the permission of the Author.
A STUDY OF GENETIC AND ENVIRON�ffiNTAL VARIATION AND COVARIATION IN PRODUCTIVE TRAITS OF
A FLOCK OF PERENDALE SHEEP
A thesis presen ted in pa r t ia l ful f i lment of the r equir ements for the degree
o f Doc to r of Philo sophy in Anima l Sc ience a t Ma ssey Univ er s i ty
ROGER PAUL LEWER
1978A B S T R A C T
Performance r ecords on 1 7 2 4 two-year , 1 1 6 0 three -year , 7 2 6 four-year and 4 4 9 f ive-year Perendale sheep from the Mas sey Unive rs ity f lock we re analysed . These records were co l l ected between birth-years 1 9 6 1 and 1 9 7 4 inc l u s ive , and therefore dif ferent numbers of years we re invo lved at each age . The data compr i sed 97 , 8 2 , 65 and 56 sire groups for the r e spect ive age s . There were 4 0 3 ewe s in 56 sire groups which had four consecutive record s be tween two- and five -years of age . There were 9 2 2
two-year , 584 three-year , 2 9 6 four -year and 1 2 5 five -year dam-daughter pairs , but the l a s t age group of reco rds wa s considered too few to j u stify analysis.
The dependent var i ables examined wer e qual i ty number (QN ) , wool character ( WC ) , s taple l ength ( S L) , greasy f l eece weight (GFW) , f ibre diameter (FD) , and number of lambs weaned per ewe joi ned and pre sent at l ambing ( NLW ) . F ibre diameter had been recorded in fewer year s than the other traits and there were conseque ntly l e ss observations ava i l abl e .
The e f fect s of the environmental factors of year , age of the darn of the ewe , rearing rank , NLW (wool traits o nl y ) , and a l l interactions among them were a s s e s sed by computing the perc entage of the total variance contribu ted by each fac tor for each variabl e . Quadratic components for these fixed e f fects were temporarily considered to be variance components for this purpo se . Year effects were the mo st important source o f
enviromental variation i n wool character i s t ics , but we re l e s s important for NLW . None of the other ma in effects were consistently important
( i . e . contributed greater than two percent o f the tota l var iation) in explaining the obs erved variab i l i ty . Year x NLW was the only interaction
i i i
which contri buted more than two percent o f the total vari ance i n any wool tra i t. The a f fected characteris t i cs were two- and three -year QN , thr ee-year FD and f ive-year SL . NLW ( f ive -year-old ewes only) was a f fe c ted by dam age x rear ing rank and year x rearing rank interac tion .
Heritab i l i ty estimates calculated by paternal hal f -s ib method (daughter-dam regression estimates in brackets ) were in the range o f 0 . 2 2 - 0 . 2 8 ( 0 . 3 9 - 0 . 6 6 ) , 0 . 1 6 - 0 . 3 9 ( 0 . 1 2 - 0 . 25 ) , 0 . 3 2 - 0 . 7 1 ( 0 . 25 - 0 . 3 6 ) , 0 . 2 4 - 0 . 53 ( 0 . 4 3 - 0 . 6 2 ) , 0 . 3 8 - 0 . 68 ( 0 . 45 - 0 . 50 ) and 0 . 0 2 - 0 . 2 4 ( - 0 . 0 2 - 0 . 0 3 ) for QN , WC , SL , GFW , FD and NLW respec tively .
D i f ferences between ages were seldom signi ficant and the rel evance o f such a compar ison was discussed . Paternal hal f -sib heri tabi l i ty
estimates for the ave rage of four mature ( two- to five-years ) produc tion records were 0 . 50 , 0 . 6 1 , 0 . 8 4 , 0 . 4 4 , 0 . 4 9 and - 0 . 0 1 for the same
respective characteristics . Binomi a l analys is methods resulted in her i tabi l i ty estimates for fert i l i ty ( twinn ing in brackets ) of 0 . 0 6 ( 0 . 1 2 ) , 0 . 0 2 ( 0 . 15 ) , -0 . 0 1 ( 0 . 1 3 ) , and -0 . 15 ( 0 . 0 1 for two- , thre e - , four - and five -year-old ewes respective l y .
