africa, Other=Mainly from southern Florida)

PAGE 50

39 Tabl e 7 . Heri t abi 1 i ti es of Euca1 y ptus grandi s traits by and over generations Individual Heritability Family Heritabilty Generati on Generati on Trait 1 2 3 3 Overall 1 2 3 5 Overall Seedl i ng Height 0.31 0.71 ( 7-nion ) Coppi ce Height 0.30 0.68 ( 3-mon ) Coppi ce Height 0.63 0.31 0.29 0.38 0.39 0.75 0.59 0.58 0.64 0.76 ( 64-mon ) Coppi ce DBH 0.57 0.33 0.29 0.33 0.39 0.73 0.61 0.58 0.61 0.76 (64-mon) Coppi ce Volume 0.37 0.27 0.25 0.25 0.31 0.63 0.56 0.54 0.54 0.59 (64-mon )

PAGE 51

I 40 through 64 month of age after harvest ranged from 0.31 to 0.39 and the family heri tabi 1 i ti es ranged from 0.65 to 0.75. When calculated separately for each generation, the first generation trees have higher heri tabi 1 i ti es compared to the other three. This explains the high genetic gains that were realized in the second generation over the first generation. Indications are that enough additive genetic variation exists in £. grandi s to yield substantial improvement in biomass productivity in a recurrent selection program. Genetic correlations among coppice growth traits were high (Table 8). The correlation between seedling height at 7 months and coppice height at 3 months was negative but not significant. This indicates that early coppice height cannot be predicted based on early seedling height. The genetic correlation between coppice height at 3 months and coppice height and DBH at 64 months was high, suggesting that early coppice growth could be a good indicator of coppice growth at later ages. One of the assumptions for the analysis of variance is that the errors are normally distributed. The frost and coppice scores in this study were tested against a normal distribution with mean and variance equal to the sample mean and variance (SAS, 1985). The normality tests on frost and coppice scores revealed that the distribution approximates a normal distribution and therefore can be analyzed as if they were conti nuous .

PAGE 52

41 Table 8. Genetic correlations and their standard errors (in parentheses) among growth traits, frost and coppice scores of Eucal yptus grandis in GP0P77 * Tr ai t Shgt Chgt H64 D64 FD CQ CD Shgt ^OTTT 0726 OTYS (0.013) (0.012) (0.013) Chgt 0.56 0.56 (0.007) (0.007) H64 0.97 (0.001) 0.20 0.81 0.80 (0.016) (0.005) (0.005) -0.40 -0.54 (0.017) (0.013) CQ 0.80 (0.007) Shgt = Seedl i ng hei ght at 7 months Chgt = Coppice height at 3 months H64 = Coppice height at 64 months D64 = Coppice DBH at 64 months FR = Frost resilience score FD = Frost defect score CQ = Coppice quality score CD = Coppice defect score

PAGE 53

42 Variation in Frost and Coppice Scores Variation in frost scores :Considerab1e genetic variation was observed for resilience to frost as well as defects in coppice attributed to frost. The mean frost resilience score based on individual trees was 1 . 75, 1 .39, 1.34 and 1.25 for first-, second-, thirdand fourth-generation trees, respectively. However, for frost defect scores the fourth-generation trees scored the worst and the first generation trees the best (Table 4). The genetic correlation between frost resilience and frost defect scores was low (Table 8). Thus selecting individuals for frost resilience would not improve the frost defect score. Comparison of trees from different geographical sources showed that trees from New South Wales were the most frost resilient, whereas trees from South Africa scored best for frost defect score (Table 9). Therefore by concentrating future introductions from New South Wales one would be able to select for higher frost resilient trees. Individual tree heri tabi 1 i ti es for frost resilience and frost defect scores were 0.29 and 0.15, respectively. These low heri tabi 1 i t es suggest low gains for frost defect scores can be expected from mass selection. The family heri tabi 1 i ti es for frost resilience and frost defect scores were 0.71 and 0.51, respectively, suggesting that higher gains can be achieved through combined selection when compared to mass selection. However, highest gains could be achieved through combined selection of the best trees

PAGE 54

I Table 9. Mean frost and coppice scores of Eucalyptus grandi s by origin in GP0P77 Trait Origin* "NSW EJTD : SA Ot her Frost resi 1.33 1.43 1.38 1.42 1 i ence Frost defects 1.65 1.40 1.36 1.42 Coppice 0.44 0.56 0.53 0.63 qua! i ty Coppice 0.86 0.99 0.99 1.10 defects * NSW = New South Wal es QLD = Queensland SA = Sout h Af ri ca Other = Mainly from southern Florida

PAGE 55

44 from the best families. This would utilize both between-and within-family variation thereby getting a higher response to sel ecti on . Variation in coppice quality : Fourth-generation trees were also the best performers for coppice quality and coppice defect scores (Table 4). In terms of geographic origin, again the best performers were the sources from New South Wales (Table 9). Individual and family heritabi 1 i ti es for coppice quality and coppice defect scores were 0.12, 0.13 and 0.45, 0.46, respectively. Low individual tree heri tabi 1 i ti es suggest the ineffectiveness of mass selection for coppice quality. Higher family heri tabi 1 i ti es indicate the potential for improving coppice quality through family selection. Genetic correlation between the two coppice traits was 0.9. This high correlation suggests that by selecting and improving one trait, we would si mul taneoul sy be improving the other. Genetic correlations between frost and coppice scores are given in Table 8. Inbreeding and Seed Orchard Development GP0P77 could be converted into a seed orchard. Six different selection strategies including individual, combined and family selection with different intensities of selection assumed that initial introductions were not related. For each selection strategy the inbreeding coefficient (F) was calculated for all possible matings

PAGE 56

45 among the selected trees that constitute the selected popul ati on . Predicted Inbr eedi ng : Inbreedi ng coefficients among all possible crosses ranged from 0 to 0.25. The frequency of F values for a 50-family seed orchard is given in Table 10. The mean inbreeding coefficient for five selection strategies ranged from 0 to 1% (Table 11). This was attributed to the fact that only a small percentage of individuals (13% in a 50 family orchard) has an F greater than zero. In addition two or three generations have passed since direct relationship by common ancestry. The presence of related families in a seed orchard does not seem to affect the predicted gains through loss from inbreeding depression. However, when predicting the genetic gains, care should be taken to adjust the gains for potential inbreeding depression that may arise from selfing. Extensive research has been done on the frequency of natural self-fertilization that occurs in pines. Franklin (1971) working with Pinus taeda reported that an average of 2.4% of self-fertilized seed is produced. He used 25 marker-carrying trees and estimates for individual trees ranged from 0 to 13.5%. Kraus and Squillace (1964) reported evidence of selfing in clonal orchards of Pi nus el 1 i otti i . According to their observations total selfing is

PAGE 57

1 46 Table 10. Frequency of different degrees of relationship in a 50-family subset of Eucal yptus grandi s seed orchard in GPOP77 Inbreeding coefficient (F) Frequency 0.250 2 0.125 13 0.063 6 0.031 93 0.016 19 0.008 28 0.000 1064 0.005 Total 1225

PAGE 58

Table 11. Mean inbreeding coefficient (F) for five selection strategies of Eucal yptus grandi s in 6P0P77 Number of Selected Families Mean F {%) 300 0 .00 100 0 .00 50 0. .12 30 0 .90 10 0 .97

PAGE 59

48 selfing is estimated to be 8% in a 20 clone seed orchard. Wright (1976) reported this figure to average about 7%. Squillace and Bingham ( 1958) recommended the elimination from seed orchards of individuals with high degree of preference for selfing. Percent selfing in £. g r a n d i s occurs at a higher rate than in most pines. In adjusting the predicted gains for selfing, Hodgson's ( 1976) and Eldridge's ( 1978) estimates of inbreeding depression were used. They predicted a height loss of 8% to 49% for selfed individuals with an inbreeding coefficient of 0.50. Using the estimated 30% selfing that occurs in grandi s . it can be deduced that 30% of the offspring from the seed orchard will have an F of 0.50 and the remaining 70% will have the estimated F for a given strategy. Thus for each selection strategy the weighted mean inbreeding was calculated for both selfing and mating of relatives. Using the conservative approach of the range of inbreeding depression reported for E^. g r a n d i s , a 50% loss in height growth for selfed progenies (F = 0.5) amounts to 10% loss for every F of 0.10. So for a given selection strategy the predicted genetic gains were reduced by a proportion equal to the mean inbreeding coefficient. Predicted geneti c gai ns : The predicted genetic gains (adjusted for inbreeding) that could be achieved through different selection strategies are given in Table 12.

PAGE 60

49 Table 12. Predicted genetic gains in 64-month coppice volume for alternative improvement strategies of Eucal yptus grandi s in 6P0P77 Sel ecti on St rat egy Geneti c gai n {%) Mass Selection 200 best trees 80 300 best trees 69 Combined Selection 10 top fami 1 i es 41 (30 t r ees per fami 1 y ) 30 t op f ami 1 i es 61 (10 t r ees per f ami 1 y ) 100 top fami 1 i es 90 (3 t r ees per fami 1 y ) 300 top fami 1 i es 86 (1 tree per family) Vegetative Propagation clonal propagation of 441 200 best trees * Gain over the population mean

PAGE 61

50 The six strategies give an effective orchard size of approximately 300 trees for an area of approximately 8.5 hectares. At 64 month coppice age, the highest gain (90%) in volume over the population mean is predicted from a combined selection of the top 100 families with 3 trees per family. Comparable gains also are predicted for other selection strategies: 86% for family selection of the top 300 families; 80% for mass selection of the top 200 trees. Prediction of higher gains through combined and family selection compared to mass selection agrees with the results* reported for ^ robusta in Florida (Dvorak et al., 1981). Diallel Analyses Results from the 10 X 10 half-diallel cross identified some of the parental lines with a high potential for genetic improvement. The 10 parents analyzed demonstrated considerable genetic variation for all the four traits studied. For seedling growth significant non-additive genetic variance was found . On the other hand, for coppice height the additive variance was significant. Genetic improvement through the use of non-additive genetic variance seems to be advantageous for early seedling growth. Height at 6 months : At 6 months after planting the best cross (1188X1189) had a mean height of 18.4 dm, while the poorest performance was demonstrated by cross 1178 X 1180 (Table 13). Individual tree heritability was 0.02,

PAGE 62

Table 13. Mean seedling height (dm) at 6 months for 45 crosses of Eucalyptus qrandis in GP0P73 Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 Total 198 14.2 15.6 14.6 14.4 14.8 16.9 14.9 12.8 14.6 132.8 1178 16.7 12.5 15.2 15.7 15.9 17.0 15.3 14.9 137.5 1179 13.4 14.9 14.9 13.8 17.2 12.7 14.7 133.9 1180 13.9 15.7 14.4 13.8 13.6 14.6 126.5 1182 13.9 15.0 14.1 14.7 14.2 130.1 1183 15.1 14.9 17.4 13.8 135.9 1185 16.1 14.8 15.1 133.8 1188 18.4 14.4 140.7 1189 _ 14.1 133.8 1191 130.3

PAGE 63

52 lower than that estimated by van Wyk (1976) for the same £. grandi s at the same age. The variance due to GCA (0.099) was lower than due to SCA variance (1.387). This suggests that greater improvement could be achieved through the use of non-additive genetic variation like clonal propagation than from the traditional method of improving through mass selection. Considerable variation also was observed in the estimates of GCA among all parents. The GCA estimates and their standard deviations for the 10 parents are given Table 14. Family 1188 has the highest (0.8510) and family 1180 had the lowest (-0.9165). The 45 crosses from the diallel exhibited considerable variation in SCA (Table 15). Cross 1179 X 1189 has the lowest SCA while cross 1188 X 1189 had the highest. The above results seem to indicate that parent 1188 is a very good general combiner and can be used as one of the parents in crosses to produce superior progeny. Cross 1188 X 1189 is the best cross among all the 45 crosses and can be used to produce better producing offspring for improving biomass production. Hei ght at 2.4 years : When compared with the performances of families for height at 6 months, considerable changes for the performances of parents as well as for crosses were observed for the same trait at age 2.4 years. Parents that ranked low for height at 6 months did relatively better and hence were ranked considerably higher for height at 2.4 years and viceversa. This suggests that

PAGE 64

53 Table 14. General combining ability (GCA) and associated variances of 10 Eucal y pt us grandi s families for seedling height at 6 months and coppice height at five months in GP0P73 Seedling height Coppice height Family at 6 months at 5 months 6CA Vari ance GCA Vari ance 198 -0.132 -0.100 -0.23 0.03 1178 0.449 0.083 0.87 0 . 74 1179 0.000 -0.118 0.28 0.06 1180 -0.917 0. 722 1.28 1 .62 1182 -0.473 0.106 -0.25 0.04 1183 0.256 -0.052 -0.57 0.31 1185 0.420 0.059 -0.49 0.22 1188 0.851 0.607 -1 .24 1.52 1189 0.003 -0.118 0.31 0.08 1191 -0.452 0.086 0.02 0.02

PAGE 65

54 Table 15. Specific combining ability of 45 Eucalyptus grandi s crosses for seedling hei ght at 6 months in GPOP73 Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 198 -0.96 0.81 0.80 0.17 -0.25 0.77 -0.68 -1.90 0.27 1178 1.35 -1.93 0.30 0.13 0.19 0.81 -0.02 0.09 1179 -0.53 0.48 -0.17 -1.50 1.40 -2.20 0.26 1180 0.45 1.50 0.02 -1.05 -0.34 1.05 1182 -0.98 0.19 -1.17 0.34 0.20 1183 -0.48 -1.08 -0.30 -0.95 1185 -0.04 -0.46 0.27 1188 2.62 -0.89 1189 -0.30

PAGE 66

] 55 height performances of families at early ages are not good indicators of height at later ages. Thus selection for height at early ages may lead to elimination of better parents. Cross 198 X 1179 demonstrated the highest mean height at 2.4 years at 86.87 dm, and the lowest mean of 41.15 dm was observed for cross 1179 X 1183 (Table 16). Heritability was slightly higher (0.06) than at age 6 months but the additive variance was not high enough to have a higher GCA variance than the SCA variance. The GCA variance was 11.29 compared to the SCA variance of 91.45. Family 1178 had the highest GCA (7.72) and parent 1182 had the lowest (-7.10) (Table 17). Table 18 summarizes the considerable variation observed in the SCA variances for the 45 cr OSS es . DBH at 2.4 years : The amount of variation in OBH at 2.4 years observed among the 45 crosses was high (Table 19). The mean DBH among the crosses ranged from 3.74 cm for cross 1179 X 1183 to 6.56 cm for cross 1183 X 1189. As observed for height at early age the variance of SCA (0.26) was higher than the variance of GCA (0.06), suggesting again the use of non-additve genetic variation in the genetic improvement for DBH at age 2.4 years is more productive. The estimates of GCA and their variances for the 10 parents is given in Table 17. Line 1185 had the lowest GCA (-0.39) and line 1189 has the highest (0.47). The GCA rankings of parents for height at 6 months and DBH at 2.4 years changed considerably. Individual tree heritability was low which is

PAGE 67

56 Table 16. Mean seedling height (dm) of 45 Eucalyptus grandi s crosses at 2.4 years in G P 0 P 7 3 Par ent 1 1 7 O 1 i / O 1 1 7 Q i i / y lion 1 1 Q 9 L IOC. 1 1 Q Q 1 1 Q K 1 1 Q Q i i OO 1 1 QQ 1 1 Q 1 113 1 1 0 L a 1 198 64.8 86.9 61.9 59.7 80.3 64.5 68.2 57.7 71.3 615.3 1178 83.3 67.3 75.0 70.8 71.9 87.0 68.0 77.6 665.6 1179 76.9 65.0 41.2 44.8 66.6 63.9 70.0 597.8 1180 67.3 83. 7 67.7 73.2 62.1 70.0 629.9 1182 47.8 59.1 47.3 72.1 53.8 547.0 1183 49.2 47.6 84.9 54.6 559 .9 1185 51 .1 77.6 63.5 549.2 1188 92.9 67.1 600.9 1189 83.0 662 .1 1191 610.2

PAGE 68

Table 17. GCA and its variances of 10 Eucal y ptus g r a n d i s parents for seedling height and DBH at age 2.4 years in GP0P73 Hei ght DBH pa p p n 1" GCA Vari ance GCA Va ri ance 198 1 .44 -2.86 -0.14 0.00 1178 7.72 54.66 0.43 0.17 1179 -0.75 -4.37 0.20 0.02 1180 3.26 5.69 0.22 0.03 1182 -7.10 45.48 -0.38 0.13 1183 -5.49 25.16 -0.23 0.04 1185 -6.82 41.58 -0.39 0.14 1188 -0.36 -4.81 -0.05 -0.01 1189 7.29 48.21 0.47 0.20 1191 0.80 -4.29 -0.03 -0.02

PAGE 69

Table 18. Specific combining ability for seedling height at 2.4 years for 45 Eucal yptus grandi s crosses in GP0P73 Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 198 -11.5 19.6 -9.9 -1.7 17.2 2.8 0.1 -18.1 2.0 1178 9.2 -10.8 7.3 1.4 3.9 12.6 -7.9 8.3 1179 7.3 5.7 -19.7 -14.8 0,6 -9.8 2.4 1180 4.1 18.8 4.1 3.2 -15.5 -1.2 1182 -6.7 5.9 -12.3 4.9 -7.0 1183 -5.6 -13.6 16.0 -7.8 1185 -8.8 10.0 2.5 1188 18.8 -0.5 1189 7.7

PAGE 70

59 Table 19. Mean seedling DBH (cm) of 45 Eucal yptus grandi s crosses in GP0P73 at age 2.4 years Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 Total 198 " 4 87 A H , fid > OH A H , . o u c 0 • > 3 4> / 0 OA > 0 0 c D . A K 0 1 1178 6. .39 5. .01 5. .22 5. .57 5. .62 5, .10 5. .50 5, .48 49, .76 1179 5, .94 5, .39 3. .74 4. .32 5, .34 5, .09 5, .42 47, .88 1180 5. .31 6. .25 4. .91 5. .44 5. .11 5, .24 48, .05 1182 4. .07 4. .55 3, .96 5. .27 4, .31 42, .38 1183 4, ,14 4. .05 6. ,56 4. .52 44. .44 1185 4. .40 5. ,57 4. .91 43. .17 1188 6. ,40 5. ,19 45. .92 1189 5. ,96 50. .02 1191 46, ,09

PAGE 71

60 Table 20. Specific combining ability of 45 Eucal y ptus grandi s crosses for 2.4-year seed! i ng DBH in GP0P73 Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 198 -0 57 1 .04 -0 39 -0 .22 0 .76 0 .13 0 .08 -0 .91 0 .08 1178 0 .61 -0. 79 0. 70 0. 79 1 .00 1 .14 0 .03 0 .50 1179 0. 38 0 .54 -1 37 -0 .63 0 .04 -0 .72 0 .10 1180 0. 44 1. 12 -0. 06 0. 12 -0 72 -0 .09 1182 -0. 35 0 .29 -0 65 0 .15 -0 .32 1183 -0 38 -0. 82 1. 18 -0. 37 1185 -0. 31 0. 35 0. 18 1188 0. 84 0. 12

PAGE 72

61 also indicated by the magnitude of the GCA variance. Thus we cannot expect much improvement through selection as the amount of additive genetic variation present for this trait is limited. The SCA for the 45 crosses given in Table 20 shows that cross 1183 X 1189 has the highest SCA (1.18), with cross 1179 X 1183 having the lowest (-1.37). Coppice height at 5 months : Mean coppice height of the 45 crosses 5 months after roguing at 3.3 years ranged from 2.05 dm for cross 1185 X 1188 to 6.13 dm demonstrated by cross 1179 X 1180 (Table 21). The three parents that ranked-highest at 2.4-year seedling height were still the top three performers for coppice height at 5 months. Unlike the other three traits studied, the variance in coppice height at 5 months seems to be predominantly additive in nature. This is demonstrated by the high heritability (0.45) as well as by the higher GCA variance (0.497) than the SCA variance (0.131). This indicates that if improvement in coppice growth is to be achieved, the use of additive genetic variance is more efficient, since the non-additve genetic variance for this trait is less. The GCA variance of the 10 parents ranged from -1.24 for parent 1188 to 1.28 for parent 1180 (Table 14). Cross 11 78 X 11 79 exhibited the highest SCA variance (0.91) and cross 1178 X 1188 had the lowest at -1.01 (Tabl e 22) . The results from the analyses of seedling traits suggest that non-additive variance is a very important

PAGE 73

1 Table 21. Mean coppice height (dm) of 45 Eucalyptus grandi s crosses at 5 months in G P 0 P 7 3 Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 Total 198 4.02 4.41 4.56 3.56 2.73 3.28 2.80 3.45 3.56 32.37 1178 5.86 5.34 4.56 4.86 4.55 2.42 4.96 4.62 41.19 1179 6.13 3.28 2.78 3.40 2.13 3.56 4.87 36.42 1180 5.54 4.70 4.04 4.18 4.73 5.25 44.47 1182 2.60 3.02 2.43 4.08 3.15 32.22 1183 2.70 2.15 4.08 3.06 29.66 1185 2.05 4.06 3.18 30.28 1188 3.64 2.48 24.28 1189 4.15 36.71 1191 34.32