Repeatabi l i ty was est imated by averagi ng the regr ess ion of l ater on earl i er record for a l l pairs of ages for wool trai ts and mature ages for reproductive characterist ics and resu l ted i n values of 0 . 4 6 (QN) , 0 . 18
( WC ) , 0 . 50 ( S L ) , 0 . 6 2 (GFW) , 0 . 7 1 ( FD ) , and 0 . 09 (NLW) . S imilar methods gave computed values for the repeatabi l i ty of the d i f ference betwee n zero and one l amb weaned of 0 . 0 1 , and between one and two l ambs weaned of 0 . 13 .
Geneti c and phenotypic corre l ations between a l l variables wer e calcula ted separately at e a c h a g e by paternal ha l f -sib and daughter-dam methods . The correlations were posi t ive for GFW - FD , GFW - S L , GFW -
WC , QN - WC and F D - S L , negat ive for GFW - QN , QN - FD and QN - S L , and mixed for FD - WC and SL - WC al though the last of these were mos t l y posi tive . The phenotypic correl ations betwe en wool tra i ts and reprod - uction were negative for NLW - GFW and NLW - S L and approxima tely zero for the remainder . Genetic corre l ations were nega tive for NLW - GFW , NLW - FD and NLW - SL , positive for NLW - QN and ma inly posi t ive for NLW - WC. Many o f the genetic correlations involving NLW had large standard errors .
App l ication of the results was discussed in terms of the s e l e c t ion indices which wer e constructed .
ACKNOWLEDGEMENTS
It is with great pl easure tha t
Iacknowledge the many peop l e who have assis ted me to c omplete this thes i s . Foremost a r e my supervisors Professor A . L . Rae and Dr . G . A . Wickham , to whom
Iowe my deepe s t
gra t i tude and apprec ia t ion for the ir cont inuous help and sugges t ions throughout the c ourse of this s tudy .
Var ious of my friends , in part icular Mr . C . J . Dodd and
Mr A . B . Pl easants were of immense assistanc e during the period of computer famil iar ization . They and o ther colleague s and fel low s tudent s have provided a sour c e of s t imula t ing d iscussion on many occasions .
Mr . L . P earson of the Compu t e r Science Depar tment of theUniversity gave invaluab l e help in probl em-solving and the creat ion
of files . helpful .
The Computer Unit staff led by Mr . C . Read were alway s
Thanks are a l so d u e to the Ma ss ey Unive r s i ty farm and t echnical s t aff who collec ted the records and to the Lib rary staff for the i r cont inuous assistanc e .
Mrs . D . S t effe r t dec ipher ed much of my early wr i t ing and typed the
draft of the thesis .All cred i t must go t o Miss J . McKegg fo r skil- fully execut ing the final typing .
My wife Raewyn and daughters Kathryn and Juliet have suffered my frequent depressions and frustrat ions without complaint throughou t this and earlier s tud ies , and provided insp ira t ion fo r their con tinua t ion . Without thei r support and encouragemen t the task may never have b e en comp l e t ed .
The very generous l eave cond i t ions ext ended by the Ministry of Agricul ture and Fisheries are gratefully acknowl edged .
V
Chapt e r
ONE
TWO
THREE
TABLE OF CONTENTS
ABSTRACT
ACKNO\.JLEDGEMENTS LIST OF TABLES
INTRODUCTI ON
REVIEW OF L ITERATURE I .
I I .
I I I .
IV .
V .
Int roduc t ion Her i tability A . Int roduc t ion .
B . Woo l Charac teris t ic s . C . Number o f Lambs Weaned . Gene tic Correl a t ions
A . I n t roduc t ion .
B . Be tween Wool Cha ra c t e r i s t ics.
C . Wool Characteris t ic s and Number of Lamb s Weaned .
Repea t ab i l i ty A . I n t roduct ion .
B . i.Joo l Charac teris t ic s . C . Number o f Lamb s Weaned . Pheno t yp i c Correlat ions A . I n t roduc t ion .
B . Wool Charac teris t ic s .
C . Wool Charac t e r is t ics and N umb er o f Lamb s Weaned .
HATERIALS
I .
TheSheep and
Envi ronmen t A . The Farm .B.
The Perenda l e .
c.
Managemen t . I I .
The Da t a .Page i i
V X
1
3 3 4 4 6 16 19 19 21
2 5 2 9 2 9 3 0 3 9 4 2 4 2 4 2
45
48 48 4 8 4 8 4 8 5 0
Chapter
FOUR
FIVE
HETHODS 56
56 56 59 64 65 65 67 68 69 69 I.
I I.
III.
IV.
V.
Analysis of Variance.
A. Least Squares and C omputing Methods
B . Fitting Models to the Data.