PAGE 74

63 Table 22. Specific combining ability of 45 Eucalyptus qrandis crosses for 5-month coppice 1 n GP0P73 Parent 1178 1179 1180 1182 1183 1185 1188 1189 1191 198 -0.43 0.56 -0.30 0.24 -0.27 0.20 0.47 -0.44 -0.03 1178 0.91 -0.62 0.13 0.75 0.37 -1.01 -0.03 -0.07 1179 0.77 -0.55 -0.73 -0.19 -0.71 -0.83 0.78 1180 0.70 0.18 -0.55 0.34 -0.67 0.15 1132 -0.39 -0.04 0.12 0.19 -0.42 1183 -0.04 0.16 0.53 -0.19 1185 -0.02 0.44 -0.15 1188 0.77 -0.10 1189 0.02

PAGE 75

1 64 source of variation that could be exploited for improving seedling growth of E_j_ g r a n d i s . Low heri tabi 1 i ti es suggest the relatively low gains possible through mass selection. This however is not the case for the early coppice height; High heri tabi Tittes fori;his trait suggest good gains can be achieved through selective breeding. In a sel ectton program that utilizes additive genetic vari ance, the variance component attributed to GCA (V ) is an g ca ' indicator of the amount of additive genetic variance. Table 23 shows the variance of SCA (V ), variance of GCA sea (Vg^^), the additive genetic variance (V^) and the dominance variance (Vp,). In this study V was u g c a significant for DBH at 2.4 years and coppice height at 5 months and the V^^^ was found to be significant for both height and DBH at 2.4 years. For all the traits except coppice height, V^^^ was lower than V^^^. Consequently, the additive genetic variance was less than the dominance variance for these traits, suggesting that breeding methods that utilizes non-additive genetic variation are more efficient than selective breeding programs that utilize additive genetic variation. A good example of such a program utilizing non-additive genetic variation is clonal propagation of selected individuals. However, for early coppice height, significant V^^,, in addition to high g ca ^ heri tabi 1 i ty, suggests that additive genetic variation can be successfully utilized through the use of traditional

PAGE 76

Table 23. Estimates of variance components and genetic variances for four growth traits in Euca1 y pt us grandi s Variance component Trait G£A JcK Additive Domi nance Seedl i ng Height at 0.06 0.22 0.24 0.88 6 months Seedl i ng ^ Height at 14.71 82.51 58.84 330.04 2.4 years Seedl i ng ^ DBH at 0.07 0.24 0.29 0.94 2.4 years Coppi ce height 0.50 0.11 1.99 0.44 at 5 months * and ** Siginificant at 5% and 1%, respectively

PAGE 77

Table 24. Rankings of general combining ability of 10 Eucal y ptus g r a n d i s parents from GP0P73 Parent Rank Seedling Seedling Seedling DBH Coppice height at height at at 2.4 years height at 6 months 2.4 years 5 months 1 1188 1189 1178 1180 2 1185 1178 1189 1178 3 1178 1180 1180 1189 4 1183 1179 198 1179 5 1189 1191 1191 1191 6 1179 1188 1188 198 7 198 198 1179 1182 8 1191 1183 1183 1185 9 1182 1182 1185 1183 10 1180 1185 1182 1188

PAGE 78

67 breeding procedures, higher genetic improvement can be achieved through the use of additive genetic variance. The GCA rankings for the four traits of the 10 parents is given in Table 24. The rankings of parents change considerably through age as well as among traits. Spearman's Rank correlations of the 10 parents based on their GCA rankings is given in Table 25. These correlations are an indication of how well the rankings of parents for the 4 traits are correlated. Significant correlations were observed for coppice height at 5 months with DBH at 2.4 years (0.79) and with seedling height at 2.4 years (0.72). The correlation was also significant between height and DBH at 2.4 years. Seedling height at 6 months was not correlated with any other trait. This suggests that 5 months is too early an age for the prediction of growth at later ages. The highly significant correlation between coppice height and seedling height at 2.4 years suggest that by selecting parents based on seedling height at 2.4 years, we would also be selecting for coppice performance. Sel ecti on Indi ces Future options for grandi s in Florida include conversion of GP0P77 into seedling seed orchard. To aid in the selection process, a selection index equations were developed to select individuals based on seedling height at 7 months, coppice height at 3 months, coppice volume, coppice quality and frost resilience scores at 64 months.

PAGE 79

Table 25. Spearman rank correlations for rankings of Eucal yptus grandi s parents for four growth traits in GP0P73 Height at DBH at Coppice height 2.4 years 2.4 years at 5 months Height at 0.05 -0.02 -0.42 6 months Height at 0.87** 0. 72* 2.4 years DBH at 0.79** 2.4 years * and ** Significant at 5% at 1%, respectively

PAGE 80

69 The phenotypic variances and covariances, genotypic variances and covariances and the relative weights (b. 's) for the ten index equations are given in Tables 26, 27 and 28 respectively. Considerable range in the values of indices found between the individuals agrees with the significant -genetic variation found in the traits. To test the significance of the relative economic weights assigned for each index equation, the ten index values obtained for each individual were subjected to Spearman's rank correlations. The results showed highly significant correlations between the ten index values within an individual suggesting that the economic weights assigned did not effect the rankings of the individuals. There are however some limitations recognized in the design of GP0P77. Slow growing trees at early stages may be suppressed by the fast growing ones. This could explain the higher growth differential between earlier and later generation trees at 54 months. Another limitation is the effect of microsite variation that is not accounted for the design. This may inflate the variance components and consequently the predicted gains. Such problems can be avoided by using a randomized block design. The experiment could be conducted on different sites thereby getting an estimate of genotype X environment interaction which seems to be important. Appendix 4 gives a listing of the 300 selected trees based on index 3 values.

PAGE 81

Table 26. Phenotypic variances (diagonal) and covariances (off diagonal) for Eucalyptus grand is traits used in selection index equati ons . * ** *** Frost Shgt Chgt Cvol ResilCoppice i ence qual i ty Shgt 18.10 12. 15 618. 11 -0. 19 -0 .02 Chgt 1127. 14 15085. 96 -2. 40 -4 .21 Cvol 817054. 98 120. 37 139 .03 Frost r esi 1 i 0. 28 0 .24 nee Coppi cequality 0.58 * Shgt = Seedling height at 7 months * Chgt = Coppice height at 3 months * Cvol = Coppice volume at 64 months

PAGE 82

Table 27. Genotypic variances (diagonal) and covariances (off diagonal) for Eucal yptus grandis traits used in selection index equations in GP0P77. * ** *** Frost resiCoppice Shgt Chgt Cvol lience Quality Shgt 5.69 -11.56 173.01 -0.17 -0.08 Chgt 332.32 5763.93 -1.16 -1.14 Cvol 256449.50 -495.00 -351.40 Frost resili0.08 0.06 nee Coppice 0.07 qual i ty * = Seedling height at 7 months ** = Coppice height at 3 months *** = Coppice volume at 64 months

PAGE 83

72 Table 28. Selection index coefficients for five Eucalyptus grandi s growth traits for ten equations in 6P0P77. I nd ex Shot Chgt Vol CO 1 -29.03 -8.00 0.66 -1781. 50 72.96 2 -19.36 -5.33 0.44 -1187. 66 48.64 3 -38.69 -10.66 0.88 -2375. 34 97.29 4 -19.27 -5.33 0.44 -1187. 72 48.65 5 -23.27 -6.39 0.53 -1424. 91 58.45 6 -28.51 -8.10 0.66 -1779. 70 72.47 7 3.93 -1.01 0.09 238. 51 10.04 8 -18.88 -5.44 0.44 -1186. 10 48.08 9 -20.45 -5.89 0.47 -1279. 80 51.92 10 4.18 -0.94 0.09 239. 96 10.38 * Shgt = Seed! i ng hei ght at 7 months Vol = Coppice Volume at 64 months FR = Frost resi 1 i ence CQ = Coppi ce quality

PAGE 84

CONCLUSIONS Considerable gains have been achieved through four generations of selection of £. grandi s for increased coppice stem size, frost resilience, coppice quality and form. Genetic variation in the base population still seems to be high enough to expect further gains in biomass productivity through continued recurrent selection. In coppice volume production at 64 months after harvest, fourth-generation families produced nearly 100 % more wood than the overall population mean and nearly 47 % more when compared to the third generation families. When the best trees were rated against f ourt h -g enerati on families, nearly two-fold differences in traits such as height and DBH were common. On a coppice volume basis, the best trees were more than three times larger at 64 months. This suggests an obvious potential for improvement in wood productivity through clonal propagation of selected individuals to capture the full genetic potential. Differences in biomass productivity between sources seems to be an important factor that should be considered in future introductions. For coppice growth at 64 months sources from Queensland performed the best. However, for 73

PAGE 85

74 frost resilience and coppice quality sources from New South Wales performed the best. Individual and family heri tabi 1 i ti es were higher than those reported for E^. grandi s . High heri tabi 1 i ti es indicate the potential of genetic improvement through mass and family selection. Inbreeding depression resulting from the mating of related individuals appears negligible with no significant effect on predicted genetic gains. The inbreeding coefficient (F) of all possible crosses in future seed orchard for GPOP77 ranged from 0 to 0.25. The mean F for all the matings was considerably low (0 to 1%). This was attributed to the low percentage of related individuals and the passing of two-three generations since direct relationship due to common ancestry. However, percent selfing occurring in grandi s is significantly higher than pines. Based on published estimates of height loss of 50 % for selfed progeny, the predicted gains in volume were reduced by a factor proportional to the estimated mean F for the next generation. A 90 % gain in coppice volume at 64 months is predicted for a combined selection of the top 100 families with three trees per family. Other selection strategies also had high predicted genetic gains. The amount of non-additive genetic variation in seedling growth is found to be high in 10 parents of grandi s studied. Parental lines 1178, 1180 and 1189 consistantly produced progeny that ranked high for seedling as well as coppice growth at later ages. The variance of

PAGE 86

1 75 SCA (V ) was higher than variance due to general S C a combining ability (Vg^^) for the seedling growth traits suggesting that genetic improvement can be achieved through techniques that utilize non-additive genetic variance. The genetic superiority exhibited by the top 200 trees in GP0P77 for 64-month coppice height provide a good source for clonal propagation. However, for early coppice height the V g c a was higher than V^^^ . Significant V^^^ along with high heri tabi 1 i ti es suggest that improvement for coppice height can be better achieved through the use traditional selectionprocedures. Spearman's rank correlations of the 10 parents showed high correlation between height at 2.4 years and early coppice height. This suggests that coppice height can be improved by selecting the parents based on the seedling height growth at 2.4 years. Selection indices were developed based on ten different economic weights being assigned to: seedling height at 7 months, coppice height at 3 months, coppice volume, frost resilience and coppice qualityat 64 months. Spearman's rank correlations of the ten index values of an individual was highly significant indicating no significant effect due to relative economic weights assigned. Based on these selection index values, individuals from GP0P77 can be selected and converted to a seedling seed orchard to provide a source for the genetic material for next generation. The best strategy for orchard development would be to retain three trees of the top 100 families.

PAGE 87

76 Predicted genetic gains for this strategy was the highest (90%) . With continued introductions to broaden the genetic base and selection for growth and frost resilience, the future looks promising for £. grandi s t o be a prime hardwood species in Florida for woody biomass production.

PAGE 88

APPENDIX 1. PEDIGREES AND ORIGINS OF 529 FAMILIES USED IN GP0P77 Table 1.1 Pedigrees and origin* of 529 Eucalyptus grandis families in GP0P77. Pedigree Origin Generation Vacant UNPLANTED SPOT 2 88) 836) 295 SLP)IMOK)GYMPIE Q 3 90) 836) 295 SLP)IMOK)GYMPIE Q 3 97) 293 SLP)URUNGA N 2 101) 836) 295 SLP)IMOK)GYMPIE Q 3 123) 294 BPP)COFFS HRBR N 2 125) 294 BPP)COFFS HRBR N 2 186) 293 SLP)URUNGA N 2 293 URUNGA N 1 294 COFFS HRBR N 1 295 GYMPIE Q 1 296 MID-NORTH COAST N 1 298 MISIONES, ARGENTINA 1 586 OSCEOLA CO 1 587 OSCEOLA CO 1 635 INDIAN RIVER CO 1 674 PLANT CITY 1 675 PLANT CITY 675 TAMPA 819 CONCORDIA ARGENTINA 829 LAKE HELL'N BLAZES 834 BAINBRIDGE GA 842) 5) 293 G73)FFF)URUNGA N 3 844) 18) 293 G73)FFF)URUNGA N 3 77

PAGE 89

Table 1.1 continued. Pedi gr ee Ori gi n Gen erat i on 845) 24) 293 G73)FFF)URUNGA N 3 846) 24) 293 G73)FFF)URUNGA N 3 847) 24) 293 G73)FFF)URUNGA N 3 848) 24) 293 G73)FFF)URUNGA N 3 849) 26) 293 G73)FFF)URUNGA N 3 850) 26) 293 G73)FFF)URUN6A N 3 853) 72) 293 G73)FFF)URUNGA N 3 854) 72) 293 G73)FFF)URUNGA N 3 856) 73) 293 G73)FFF)URUNGA N 3 857) 73) 293 G73)FFF)URUNGA N 3 858) 73) 293 G73)FFF )URUNGA N 3 859) 75) 293 G73)FFF)URUNGA N 3 860) 75) 293 G73)FFF)URUNGA N 3 861) 76) 293 G73)FFF)URUNGA N 3 864) 77) 293 673)FFF)URUNGA N 3 865) 77) 293 G73)FFF)URUNGA N 3 867) 78) 293 G73)FFF)URUNGA N 3 868) 78) 293 G73)FFF)URUNGA N 3 869) 79) 293 G73)FFF)URUNGA N 3 870) 79) 293 G73)FFF)URUNGA N 3 871) 79) 293 G73)FFF)URUNGA N 3 873) 80) 293 G73)FFF)URUNGA N 3 874) 80) 293 G73)FFF)URUNGA N 3 875) 80) 293 G73)FFF)URUNGA N 3

PAGE 90

1 Tabl e 1 . 1 conti nued . Pedi gree " Ori gi n Generati on 876) 97 ) 293 G73)SLP)URUNGA N 3 877) 97 ) 293 G73)SLP)URUNGA N 3 879) 97 ) 293 G73)SLP)URUNGA N 3 880) 97 ) 293 G73)SLP)URUNGA N 3 881) 97 ) 293 G73)SLP)URUNGA N 3 883) 107 ) 293 G73)SLP)URUNGA N 3 884) 107 ) 293 G73)SLP)URUNGA N 3 885) 109 ) 293 G73)SLP)URUNGA N 3 887) 109 ) 293 G73)SLP)URUNGA N 3 888) 109 ) 293 G73)SLP)URUNGA N 3 889) 118 ) 293 G73)SLP)URUNGA N 3 890) 119 ) 293 G73)SLP)URUNGA N 3 892) 119; 293 G73)SLP)URUNGA N 3 893) 273 G73)PINECREEK N 2 894) 293 G73)URUNGAN 2 895) 293 G73)URUNGA N 2 896) 81) 294 G73)B IG)C0FFS HRBR N 3 897) 81) 294 G73)BIG)C0FFS HRBR N 3 898) 84) 294 G73)BIG)C0FFS HRBR N 0 899) 84) 294 G73)BIG)C0FFS HRBR N 3 900) 84) 294 G73)BIG)C0FFS HRBR N 3 901) 85) 294 G73)BIG)C0FFS HRBR N 3 902) 85) 294 G73)BIG)C0FFS HRBR N 3 903) 85) 294 G73)BIG)C0FFS HRBR N 3

PAGE 91

Table 1.1 continued. Pedi gree Uri gi n Generati on 904) 112) 294 G73)SLP)C0FFS HRBR N 3 905) 112) 294 G73)SLP)C0FFS HRBR N 3 906) 113) 294 673)SLP)C0FFS HRBR N 3 907) 113) 294 G73)SLP)C0FFS HRBR N 3 908) 114) 294 G73)SLP)C0FFS HRBR N 3 909) 114) 294 G73)SLP)C0FFS HRBR N 3 910) 114) 294 G73)SLP)C0FFS HRBR N 3 911) 114) 294 G73)SLP)C0FFS HRBR N 3 912) 114) 294 G73)SLP)C0FFS HRBR N 3 913) 114) 294 G73)SLP)C0FFS HRBR N 3 914) 115) 294 G73)SLP)C0FFS HRBR N 3 915) 115) 294 G73)SLP)C0FFS HRBR N 3 916) 116) 294 G73)SLP)C0FFS HRBR N 3 917) 116) 294 G73)SLP)C0FFS HRBR N 3 918) 116) 294 G73)SLP)C0FFS HRBR N 3 919) 116) 294 G73)SLP)C0FFS HRBR N 3 920) 117) 294 G73)SLP)C0FFS HRBR N 3 921) 117) 294 G73)SLP)C0FFS HRBR N 3 923) 117) 294 G73)SLP)C0FFS HRBR N 3 924) 117) 294 G73)SLP)C0FFS HRBR N 3 925) 117) 294 G73)SLP)C0FFS HRBR N 3 926) 117) 294 G73)SLP)C0FFS HRBR N 3 927) 117) 294 G73)SLP)C0FFS HRBR N 3 928) 120) 294 G73)SLP)C0FFS HRBR N 3

PAGE 92

Table 1.1 continued Pedi gree ~~~~~~~~ Ori gi n Generati on 931 ) 120 J 294 G73 )SLP)COFFS HRBR N 3 933 ) 121 J 294 G73 )BPP)COFFS HRBR N 3 934 ) 121 ) 294 G73 )BPP)COFFS HRBR N 3 935) 121 ) 294 G73 )BPP)COFFS HRBR N 3 936) 121 ) 294 G73 )BPP)COFFS HRBR N 3 93 7) 121 ) 294 G73 )BPP)COFFS HRBR N 3 938) 123 ) 294 G73 )BPP)COFFS HRBR N 3 941 ) 125) 294 G73 )BPP)COFFS HRBR N 3 943 ) 126) 294 G73 )BPP )COFFS HRBR N 3 944 ) 12 7) 294 673 )BPP)COFFS HRBR N 3 946 ) 128) 294 G 73 )BPP )COFFS HRBR N 3 94 7) 128) 294 G73 )BPP )COFFS HRBR N 3 948 ) 129) 294 G73 iB PP )COFFS HRBR N 3 9 49) 129 ) ^ r\ A 294 673' BPP)COFFS HRBR N 3 950 ) 129 ) 294 673] BPP)COFFS HRBR N 3 9 52 ) 132 ) 294 673] SFDP)COFFS HRBR N 3 953 ) 132 ) 294 673 SFDP )COFFS HRBR N 3 954) 132 ) 294 673) SFDP)COFFS HRBR N 3 955) 132 ) 294 673; SFDP)COFFS HRBR N 3 956) 132) 294 673) SFDP)COFFS HRBR N 3 958) 134) 294 673) SFDP)COFFS HRBR N 3 959 ) 134) 294 673) SFDP)COFFS HRBR N 3 960) 134) 294 673) SFDP)COFFS HRBR N 3 961) 134) 294 673) SFDP)COFFS HRBR N 3

PAGE 93

1 82 Tabl e 1.1 conti nued . Hedi gree ~ ' Origin Generati on rt ^ o \ 962 ) 134 ) 294 G73)SFDP)C0FFS HRBR N 3 963 ) 136) 294 G73)SFSP)C0FFS HRBR N 3 964 ) 137) 294 G73)SFSP)C0FFS HRBR N 3 y 65 ) 138) 294 G73)SFSP)C0FFS HRBR N 3 966 ) 138) 294 G73)SFSP)C0FFS HRBR N 3 96 7) 138) 294 G73)SFSP)C0FFS HRBR N 3 968) 138) 294 G73)SFSP)C0FFS HRBR N 3 969 ) 138) 294 G73)SFSP)C0FFS HRBR N 3 9 71) 228) 294 G73)SLHQ)C0FFS HRBR N 3 9 72 ) 228) 294 G73)SLHQ)C0FFS HRBR N 3 9 73 ) 228) 294 G73)SLHQ)C0FFS HRBR N 3 9 74 ) 234)1736) 294 G73)F69)B IG)COFFS HRBR N 4 975) 234)1736) 294 G73)F69)B IG)COFFS HRBR 4 976) 234)1736) 294 G73)F69)BIG)C0FFS HRBR 4 9 77) 235)1736) 294 G73)F69)B IG)COFFS HRBR 4 9 78 ) 235)1736) 294 G73)F69)BIG)C0FFS HRBR 4 9 79 ) 235)1736) 294 G73)F69)BI6)C0FFS HRBR 4 981 ) 236)1736) 294 G73)F69)BIG)C0FFS HRBR 4 982) 236)1736) 294 G73)F69)BIG)C0FFS HRBR 4 983) 237)1736) 294 G73)F69)BIG)C0FFS HRBR 4 984) 237)1736) 294 G73)F69)BIG)C0FFS HRBR 4 985) 237)1736) 294 G73)F69)BIG)C0FFS HRBR 4 986) 86) 835) 295 G73)SLP) IMOK)GYMPIE Q 4