C. Significance Testing.
Accounting for Observed Variance.
A. Variance Component Estimation.
B. Proportion -of-variance Calculations.
Environmental Effects.
Genetic and Phenotypic Parametets.
A. Paternal Half-sib Analysis.
L \.Jool Traits and Number of Lambs Weaned. 69
2. Fertility and Twinning. 72
B .
Daught0r-dam R�gression Analysis.C. �epeatability.
1. Wool Characteristics.
2. Number of Lambs Weaned.
Selection.
77 80 80 81 83
RESULTS AND DISCUSSION. 84
84 84 95 97 99 I.
I I.
Ill.
Analysis of Variance and Environmental Effects.
A. Quality Number.
B .
Wool Character.c. Staple Ll�ngth.
D. Fleece \�eight.
E. Fibre Diameter.
G. Number of Lambs \.Jeaned.
Selection.
Within-age Study.
A. Heritabilities.
1. Quality Number.
2. Wool Character.
3. Staple Length.
103 104
108 110 llO 111 ll4 114
vii
Chapter
IV
4.
Fleece Weight . 5 . Fibre Diame t er .
6 . Number
ofLamb s Weaned .
7.Fer til i ty and Twinn ing .
8.Conclusion s .
B . Gene tic Correl a t ion s .
1.
Be tween Wool Charac teristic s . ( a ) Greasy F l e e ce Weight with o ther Woo l Cha ra c t e r i s t ics .
(b) Qua l i t y Number wi th Other Wool Cha rac ter i s t ic s .
( c ) Fib re Diame ter with Other Woo l Charac t er i s t ic s .
(d) S ta p l e Len g t h wi th Wool Charac ter .
2 . Between Numbe r of Lambs Weaned and Wool Chara c t e ris t ic s .
3. Conclusions ,
C. Phenotypic Correla tions.
1. Between Wool Characterist ics.
( a ) Greasy Fleece We ight with Other Wool Characteris t i cs.
( b ) Qual ity Number with Other Wool Characteristics.
( c ) Fibre Diame t e r with Other Wool Characteristics.
(d) S tapl e Length with Wool Charac ter .
2. Between Number of Lambs Weaned and Wool Charateristics .
3. Conclusions .
Between-age and Lifetime S tudy.
A . Repeatability .
1. Bartlett ' s Test of Homogenei ty o f Varianc e .
Page 115 116 117 119 120
122 122 122 122
125
126
127
128 131 132 132
132
135
135
136
136 138
139 139
139
Chapter
V .
2. Repeatabil ity Est imates .
( a ) Qua l i ty Number . (b) Wool Character . ( c ) S tapl e Leng th . (d) Fl eece Weight . ( e ) Fi bre Diame ter .
( f ) Number of Lambs Weaned .
( g ) Number o f Lambs Weaned at Dif ferent Production Levels .
( h ) Conclusions .
Page 139 142 142 143 143 143 144
144 147
B . Heri tabi l i ty of Annua l Average Performanc e . 148 c. Genetic Correlations Between Ages . 150 App l ic a tions and Conc luding Comments .
A . Some Aspects of Selection . B . Selec tion I ndices .
155 155 158
B IBLIOGRAPHY 167
ix
Tab l e
2 . 1 2 . 2 2 . 3 2 . 4 2 . 5
LIST OF TABLES
E s t ima t es of her itab i l i t y of greasy fl eece we igh t . E s t ima tes of heritability of qual ity number .
Es tima tes of heritab i l i t y of woo l charac ter . Estima tes of heri tab il i t y of s taple l ength.
Estima tes of heri tab il i ty of fibre diame ter . 2 . 6 E s t ima tes of her itability of greasy fleece weight
at different ages .
2 . 7 Estima t es of heritability of s taple length at d ifferent ages .
2 . 8 Est ima t es of her itability of fibre diame ter at d ifferent
age s .2 . 9 Estima t e s of heritability of quality numb er at different ages .
2 . 1 0
2 . 1 1 2 . 1 2 2 . 1 3
Es t ima tes of heritab il i ty o f wool character at different ages .
Es t ima t e s of heri tab ility of number of lambs weaned .
Es t ima t es of gene tic corre l a t ions . E s t ima tes of gene tic correla tions . lamb s weaned with wool t ra i t s .
Wool trai ts . Numb er of
2 . 1 4 Estima tes of repeatability of greasy fleece weigh t .