PAGE 94

Tabi e 1 . 1 conti nued . Pedi g r ee Ori gi n Generati on 987) 86) 835) 295 G73)SLP) IMOK)GYMPIE Q 4 988) 86) 835) 295 G73)SLP) IMOK)GYMPIE Q 4 989) 96) 835) 295 G73)SLP) IMOK)GYMPIE Q 4 990) 96) 835) 295 G73)SLP) IMOK)GYMPIE Q 4 991) 96) 835) 295 G73)SLP) IMOK)GYMPIE Q 4 992) 96) 835) 295 G73)SLP) IMOK)GYMPIE Q 4 993) 96) 835) 295 G73)SLP) IMOK)GYMPIE Q 4 994) 88) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 996) 88) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 997) 88) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 998) 88) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1000) 88) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1001) 88) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1002) 90) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1003) 90) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1004) 90) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1005) 91) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1006) 91) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1007) 91) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1009) 91) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1010) 101) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1011) 101) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1012) 101) 836) 295 G73)SLP) IMOK)GYMPIE Q 4 1016) 92) 837) 295 G73)SLP) IMOK)GYMPIE Q 4

PAGE 95

Tab! e 1 .1 conti nued . Pedi gree Ori gi n Generati on 1017) 92) 837) 295 G73)SLP) IM0K)GYMPIE Q 4 1018) 92) 837) 295 G73)SLP) IMOK)GYMPIE Q 4 1019) 92) 837) 295 G73)SLP) IMOK)GYMPIE Q 4 1020) 92) 837) 295 G73)SLP) IMOK)GYMPIE Q 4 1020) 92) 837) 295 G73)SLP) IMOK)GYMPIE Q 4 1022) 92) 837) 295 G73)SLP) IMOK)GYMPIE q 4 1023) 92) 837) 295 G73)SLP) IMOK)GYMPIE Q 4 1024) 92) 837) 295 G73)SLP) IMOK)GYMPIE Q 4 1025) 92) 837) 295 G73)SLP) IMOK)GYMPIE Q 4 1026) 103) 837) 295 G73)SLP) IMOK)GYMPIE Q 4 1027) 98) 838) 295 G73)SLP) IMOK)GYMPIE Q 4 1028) 98) 838) 295 G73)SLP) IMOK)GYMPIE Q 4 1029) 95) 295 G73)SLP)GYMPIE Q 3 1030) 95) 295 G73)SLP)GYMPIE Q 3 1031) 95) 295 G73)SLP)GYMPIE Q 3 1033) 295 G73)6YMPIE Q 2 1034) 295 G73)GYMPIE Q 2 1035) 100) 296 G73)SLP)N0RTH COAST N 3 1036) 100) 296 G73)SLP)N0RTH COAST N 3 1038) 100) 296 G73)SLP)N0RTH COAST N 3 1039) 100) 296 G73)SLP)N0RTH COAST H 3 1040) 296 G73)N0RTH COAST N 2 1041) 296 G73)N0RTH COAST N 2 1043) 104) 298 G73)SLP)ARGENTINA 3

PAGE 96

1 Table 1.1 continued . Pedi gree Uri gi n Generati on 1045) 104) 298 G73)SLP)ARGENTINA 3 1046) 104) 298 G73)SLP)ARGENTINA 3 1047) 185 G73)0RLAND0 2 1049) 189 G73)0RLAND0 2 1052) 212 G73)MULBERRY 2 1053) 212 G73)MULBERRY 2 1054) 217 G73)PALMDALE 2 1055) 217 G73)PALMDALE 2 1056) 219 G73)CABBAGE VAT 2 1057) 220 G73)CABBAGE VAT 2 1058) 220 G73)CABBAGE VAT 2 1059) 222 G73)CABBAGE VAT 2 1060) 222 G73)CABBAGE VAT 2 1063) 224 G73)CABBAGE VAT 2 1065) 226 G73)TICE 2 1067) 226 G73)TICE 2 1068) 227 G73)TICE 2 1069) 227 G73)TICE 2 1070) 227 G73)TICE 2 1071) 227 G73)TICE 2 1072) 227 G73)TICE 2 1077) 248 G73) ITALY 2 1078) 248 G73) ITALY 2 1079) 250 G73)PINE CREEK N 2

PAGE 97

Tabi e 1.1 conti nued . Pedi gree Uri gi n Generati on 1080) 250 G73 )PINE CREEK N 2 1081) 250 G73 )PINE CREEK N 2 1082) 250 G73 )PINE CREEK N 2 1083) 251 G73 )PINE CREEK N 2 1085) 251 G73 )PINE CREEK N 2 1086) 252 G73 )PINE CREEK N 2 1087) 252 673 )PINE CREEK N 2 1088) 252 G73 )PINE CREEK N 2 1091) 253 G73 )PINE CREEK N 2 1092) 254 673 )PINE CREEK N 2 1093) 256 673 )PINE CREEK N 2 1094) 258 673' IPINE CREEK N 2 1095) 259 673 )PINE CREEK N 2 1096) 260 673; PINE CREEK N 2 1097) 262 673 IPINE CREEK N 2 1098) 267 673] PINE CREEK N 2 1099) 268 673] PINE CREEK N 2 1100) 273 G73] PINE CREEK N 2 1 1 ni ^ X 1 u 1 y CIO P T N F r D p F / r\ t C. N N Pi o C 1102) 277 673) PINE CREEK N 2 1103) 277 G73) PINE CREEK N 2 1104) 278 673) PINE CREEK N 2 1105) 278 673) PINE CREEK N 2 1107) 257 673) N 2

PAGE 98

Tab! e 1.1 conti nu ed . Pedi gr ee ' Ori gi n Generati on 1108) 257 G73)N 2 1109) 263 G73)NEWRY N 2 1110) 264 G73)NEWRY N 2 nil) 265 G73)NEWRY N N 1112) 274 G73)NEWRY N 2 1113) 274 G73)NEWRY N 2 1114) 275 G73)NEWRY N 2 1115) 276 G73)NEWRY N 2 1117) 269 G73)B0AMBEE N 2 1118) 269 G73)B0AMBEE N 2 1119) 271 673)C0NGL0MERATE N 2 1120) 271 G73)C0NGL0MERATE N 2 1122) 272 G73)C0NGL0MERATE N 2 1123) 281 G73)C0NGL0MERATE N 2 1124) 283 G73)C0NGL0MERATE N 2 1125) 283 G73)C0NGL0MERATE N 2 1127) 285 G73)C0NGL0MERATE N 2 1128) 285 G73)C0NGL0MERATE N 2 1129) 285 G73)C0NGL0MERATE N 2 1130) 287 G73)C0NGL0MERATE N 2 1131) 287 G73)C0NGL0MERATE N 2 1132) 288 G73)C0N6L0MERATE N 2 1133) 288 G73)C0NGL0MERATE N 2 1134) 288 G73)C0NGL0MERATE N 2

PAGE 99

Tabi e 1.1 conti nu ed . Pedi gree Or i g i n Generati on 1135) 290 G73)C0NGL0MERATE N 2 1136) 290 673)C0NGL0MERATE N 2 1137) 290 G73)C0NGL0MERATE N 2 1138) 291 673)C0NGL0MERATE N 2 1139) 292 G73)C0NGL0MERATE N 2 1140) 292 G73)C0NGL0MERATE N 2 1141) 292 G73)C0NGL0MERATE N 2 1142) 307 G73)W00NDUM Q 2 1143) 307 G73)W00NDUM Q 2 1144) 309 G73)W00NDUM Q 2 1145) 309 G73)W00NDUM Q 2 1146) 309 G73)W00NDUM Q 2 1147) 309 G73)W00NDUM Q 2 1148) 310 G73)W00NDUM Q 2 1149) 310 G73)W00NDUM Q 2 1150) 311 G73)W00NDUM Q 2 1151) 311 G73)W00NDUM Q 2 1152) 312 G73)W00NDUM Q 2 1153) 312 G73)W00NDUM Q 2 1154) 312 G73)W00NDUM Q 2 1155) 313 G73)GYMPIE Q 2 1156) 313 G73)GYMPIE Q 2 1158) 317 G73)GYMPIE Q 2 1160) 319 G73)KENILW0RTH Q 2

PAGE 100

Tabl e 1.1 conti nued . 89 Pedi gree Uri gi n Generati on 1161 ) 320 G73)BELLTH0RPE Q 2 1162 ) 321 G73)BELLTH0RPE Q 2 1163 ) 324 G73)P0M0NA Q 2 1164 ) 325 G73)P0M0NA Q 2 1165 ) 326 G73)P0M0NA Q 2 1166 ) 326 673)P0M0NA Q 2 1168 ) 327 673)P0M0NA Q 2 1169 ) 328 G73)P0M0NA Q 2 1170 ) 328 G73)P0M0NA Q 2 1171 ) 329 G73)P0M0NA Q 2 1172 ) 330 G73)P0M0NA Q 2 1173 ) 331 G73)P0M0NA Q 2 1174 ) 331 G73)P0M0NA Q 2 1175^ 331 G73)P0M0NA Q 2 1176] 331 G73)P0M0NA Q 2 1177] 331 G73)P0M0NA Q 2 1192] 195 G73)RSA 2 1193) 1013 G73)RSA 2 1194) 1013X1187 G73)RSA 2 1195) 1013X198 G73)RSA 2 1196) 1013X1178 G73)RSA 2 1197) 1178X1186 G73)RSA 2 1198) 1178X1186 G73)RSA 2 1199) 1179X1178 G73)RSA 2

PAGE 101

Tabi e 1 . 1 conti nued . Pedi gree Ori gi n Gen erat i on 1200)1179X1178 G73)RSA 2 1478) 1179X1178 G73)RSA 2 1479) 1179X1178 G73)RSA 2 1480) 1179X1013 G73)RSA 2 1481) 1179X1180 G73)RSA 2 1482) 1179X1180 G73)RSA 2 1483) 1179X1184 G73)RSA 2 1484) 1179X1184 G73)RSA 2 1485) 1179X1187 G73)RSA 2 1486) 1179X198 G73)RSA 2 1487) 1179X1189 G73)RSA 2 1488) 1179X1189 G73)RSA 2 1489) 1179X1191 G73)RSA 2 1490) 1179X1191 G73)RSA 2 1491) 1180 G73)RSA 2 1492) 1180X1013 G73)RSA 2 1493) 1180X1013 G73)RSA 2 1494) 1180X198 G73)RSA 2 1495) 1181 G73)RSA 2 1496) 1182X1179 G73)RSA 2 1497) 1182X1180 G73)RSA 2 1498) 196 G73)RSA 2 1499) 196X1178 G73)RSA 2 1500) 196X1178 G73)RSA 2

PAGE 102

Tabl e 1 . 1 conti nued . Pedi gree Ori gi n Gen erat i on 1501) 196X1178 G73 )RSA 2 1502) 196X1184 G73 )RSA 2 1503) 196X198 G73 )RSA 2 1504) 196X1198 G73 )RSA 2 1505)1183X1178 G73 )RSA 2 1506)1183X1178 G73 )RSA 2 1507)1183X1180 G73 )RSA 2 1508)1183X1180 673 )RSA 2 1509)1183X1184 G73 )RSA 2 1510)1183X1185 G73 )RSA 2 1511)1183X 198 G73 )RSA 2 1512)1183X1189 G73 )RSA 2 1513)1184X1180 G73 )RSA 2 1514)1184X1180 673' )RSA 2 1515)1184X1180 673 )RSA 2 1516)1184X1188 673' RSA 2 1517)1184X1188 673 IRSA 2 1519)1184X1189 673] RSA 2 LOCV) ) lioDAii / o b / J , n c ft R SA 2 1521)1185X1178 673] RSA 2 1522)1185X1180 673] RSA 2 1523)1186X1180 673 ) RSA 2 1525)1187X1188 673) RSA 2 1526)1187X1188 673) RSA 2

PAGE 103

Tabi e 1.1 continued. Pedi gree Ori gi n Generati on 1527)1187X1189 G73)RSA 2 1528)1188X1178 G73)RSA 2 1529)1188X1178 G73)RSA 2 1530)1188X1178 G73)RSA 2 1531)1188X1178 G73)RSA 2 1532)1188X1180 G73)RSA 2 1532)1188X1180 G73)RSA 2 1534)1188X1180 G73)RSA 2 1535)1188X1189 G73)RSA 2 1536) 198 G73)RSA 2 1537) 198 G73)RSA 2 1538) 198X1178 G73)RSA 2 1539) 198X1185 G73)RSA 2 1540) 198X1189 G73)RSA 2 1541) 198X1189 G73)RSA 2 1542)1189X1178 G73)RSA 2 1543)1189X1180 G73)RSA 2 1544)1189X1182 G73)RSA 2 1545)1189X1185 673)RSA 2 1546)1189X1185 G73)RSA 2 1547)1190 G73)RSA 2 1548)1190 G73)RSA 2 1549)1191X1178 G73)RSA 2 1550)1191X1178 G73)RSA 2

PAGE 104

Tabi e 1.1 conti nu ed . Pedi gree ~ Ori gi n Generati on 1551)1191X1178 G73)RSA 2 1552)1191X1178 G73)RSA 2 1553)1191X1185 G73)RSA 2 1554)1191X1186 G73)RSA 2 1555)1191X1189 G73)RSA 2 1556)9999 G73)IDENTITY LOST 2 1558)1557 G73)HYBRID COMPOSITE 2 1560)1557 673)HYBRID COMPOSITE 2 1562)1557 G73)HYBRID COMPOSITE 2 1626 BRASIL 1627 BRASIL 1628 GLADFIELD Q 1629 GLADFIELD Q 1630 GLADFIELD Q 1633 GLADFIELD Q 1634 GLADFIELD Q 1635 GLADFIELD Q 1636 GLADFIELD Q 1637 GLADFIELD Q 1 1638 KENILWORTH Q 1639 CONONDALE Q 1640 CONONDALE Q 1641 CONONDALE Q 1642 CONONDALE Q

PAGE 105

Tabi e 1 . 1 conti nued . Pedi gree Uri gi n Generati on 1643 CONONDALE Q 1 1644 CONONDALE Q 1 1645 CONONDALE Q 1 1646 CONONDALE Q 1 1647 CONONDALE Q 1 1648 CONONDALE Q 1 1649 CONONDALE Q 1 1650 CONONDALE Q 1 1651 . CONONDALE Q 1 1652 CONONDALE Q 1 1653 CONONDALE Q 1 1654 CONONDALE Q 1 1655 CONONDALE Q 1 1656 DANBULLA Q 1 1657 DANBULLA Q 1 1658 CONONDALE Q 1 1659 CONONDALE Q 1 1660 CONONDALE Q 1 1661 CONONDALE Q 1 1662 COOLOOLABIN Q 1 1663 COOLOOLABIN Q 1 1664 MAPLETON Q 1 1665 MAPLETON Q 1 1566 COOLOOLABIN Q 1

PAGE 106

Table 1.1 continued . Pedi gree Ori gi n Generati on 1667 COOLOOLABIN Q 1 1568 COOLOOLABIN Q 1 1669 COOLOOLABIN Q 1 1670 COOLOOLABIN Q 1 1671 COOLOOLABIN Q 1 1672 WEST COOROY Q 1 1673 WEST COOROY Q 1 1674 WEST COOROY Q 1 1675 WEST COOROY Q 1 1676 WEST COOROY Q 1 1677 WEST COOROY Q 1 1678 . WEST COOROY Q 1 1679 WEST COOROY Q 1 1680 WEST COOROY Q 1 1681 CREDITON Q 1 1682 CREDITON Q 1 1683 CREDITON Q 1 1684 CREDITON Q 1 1685 CREDITON Q 1 1586 COMO Q 1 1687 COMO Q 1 1688 COMO Q 1 1689 COMO Q 1 1690 COMO Q 1

PAGE 107

Tab! e 1 . 1 conti nued . Pedi gr ee Qri gi n Gen erat i on 1691 COMO Q 1692 COMO Q 1693 COMO Q 1694 COMO Q 1695 COMO Q 1696 DANBULLA Q 1697 DANBULLA Q 1698 DANBULLA Q 1699 DANBULLA Q 1700 MT MEE Q 1701 MT MEE Q 1702 MT MEE Q 1 1703 MT MEE Q 1704 MT MEE Q 1705 MT MEE Q 1706 MT MEE Q 1707 MT MEE Q 1 1708 MT MEE Q 1709 CONONDALE Q 1710 CONONDALE Q 1 1711 CONONDALE Q 1712 MT GLORIOUS Q 1713 BELLTHORPE Q 1714 BELLTHORPE Q

PAGE 108

Tab! e 1 .1 conti nued . Pedi gr ee Ori gi n Generati on 1715 BELLTHORPE Q 1 1716 KENILWORTH Q 1 1717 KENILWORTH Q 1 1718 KENILWORTH Q 1 1719 KENILWORTH Q 1 1720 HWY 31 CHARLOTTE CO 1 1721) 1736) 294 F69)B IG)COFFS HRBR N 3 1722) 1736) 294 F69)B IG)COFFS HRBR N 3 1723) 1736) 294 F69)B IG)COFFS HRBR N 3 1724) 1736) 294 F69)B IG)COFFS HRBR N 3 1725) 1736) 294 F 69 ) B I G ) C OFF S HRBR N 3 1726) 1736) 294 F69 )B IG )COFF S HRBR N 3 1727) 1736) 294 F69 )B IG )COFFS HRBR N 3 1728 MANATEE CO 1 1729 ORANGE CO 1 1730 POPAYAN COLOMBIA 1 1731 POPAYAN COLOMBIA 1 1732 POPAYAN COLOMBIA 1 1733 CALI COLOMBIA 1 1734 BELLE GLADE 1 1735 HAWAII 1 1737 STYX RIVER N 1 1738 RSA 1 1739 RSA 1

PAGE 109

Table 1.1 continued . Pedi gree Uri gi n Generati on 1740 RSA 1741 RSA 1742 RSA 1743 RSA 1744 RSA 1745 RSA 1746 RSA 1747 RSA 1748 RSA 1749 RSA 1750 RSA 1751 RSA 1752 RSA 1753 RSA 1754 RSA 1755 RSA 1756 RSA 1757 RSA 1758 RSA 1759 RSA 1760 RSA 1761 RSA 1763 LOUWS CREEK TIMBER RSA 1764 CHOMA ZAMBIA

PAGE 110

1 4Tabl e 1.1 continued . Pedi gree _ Ori gi n Generati on 1765 RHODESIA 1 1766 RSA 1 1767 RHODESIA 1 * Q = Queensland RSA = Republic of South Africa N = New South Wal es G73 = GP0P73 SLP = Silver Lake Prairie

PAGE 111

APPENDIX 2. DESCRIPTION OF MEASUREMENTS OBTAINED AT 64-MONTH COPPICE AGE 1. Prior to harvest in February 1978, height measurements were obtained at age 7 months. 2. In November 1978, three months after harvest the height of the tallest stem on each stool was measured. 3. In March 1981, 30 months after harvest the following measurements were taken : a. DBH of the three largest stems per stool b. Height of the three tallest stems on every fifth stool . 4. The occurence of severe freeze in January 1982 permitted study of genetic variation in frost resilience of £. grandi s . In December 1983, ^ 64 months after harvest the following measurements were obtained from all 15,510 stools : a. Height of the tallest stem b . DBH of the tal 1 est stem c. Number of stems per stool d. Frost resilience score 0 = no frost damage 1 = 100% recovery of original height from frost 2 = 50% recovery of original height from frost 3 = height recovery is less than 50% 8 = missing tree 100

PAGE 112

101 9 = dead tree e. Frost defect score 0 = no major defect 1 = undesirable branching, basal sprouting and/or dominant leader killed 2 = freeze scar or lesion 3 = undesirable response: crook, feathering, forking or clawing 8 = mi ssi ng tree 9 = dead tree e. Coppice quality score 0 = fully desirable coppice production 1 = basically desirable but few minor defects 2 = still acceptable but a number of additional undesirable coppicing characteristics 3 = cul 1 (undesi rabl e) 8 = mi ssi ng tree 9 = dead tree g. Coppice defect score 0 = no practical defects 1 = defects only on minor coppice stems 2 = major coppice stem has defects: crook, forking sweep or 1 ean 3 = both major and minor stems defective 8 = mi ssi ng tree 9 = dead tree

PAGE 113

5 APPENDIX 3. ANALYSIS OF VARIANCE TABLES FOR GROWTH, COPPICE AND FROST SCORES Table 3.1 Analysis of variance table for seedling height at 7 months. Source* DF SS EMS G 3 3968.1 V +20.15V^/px+2121.4Vp e F(Gj G F (G) 523 21524.0 Error 8555 142578.2 Total 9081 168070.3 *G is the generation number and F(G) is family within generation. 102 I

PAGE 114

103 Table 3.2 Analysis of variance table for coppice height at 3 months. * Source DF SS EMS G 3 155993.8 V +20 e .15Vp(Q)+2121.4Ve F(G) 523 1293889.7 V +17 e •^^^F(G) Error 8555 8931931.6 V e Total 9081 10381815.1 * G is the generation number and F(G) is family within generati on