2 . 1 5 2 . 1 6
2 . 1 7 2 . 1 8 2 . 1 9 2 . 20 2 . 2 1
Estima t e s of repeatab il i t y of woo l character . E s t imat e s of repeatab ility o f q ua l i ty number . Estimat e s of repeatab i l i t y o f s taple l ength . Estimat e s of repeatab ility of fibre diame t e r .
Estimat e s of repeatability of numbe r o f lamb s weaned . E s t ima t e s of pheno typic correlat ions .
E s t ima t e � of pheno typ i c cor relations . l amb s weaned with woo l traits .
Woo l trai t s . Number of
Pag�
8
11
11
12
12
14
14
15
15
15
17
22
26
32 36 36
37
38
40
43
47
Tab l e
3 . 1 3 . 2 3 . 3 3 . 4
3 . 5
4. 15 . 1
5 . 2
5 . 3
5 . 4
5 . 5
5 . 6
5 . 7
5 . 8
5 . 9
Dis t r ibut ion o f Ages wi thin year s . Numb e r o f sire groups analysed . Daugh t er-dam pairs .
Numb e r of ewes with records a t b o th age
1and age 2 .
Qua l i t y numbe r codes .
Main e f f e c t s and nested fac t o r . Calc ula tion o f redu c t ions in su�s o f s quares and expe c ta t ions of mean squares .
Two-year-old ewe s . Lea s t sqt:arcs e s t ima tes and s t andard errors o f envi ronmen tal e f f ec ts .
Three-year-old ewe s . LEa s t squares e s t ima tes and s tandard errors of environmen tal e f f ec ts .
Four-year-o ld ewe s . Lea s t s quares e s t ima tes and s t andard errors of env ironmen tal e f f ec ts .
F ive-year-old ewes . Leas t s quares e s t ima tes and s t andard errors o f environmen tal e f f e c t s .
Two-yea r-old ewes . P e r c entage o f var iance
acc ounted for by each f a c tor for each var iab l e . Three-year-old ewes . Percentage o f varianc e
a c c oun t ed for b y each fac tor for each variab l e . Four-year-old ewe s . P ercen tage o f variance
ac count ed for by each factor for each variab l e . Five-year-old ewes . P e r c en tage o f var iance
a ccount ed for by each f a c to r for each variab l e . Two-year-old ewes . Lea s t squares e s t ima tes o f
environmental e f f ec t s . o n
qua l i tynumber .
Y ear x NLW interac t ion
Page
51 52 52
53
53
62
85
86
87
88
89
90
91
92
96
xi
Tab l e 5 . 1 0
5 . 1 1
5 . 1 2
5 . 1 3
5 . 1 4
5 . 1 5
5 . 1 6
5 . 1 7 5 . 1 8
5. 19 5. 20
5. 21
5. 22
5.23
Five-year-old ewe s . Lea s t squares e s t ima tes o f environmen tal e f f e c t s . Y ear x rearing rank int erac t ion e f f e c t on NLW .
Five-year-o ld ewe s . Lea s t s quar es e s t ima tes o f environmen tal e f f ec t s . Dam-age x rear ing rank in t e rac t ion e f fec t on NLW .
Dif ferenc e in s tandard d evia t ion un i t s be tween the p8rformance of all animals of age i , and the age i
107
107
performanc e o f anima l s which succeed t o age j .
109Heritab i l i t y es t ima t e s , s tandard erro rs and approxima t e
biasses .
112Compara t ive her i t abil i ty e s t ima tes from models
includ ing a�d excluding impor tant in t erac t ions .
121Estima tes o f gene t ic correlations and s tandard errors .
Wool trai t s .
123Es t ima tes o f gene t i c c o rrela t ions a�d s t andard errors . Number o f lamb s weaned with woo l
tra its .Es t ima t es o f pheno t yp i c correla t ions . Es t ima tes o f pheno typic co rrela t ions .
weaned wi th woo l chara c teris t ics .
Wo ol t rai t s .
Numbe r o f lamb sBartl ett's test o f homogene ity of variance Regre s s ion of l ater on earl ier record .
Regre s s ion of l ater on e ar l ier record . NLW at d i f ferent production leve ls.
Heritabi l ity e stimate s, standard errors and approximate biasse s . Average annual per formance
Genetic correlations between ages and between each age and average annual performance .
129
133
137 140 141
14 5
149
152
Tabl e
5.24 Se lection i ndex parameters
5.25 Selec t ion index solutions . Hogget .
5.26 Selection i ndex solutions . Two-year -old ewes .
x i i i
160
162
164