PAGE 115

104 Table 3.3. Analysis of variance table for coppice height at 64 months. Source * DF SS EMS G 3 1920784.3 V +20 e .15Vp^gj+2121.4Vg F(G) 523 1636210.4 V +17 e •^^^F(G) Error 8555 9401439.4 V e Total 9081 12958434.1 * 6 is the generation number and F ( G ) is f ami 1 y wi t hi n generati on

PAGE 116

105 Table 3.4 Analysis of variance table for coppice DBH at 64 months . Source * DF SS EMS G 3 2439267.1 V +20. e 15VF(G)^2121,4Vg F(G) 523 2280967.2 V^+17. e Error 8555 13073799.2 V e Total 9081 17794023.4 * G is the generation number and F(G) is family within generati on

PAGE 117

106 Table 3.5 Analysis of variance table for coppice volume at 64 mont hs . * Source DF SS EMS G 3 8584961.3 V +20. e 15Vp(3)+2121.4Ve F(G) 523 10099101.7 V +17. e Error 8555 67761992.4 V e Total 9081 86446055.4 * G is the generation number and F(G) is family within generati on

PAGE 118

107 Table 3.6 Analysis of variance table for frost resilience score at 64 months . Source * DF SS EMS G 3 164.5 V^+20. e 15Vp(s)+2121.4V3 F(G) 523 317.1 V +17. e Error 8555 2183.6 V e Total 9081 2665.2 * G is the generation number and F{G) is family within g ener at i on

PAGE 119

108 Table 3.7 Analysis of variance table for frost defect score at 64 months . Source * DF SS EMS G 3 484.5 V +20. e 15Vp(3j+2121.4Vg F(G) 523 1193.9 V +17. e 21Vp(G) Error 8555 11586.6 V e Total 9081 13264.9 * 6 is the generation number and F(G) is family within generati on

PAGE 120

109 Table 3.8 Analysis of variance table for coppice quality score at 64 months . Source DF SS EMS G 3 F(G) 523 Error 8555 Total 9081 179.9 459.7 4795.1 5434.7 V +20.15Vc,^,+2121.4Vp * G is the generation number and F(G) is family within generati on

PAGE 121

1 no Tab! p 3 9 An;*! \yci c r\'f r\ f 1 a 1 Jr 3 1 o O 1 score at 64 vari ance months . taoie ror coppice aerect Source * DF SS EMS G 3 283.0 V +20 e .15Vp(gj+2121.4Vg F(G) 523 570.1 V +17 e F(G) Err or 8555 5875.9 V e Total 9081 6729.2 * G is the generation number and F(G) is family within generati on

PAGE 122

APPENDIX 4. IDENTIFICATION OF 300 SELECTED TREES USING SELECTION INDEX 3 Table 4.1 Location and family of 300 selected trees from GP0P77 based on selection index (Index 3) equation. Row Col Family Row Col Family 131 163 1498 122 220 1019 131 133 907 130 205 74 123 213 1538 83 224 1478 134 159 101 126 203 1002 116 206 859 116 128 962 116 171 965 82 225 1019 132 190 1012 61 169 1764 116 125 1535 68 149 1164 141 124 1161 116 193 931 140 134 1020 141 148 958 127 205 943 112 158 1019 10 223 867 133 149 1480 49 145 998 140 149 961 136 132 847 55 188 1004 138 183 1055 90 151 1104 60 121 88 114 207 1155 131 141 1499 123 207 847 131 161 1693 66 171 960 32 137 1175 141 185 1057 139 121 1012 67 161 893 129 129 1199 132 131 1560 128 201 1517 131 140 293 111

PAGE 123

Table 4.1 continued. Row Col Fami 1 y Row Col Fami 1 y 136 19 7 1200 66 202 955 74 128 1158 127 174 931 126 132 1081 10 137 985 125 170 988 42 190 88 114 198 907 130 151 1481 49 192 1007 39 145 1544 117 221 849 113 133 1004 128 124 927 70 189 1022 138 225 1149 60 151 1498 140 188 1012 132 147 1124 112 206 869 118 216 1118 66 221 1124 121 216 987 40 174 1138 128 139 846 135 123 1547 128 188 963 119 153 938 125 173 1145 127 140 1683 122 206 861 116 198 1134 86 132 1018 7 129 1764 112 173 908 135 166 873 131 139 906 121 121 1492 132 121 1149 136 220 926 82 184 1018 6 123 1166 132 194 880

PAGE 124

113 Table 4.1 continued. Row Col Fami 1 y Row Col Fami ly 126 182 857 139 127 987 132 129 1054 134 137 1679 84 199 962 133 182 1488 113 206 1489 131 144 1503 15 225 1515 127 223 1523 116 124 1740 53 141 981 123 132 943 125 163 884 129 200 1011 128 120 1507 113 196 1065 127 211 859 131 138 1173 70 132 1002 121 222 1146 102 168 1071 63 134 910 112 222 1020 140 131 918 90 218 1119 58 127 921 132 198 1484 131 198 1057 124 185 1002 129 157 1115 132 132 1484 100 143 1540 122 132 11 ?4 115 224 987 131 206 1192 107 219 914 18 209 1007 140 144 1019 116 219 1040 125 198 858 133 218 1124 134 142 981 80 219 998

PAGE 125

114 Tabi e 4.1 conti nued Row C9I Fami 1 y 71 155 1137 122 129 987 66 142 1702 137 140 1124 126 135 1020 135 128 1038 121 170 90 122 162 1023 131 142 1114 122 123 1485 130 168 931 135 211 90 39 131 1021 124 166 965 138 133 941 131 214 74 137 148 888 114 223 1027 116 177 881 26 225 1113 53 200 1052 133 123 1547 Row Col Fami 1 y 61 184 1154 138 140 850 121 123 1680 71 123 1693 72 145 1171 31 148 1052 135 188 1491 67 218 1022 112 193 586 122 224 854 49 171 917 129 223 1528 124 167 1166 121 209 1556 110 135 1685 137 167 911 96 1 72 1646 135 136 1085 138 184 1496 140 151 1515 46 219 1150 50 125 1020

PAGE 126

115 Table 4.1 continued. Row Col Fami 1 y Row Col Fami 1 y 102 126 993 29 171 950 129 136 1673 72 225 884 123 207 847 137 189 1022 65 170 674 125 221 1544 57 144 1488 96 220 101 130 224 1071 133 134 1174 120 169 938 127 181 849 130 169 996 71 149 1522 125 136 1497 120 157 877 78 224 1693 60 136 674 64 173 1641 125 168 1763 139 142 1492 135 184 895 135 176 1045 141 138 1028 57 153 1140 116 153 986 137 122 1160 102 134 1022 116 119 1501 126 154 991 127 127 1038 128 1 7 Q Q 3 0 O 135 121 1497 111 180 1166 131 118 1544 134 125 954 131 216 1134 121 218 1500 134 186 1545 68 193 1506 76 159 1156 3 131 844

PAGE 127

116 Tab1 e 4. 1 conti nued . Row Col Fami 1 y Row Col Fami 1 y 127 155 876 71 206 674 69 141 888 65 143 1749 120 133 846 115 135 944 29 224 1496 119 124 1003 131 137 917 128 116 1560 66 205 986 53 218 1092 123 138 1765 50 133 1720 19 180 998 70 148 892 30 135 1007 68 166 1055 105 224 1023 139 158 1069 48 204 1710 121 125 1177 136 195 954 68 141 1145 121 122 1708 87 192 1118 124 205 869 99 193 1083 134 220 1137 69 143 1165 134 127 928 116 126 1130 30 173 1533 125 1 50 154 7 68 169 1763 58 143 1514 123 181 1006 10 224 1077 131 134 1142 78 169 1630 71 216 97 129 175 1714 43 125 1132 110 131 1022

PAGE 128

117 Tabi e 4.1 -continued. Row Col Family Row Col Fami 1 y 1 1 O IOC 13 5 1 764 20 185 1549 o n oU 174 1131 122 171 918 120 140 924 67 127 674 ion i d y 134 1538 61 132 1638 OU 1 72 1693 64 168 1160 ion one 206 1764 104 225 896 lib ICC 16 5 1035 132 184 1169 ion 140 1734 133 143 1038 ion 189 1516 58 188 1007 i 0 '^ 121 884 127 153 1043 0 b 205 1479 64 166 1149 O rt o 203 1 764 132 142 1681 1 O 1 131 175 876 72 186 1133 1 o n 129 170 1547 72 131 892 21 157 1063 79 123 1641 67 168 1499 44 163 1200 139 209 994 130 127 958 14 170 987 127 133 950

PAGE 129

LITERATURE CITED Ades, P. K., and I. P. Burgess. 1983. Provenance trials of Eucal yptus g randi s in northern New South Wales. Silvicultura 3: 396-397. Assis, T. F. D., and A. Brune. 1983. Heri tabi 1 i ti es and correlations in progenies of Eucal yptus g randi s from Australia, South Africa and Brazil. Silvicultura 3:524-525. Becker, W. A. 1984. Manual of Quantitative Genetics. Washington State Univ., Pullman, WA. 188p. Borges, R. C. G., and A. Brune. 1983. Heritability estimates and correlations between characters in Eucal yptus grandi s . Silvicultura 3: 525-527. Brim, C. A., H. W. Johnson and C. C. Cockerham. 1959. Multiple selection criterion in soybeans. Agron. J. 51: 42-46. Brown, A. H. D., A. C. Matheson and K. G. Eldridge. 1975. Estimates of the mating system of Eucal yptus obliqua using allozyme polymorphisms. Aust"i J . Bot . 23: 931-949. Burgess, I. P. 1983. The natural occurrence of Eucal ypt us g randi s , its distributions in natural forests, its characteristics and conservation. Silvicultura 3: 397-399. Carr, S. G. M., and D. J. Carr. 1972. Eucalyptus cal ophyl 1 a var. mai d eni a na Hochr. : a male tree. Proc. Royal Soc. Vict. 85: 49-50. Davis, G. L. 1969 . Floral morphology and the development of gametophytes in Eucal yptus st el 1 ul ata Si eb . Aust. J. Bot. 1 7: 177-190. Darrow, K. W., and Roeder, K. R. 1983. Provenance studies of Eucal yptus grandi s (Hi 11) ex. Maiden in South Africa. Silvicultura 3: 402-40^. Dvorak, W. S., E. C. Franklin and G. F. Meskimen. 1981. Breeding strategy for robusta in southern Florida. 118

PAGE 130

Proc. 16th S. For. Tree Improv. Conf., Balcksburg, VA., pp. 116-122. Eldridge, K. G. 1970. Breeding system of E^. regnans . Proc. lUFRO Sect. 22 Meeting, Varparanta, Fi nl and . Vol . 1, 12pp. Eldridge, K. G. 1975, Eucal yptus species. ln_: Seed orchards. R. Faulkner (ed.) 134-139, U. K. For. Comm. Bun. No. 54. H. M. S. 0., London. Eldridge, K. G. 1976. Breeding system, variation and genetic improvement of tropical eucalypts. ^JII " Tropical trees : Variation, breeding and conservation ", J. Burley and B. T. Styles (ed.), pp. 101-108. Academic Press, London. Eldridge, K. G. 1978. Genetic improvement of eucalypts. Silv. Gen. 27: 205-209. Falconer, D. S. 1981. Introduction to Quantitative Genetics. Longman Press, London. 334p. Food and Agricultural Organization of the United Nations. 1979 . Eucalypts for Planting. FAO forestry series No. 11, FAO, Rome, Italy. 676p. Franklin, E. C. 1969. Inbreeding depression in metrical traits of loblolly pine ( Pi nus taeda L.) as a result of self pol 1 i nati on . Tech . Rep. No. 40. N. C. State Univ. Co-op. program, Raleigh, NC . 19pp. Franklin, E. C. 1971. Estimates of frequencies of natural selfing and of inbreeding coefficients in loblolly pine. Silv. Gen. 13: 7276. Franklin, E. C. 1978. Exotics for hardwood timber production in southeastern United States. Proc. 2nd Sym. Southeast, area S and PF , Atlanta, GA. p. 171-179 Fowler, D. P. 1964. Effects of inbreeding in red pine. Pi nus r esi nosa Ait. Silv. Gen. 14: 12-23. Gansel, C. R. 1971. Effects of several levels of inbreeding on growth and oleoresin yield in slash pine. Proc. 11th S. For. Tree Improv. Conf., pp. 173-177. Gaylor, D. W., H. L. Lucas and R. L. Anderson. 1970. Calculations of expected mean squares by the abbreviated Doolittle and square root method. Biometrics 26: 641-655.

PAGE 131

120 Geary, T. F., G. F. Meskimen and E. C. Franklin. 1983. Growing Eucal yptus for industrial wood production. Gen Tech. Rep. SE-23. USDA For. Serv., Southeast. For. Exp. Sta., 43p. Hallauer, A. R. and J. B. Miranda Fo. 1981. Quantitative Genetics in Maize Breeding. Iowa St. Univ. Press. Ames, Iowa . Harris, D. L. 1964. Expected and predicted progress from index selection involving estimation of population parameters. Biometrics 20: 46-72. Hazel, L. N., and J. L. Lush. 1942. The ef f eci ency of three methods of selection. J. Hered. 33: 393-394. Hodgson, L. M. 1974. Breeding of eucalypts in South Africa. S. A. For. J. 89: 13-15. Hodgson, L. M. 1976. Some aspects of flowering and reproductive behavior in Eucal yptus grandi s (Hill) Maiden at J. D. M. Keet For. Res. Sta., S. A. For. J. 97: 18-28. Hodgson, L. M. 1977. Methods of seed orchard management for seed production and ease of reaping in Eucal yptus grandi s . S. A. For. J. 100: 38-42. Hopper, S. D., and G. F. Moran. 1981. Bird pollination and the mating system of Eucalyptus stoatei . Aust. J. Bot. 29: 625-638. Jahromi , S. T. 1983. Variation in Eucal yptus vi mi nal i s with respect to cold resistance and growth. lUFRO Sym. Breeding and Yield of Fast Growing Trees, Brazil. Vol . 3, pp. 502-504. King, J. P. 1983a. Provenance studies of Eucal yptus s a 1 i g n a and Eucalyptus grandi s in Hawaii . Si 1 vi cultura 3: 412-413. King, J. P. 1983b. Selection of Eucalyptus species for northern California. lUFRO Sym. Breeding and Yield of Fast Growing Trees, Brazil. Vol. 3. pp. 453-455. Kraus, J. F., and A. E. Squillace. 1964. Sel fi ng vs. outcrossing under artificial conditions in Pi nus el 1 i otti i Engelm. Si 1 v . Gen. 13: 72-76. Li, C. C. 1976. First Course in Population Genetics. Boxwood Press, Pacific Grove, CA. 594p.

PAGE 132

121 Logan, W. E. M. 1967. Policy. Unaslyva 21: (3/4). pp. 8-22. Meskimen, G. F. 1983. Realized gain from breeding Eucalyptus grandi s i n Florida. Gen Tech. Pap., Pac. Southwest For. and Range Exp. Sta., For. Serv., USDA, PSW-69, pp. 121-128. Moran, G. F., and A. H. D. Brown. 1980. Temporal heterogeneity of outcrossing rates in alpine ash (Eucalyptus del egat ensi s ) . Theor, Appl . Genet. 57: 101-105. Mullin, L. J., and J. Gough. 1983. Variation in survival and height growth in 18-month old progenies of Eucal yptus grandi s i n Zimbabwe. Silvicultura 3: 413-414. Namkoong, G., and E. B. Snyder. 1969. Accurate values for selection intensities. Si 1 v . Gen. 18: 1 72-1 73 . Phillips, M. A., and A. H. D. Brown. 19 77. Mating system and hybridity in Eucal yptus pauci fl ora . Aust . J. Biol. Sci . 30: 337-344. Pryor, L. D. 1957. Selecting and breeding for cold resistance in Eucal yptus . Si 1 v . Gen. 6: 98-109. Pryor, L. D. 1961. Inheritance, selection and breeding in Eucal yptus . Report and Documents, 2nd World Eucalyptus Conf., San Paulo, Brazil, Vol. 1, pp. 297-304. Rockwood, D. L., C. W. Comer, S. V. Kossuth and G. F. Meskimen. 1984. Eucalyptus for biomass production in Florida. Ann. Rep . Oak Ridge Nat. Lab., 18p. Rockwood, D. L., and T. F. Geary. 1982. Genetic variation in biomass productivity and coppicing of intensively grown £. grandi s in southern Florida. Proc. 7th N. Am. For. Biol. Wrksp., Lixington, KY. Rockwood, D. L., and G. F. Meskimen. 1981. Genetic, spacing and genetic X spacing interaction influences on growth of E_. grandi s in southern Florida. Proc. 16th S. For. Tree Improv. Conf., Atlanta, GA. pp. 77-85. SAS Institute Inc. 1985. SAS User's Guide: Statistics, 1985 edition. SAS Institute, Gary, NC . 957p. Smith, H. F. 1936. A discriminant function for plant selection. Ann. Eugen. 7: 240-250.

PAGE 133

122 Snyder, E. B. 1968. Seed yield and nursery performance of sel f -poll i nat ed slash pines. For. Sci . 14: 68-74. Snyder, E. B. 1972. Five year performance of selfpollinated slash pines. For. Sci. 18: (3) p. 246. Sprague, G. F., and L. A. latum. 1942. General vs. specific combining ability in single crosses of corn. J. Amer. Soc. Agron. 34: 923-932. SquiUace, A. E., and R. T. Bingham. 1958. Selective f erti 1 i zati on i n Pi nus monti col a Dougl . Si 1 v . Gen . 7(6): 188-196. SquiUace, A. E., and J. F. Kraus. 1962. Effects on inbreeding on seed yield, germination, rate of germination and seedling growth in slash pine. Proc. of a For. Gen. Wrksp., Macon, GA . pp. 59-63. SquiUace, A. E., and J. F. Kraus. 1963. The degree of natural selfing in slash pine as estimated from albino frequencies. Si 1 v . Gen. 12(2): 46-50. Turner, H. N., and S. S. Y. Young. 1969. Quantitative Genetics in Sheep Breeding. Cornell Univ. Press, Ithaca, NY. van Buijtenen, J. P. 1971. Seed orchard design, theory and practice. Proc. 11th Conf. on S. For. Tree Improv . , pp. 197-206. van Vleck, D. 1981. Notes on the Theory and Application of Selection Principles for the Genetic Improvement of Animals. Dept. Animal Sci., Cornell Univ., Ithaca, NY. 260p. van Wyk, G. 1976. Early growth results in a diallel progeny test of Eucalyptus grandi s (Hill ) ex . Maiden. Si 1 v . Gen. 25: 126-132. van Wyk, G. 1977. Progress with the Eucalyptus grandi s breeding programme in the Republic of South Africa. Proc. 3rd World Consul. For. Tree Breeding, Canberra, Australia, pp. 639-643. van Wyk, G. 1981. Inbreeding effects in Eucal y ptus g randi s in relation to the degree of relatedness. S. A. For. J. No. 116. pp. 60-63. Venkatesh, C. S., R. S. Arya and V. K. Sharma. 1973. Natural selfing in planted Eucalyptus and its estimation. J. Plantation Crops 1:

PAGE 134

Walters, G. A. 1980. Saligna Eucal y pt us growth in a 15-year old spacing study in Hawaii. USDA For. Serv . Res. Pap. PSW-151, 6p. Williams, J. S. 1962. The evaluation of a selection index. Biometrics 18: 375-393. Wright, J. W. 1976. Introduction to Forest Genetics. Academic Press, New York, 463p.

PAGE 135

BIOGRAPHICAL SKETCH Kondala Vikram Reddy, son of K. V. Ranga Reddy, was born Febraury 1, 1959, in Hyderabad, India. He received his early education from Little Flower High School in Hyderabad and graduated in June 1976. He obtained his bachelor's degree in agricultural sciences from Andhra Pradesh Agricultural University, Hyderabad. He came to the United States in August 1980 to attend the University of Illinois, Urbana, Illinois and in August 1982 received a Master of Science degree in forest genetics. He continued his graduate study at the University of Florida in the School of Forest Resources and Conservation in pursuit of the Doctor of Philosophy degree in forest genetics. 124

PAGE 136

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Phi 1 osophy . Associate Professor of Forest Resources and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Phi 1 osophy . Ray E^oddard Pro^j^sor of Forest Resources and Conservati on I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Phi 1 osophy . Susan V. Kossuth Associate Professor of Forest Resources and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Phi 1 osophy . Charles J . Wi 1 cox Professor of Dairy Science

PAGE 137

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Phi 1 osophy . I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Phi 1 osophy . This dissertation was submitted to the Graduate Faculty of the School of Forest Resources and Conservation in the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. Frank G. Marti n Professor of Statistics Randy C. ('ihow Associ at e Pr of essor of Computer and Information Sciences December 1985 Director, Forest Resour/fes and Conservation ^ — '


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ENABE1H7X_OR0MI1 INGEST_TIME 2015-03-09T21:00:04Z PACKAGE AA00028884_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES