‘ ' w , ASSESSMENT OF GENETIC vARLATIoN Eff ’_
, , INAMULTI PLANTATION TESTOF
, .   HALF-SIB FAMILIESOFT’SCQTCH-VHNET‘T" :,

      

 

   

 

 

     
 

 

 

 

 

   

       

 

J

   

         

 

.‘«u-.-

   

   

a

 

 

1

   

‘.

 

(

   
   

_ ,, .y 4","-
.F'n' {é}: "If,”

  
 
  

    
    
  
 
 
 
   

     
  
  

  
 

 

~.
”fir’ur‘... rr /- "
I'" {3‘ ..,....,..-. -» ~“'n--"m"‘" .'.'
....-. "- "'_"',,,.,,‘..,,‘ «D u-v:~’
m

    

. ~0-

 

 

 

    
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

WWWTmWWHTBMTNTHT RA RY

4Michigan State
University ti

 

This is to certify that the

thesis entitled

Assessment of Genetic Variation in a
Multi—plantation Test of Half-sib
Families of Scotch Pine

presented by

George Edward Howe

has been accepted towards fulfillment
of the requirements for

Ph.D.

Forestry

degree in

    

Major profe
/,// Jonathan W. Wright

0-7639

 

 

ABSTRACT

ASSESSMENT OF GENETIC VARIATION IN A
MULTI—PLANTATION TEST OF HALF-SIB
FAMILIES OF SCOTCH PINE

BY

George Edward Howe

The objectives of this study were to determine
relative amounts of within- and between-stand genetic
variation, estimate gains from half-sib family selection,
and make recommendations for future genetic improvement

work in Scotch pine (Pinus sylvestris L.). The material

 

used for the within—stand assessment was 140 open-
pollinated half-sib families collected from nine stands
in Norway, Belgium, and East Germany. Between-stand dif-
ferences were assessed using bulked progenies from stands
surrounding those from which the one-parent progenies
were collected.

The 2-0 half-sib families were planted in nine
randomized complete block experiments at each of three
sites in Michigan in spring, 1961. The stand progenies
were planted in the same spring in a provenance test at

each of the same three sites.

 

George Edward Howe

Eight height measurements, age-ll diameter, and
eleven other traits were analyzed in 1969 and 1970 for
each of the nine groups. In addition to considering the
five East German groups of half-sib families individually,
all 100 East German families were analyzed together for
each of four commercially important traits.

Eight traits displayed significant genetic varia-
tion among the Norwegian families, including total height,
diameter, and frequency of Zimmerman moth attack. Ex-
pected gains from half-sib family selection were high for
all of these traits, based on a selection intensity of
50% of the families. Parent-progeny correlations indi-
cated that mass selection in this group would have been
ineffective as an improvement technique. Between-stand
differences were up to 18 times larger than within, and
indicated that continued selection should be concentrated
in the best stands.

There was little genetic variation in either of
the Belgian groups from planted parental stands. One of
these, however, was the fastest growing of all nine groups,
but exhibited no significant within-stand variation in
height, which made it of limited value to the tree breeder.
In the third Belgian group there was usable genetic vari-
ation in total height and diameter. A 50% thinning of
the shortest families will result in a predicted gain in

height of 3.6% in the next generation. Parent-progeny

 

George Edward Howe

correlations were non-significant, showing that mass-
selection would not have been effective in the Belgian
population.

Between-stand differences were non-significant in
East German Scotch pine, so all 100 East German families
were analyzed together, as though sampled from one stand.
This all-East German analysis provided a different picture
of variation than did any one of the five East German
groups by itself. Decisions based on 20-family results
would have led to incorrect thinning of two of the groups,
and Zimmerman moth attack would have been ignored because
of non-significance. The all-East German analysis re-
vealed significant differences in Zimmerman moth attack
and indicated no differences in height growth, a trait
which had shown differences in two of the five East Ger-
man groups individually. It was concluded that small
sample sizes lead to distorted assessment of populational
variation.

Heritability estimates in forestry are strictly
applicable only to the samples for which they are calcu-
lated, because sample sizes have been too small for broad
application.

Genetic gains in forestry have come largely from
sources of variation other than those recoverable by mass
selection. These sources have been primarily provenance

selection and family selection. Mass selection should

 

George Edward Howe

occupy only a small part of improvement programs in
forestry.

The precision of genetic tests may be increased
by improving cultural practices, increasing replication
or otherwise modifying experimental design, and enlarg-
ing sample size. For a given sample size, improved cul-
tural practices are usually less expensive than increasing

replication.

 

 

ASSESSMENT OF GENETIC VARIATION IN A
MULTI-PLANTATION TEST OF HALF-SIB

FAMILIES OF SCOTCH PINE

BY

George Edward Howe

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Forestry

1971

67(75):? &

ACKNOWLEDGMENTS

I express my sincere gratitude to Professor
Jonathan W. Wright (chairman) for providing the experi-
mental material and the many intense hours of guidance
for this study and training as a tree breeder. I am
grateful to the other members of my committee--Professors
M. W. Adams, Charles Cress, J. W. Hanover, and S. N.
Stephenson--for their advice and critical review of this
thesis, and to Dr. Wayne Myers of the Forestry Department
for his help in computer usage. My thanks are extended
to Myra Bair, Frances Howe, Timothy LaFarge, Hildegarde
Lindsey, and David Reicosky for help in various phases
of this study.

Financial assistance for this study was provided
from funds for the NC-Sl Project, and by the Michigan

State University Forestry Department.

ii

TABLE OF CONTENTS

LIST OF TABLES AND FIGURES . . . . .

INTRODUCTION 0 O O O O O O O O O O O 0

MATERIALS AND METHODS . . . . . . . .

RESULTS AND DISCUSSION . . . . . . . .

Appearance of the plantations in 1969

Families #275 to #284 .

Families #285 to #294 . . . . . . .
Families #295 to #304 . . . . . .
Families #531 to #540 . . . . . .
All Belgian Families . . . .
Families #321 to #340 . . . . .
Families #341 to #360 . . . . . . .
Families #361 to #380 . . . .
Families #381 to #400 . . . .
Families #501 to #520

The All-East German Analysis

All East German Families . .
HERITABILITY ESTIMATES IN FORESTRY
THE EFFICACY OF MASS SELECTION .
INCREASING PRECISION OF GENETIC TESTS
LIST OF REFERENCES . . . . . . . . .
APPENDIX . . . . . . . . . . . . . .

VITA O O O O O O O O O 0

iii

Page

11

ll
35
40
43
44
46
47
49
51
52
53
54
56

58
61
65
67
70

83

 

LIST OF TABLES AND FIGURES

Figure Page

1.

Outline map of Michigan showing the locations
of the Russ (R), Kellogg (K), and Dunbar (D)
test plantations . . . . . . . . . . . . . . . . 5

Tables

1.

Identification number and origin information
for 140 half-sib Scotch pine families from
nine stands . . . . . . . . . . . . . . . . . . 4

Traits evaluated in the study of 140 half-sib
families of Scotch pine from nine stands . . . . 9

Means, mean square ratios, coefficients of
variation, variance component ratios and
heritability estimates for height, diameter

and cone bearing in Norwegian families #275

to #284 . . . . . . . . . . . . . . . . . . . . 12

Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for color, needle

retention, and Zimmerman moth attack in

Norwegian families #275 to #284 . . . . . . . . 13

Simple correlations among variable traits of
families #275 to #284 grown from seed collected
at N. H¢land, southern Norway . . . . . . . . . 14

Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for height, branching,

and cone bearing in Belgian families #285 to

#294 . . . . . . . . . . . . . . . . . . . . . . 15

Simple correlations among variable traits of

families #285 to #294 grown from seed collected
at Achel, Limburg, Belgium . . . . . . . . . . . 16

iv

Table

10.

ll.

12.

l3.

14.

15.

16.

Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for height, diameter,
and needle retention in Belgian families
#295 to #304 . . . . . . . . . . . . . . . .

Simple correlations among variable traits of
families #295 to #304 grown from seed
collected at HechteL.Limburg, Belgium . . .

Means, mean square ratios, coefficient of
variation, variance component ratios, and
heritability estimate for needle retention
in Belgiam families #531 to #540 . . . . . .

Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for diameter and
cone-bearing in East German families #321
to #340 . . . . . . . . . . . . . . . . . .

Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for color and needle
retention in East German families #321 to
#340 . . . . . . . . . . . . . . . . . . . .

Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for color, branching,
age-5 height, and age-9 height in East
German families #341 to #360 . . . . . . . .

Simple correlations among variable traits of
families #341 to #360 grown from seed
collected at Neustrelitz, East Germany . . .

Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for cone bearing and
foliage color in East German families #361
to #380 . . . . . . . . . . . . . . . . . .

Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for needle retention
and age-5 height in East German families
#361 to #380 . . . . . . . . . . . . . . . .

Page

17

18

19

20

21

22

23

24

25

 

Table Page

17. Simple correlations among variable traits of
families #361 to #380 grown from seed collected
at Gfistrow, East Germany . . . . . . . . . . . . 26

18. Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for height growth,
needle retention, and age-10 height in East
German families #381 to #400 . . . . . . . . . . 27

19. Simple correlations among variable traits of
families #381 to #400 grown from seed collected
at Nedlitz, East Germany . . . . . . . . . . . . 28

20. Means, mean square ratios, coefficient of
variation, variance component ratios, and
heritability estimate for height growth in
East German families #501 to #520 . . . . . . . 29

21. Simple correlations among traits of families
#501 to #520 grown from seed collected at
Joachimsthal, East Germany . . . . . . . . . . . 30

22. Variation in height, diameter and cone-bearing
of Scandinavian provenances #543 to #546 . . . . 31

23. Simple correlation among traits of Scandina-
vian provenances #543 to #546 . . . . . . . . . 32

24. Mean, mean square ratio, coefficient of vari-
ation, variance component ratios, and
heritability estimate for Zimmerman moth
attack in East German families #321 to #400
and #501 to #520, grown at Russ Forest . . . . . 33

25. Variation in height, diameter and cone-bearing
in families #275 to #284 grown from seed
collected in N. H¢land, southern Norway . . . . 70

26. Variation in color, needle retention and
Zimmerman pine moth attack in families #275
to #284 grown from seed collected in N.
H¢1and, southern Norway . . . . . . . . . . . . 71

27. Variation in height, branchiness and cone-

bearing in families #285 to #294 grown from
seed collected at Achel, Limburg, Belgium . . . 72

vi

Table Page

28. Variation in height, diameter, and needle
retention in families #295 to #304 grown from
seed collected at Hechtel, Limburg, Belgium . 73

29. Variation in needle retention in families #531
to #540 grown from seed collected at Campine,
Belgium . . . . . . . . . . . . . . . . . . . 74

30. Variation in diameter and cone-bearing in
families #321 to #340 grown from seed col-
lected in Rovershagen, East Germany . . . . . 75

31. Variation in color and needle retention in
families #321 to #340 grown from seed col-
lected at Rovershagen, East Germany . . . . . 76

32. Variation in color, branchiness, height at
age-5, and height at age 9 in families #341
to #360 grown from seed collected at
Neustrelitz, East Germany . . . . . . . . . . 77

33. Variation in cone-bearing and color in
families #361 to #380 grown from seed col-
lected at Gfistrow, East Germany . . . . . . . 78

34. Variation in needle retention and age-5
height in families #361 to #380 grown from
seed collected at Gfistrow, East Germany . . . 79

35. Variation in height growth, needle retention
and age-10 height in families #381 to #400
grown from seed collected at Nedlitz, East
Germany . . . . . . . . . . . . . . . . . . . 80

36. Variation in height growth in families #501
to #520 grown from seed collected at
Joachimsthal, East Germany . . . . . . . . . . 81

37. Variation in Zimmerman moth attack among East

German families #321 to #400 and #501 to
#520, grown at Russ Forest . . . . . . . . . 82

vii

 

INTRODUCTION

Scotch pine (Pinus sylvestris L.) is the most im-

 

portant planted Christmas tree species in the United
States, and is widely planted as an ornamental. It holds
potential as a pulp and lumber species in North America,
as it does in its native Europe. For these reasons,
Scotch pine has become one of our most widely-planted
exotic tree species, and is receiving increasing attention
from tree breeders in the northeastern U.S. and south—
eastern Canada.

In Scotch pine, as in any organism, genetic im-
provement is dependent upon the amount of genetic variation
present in the population. Considerable variation between
races of Scotch pine has been demonstrated by Langlet
(1937), Wright and Bull (1963), Nanson (1968), and Wright
g£_al. (1966). What accounts for race formation? What is
the nature and extent of within-stand genetic variation
and how does it relate to between-stand variation? These
_questions must be dealt with in assessing family selection
as an improvement technique, which is the focus of the
present study.

Specifically, the objectives of this study were

(1) to determine the relative amounts of within- and

 

between-stand genetic variation in variable traits, (2)
to make estimates of genetic gain by half-sib family
selection, and (3) to make recommendations for a program
for the future genetic improvement of Scotch pine. The
term half-sib families is used throughout, although it is
reCOgnized that some open-pollinated family members may

be full-sibs.

MATERIALS AND METHODS

The material used in this study is the same as in
the study reported on by Wright (1963). One open-
pollinated seedlot (family) from each of ten or twenty
randomly-chosen trees in each of eight European stands of
Scotch pine (Table l) was collected in the fall of 1958.
A ninth group, in Norway (Table l), was represented by
nine half-sib families from one stand and a six-tree
bulked sample (#284) from a nearby stand. The latter was
included by mistake. All parental stands were 8 to 10
acres in area and fully stocked.

All seedlots were sown in the Michigan State
University (MSU) forest tree nursery in the spring of
1959. In the spring of 1961 the nine groups of families
were planted in nine randomized complete block experiments
at the MSU Fred Russ Forest, Cass County, southwestern
Michigan (Figure 1). Each experiment contained five
blocks (replicates), and each block contained one 4-tree
r0w plot of each family.

The nine experiments were also planted in the
same spring at the MSU W. K. Kellogg Forest, Kalamazoo
County, south central Michigan (Figure l), and the MSU

Dunbar Forest, Chippewa County, eastern Upper Peninsula

 

.pcmum Umpcmam UHOIHMdMTooa mnummc Eomm Goflumumcommu HMHSDMZm

N
.Ucmum annmmc m Eonm mHQEMm omMHSQ couple m mm3 www¢a

.czocxcs monopm UmucmHm mo cflmfluo

 

wcmfiuww .m

 

 

m>flumz .mvoma .mmomm .HmcumEHAOMOb Momunom .o ommlaom
mamfiumw .m

mcmbcmam .mHoNH .m omm .Nuflacmz xowucom .o oovnamm
acmsumw .m

m>flumz .Haoma .maomm .3ouummo nomenom .o ommuamm
hcmfiumw .m

m>Humz .m omH .mmomm .mpflamnpmsmz xomucom .o ommuavm
mamEnmw .m

m>flumz .maoma .Haoam .qmmmnmum>wm xomunom .o ovmuamm
Esflmamm

Ncmucmam IIIIIIIIII .mchEMU xsoaow .« owmnamm
Esflmamm

m .Hmom .s .Hm .musnsaq mcaflfinsmn we .4 somumam
Edflmamm

NompcmHm .om°m . .haoam .mnsnfiflq occHHbEMh mp .¢ vmmlmmm
mmzuoz

m>Humz .am0HH .omoom .eemanm .z emesm .9 Hammumsm

cflmfluo .m.mcoq .z.umq c0flumooq Eoum Hogans

©m>flmomu comm NHHEmm

ccmum ucmnmm

 

.mccmum mcflc Eoum moflHHEmm mafia couoom
nflmumamc ova How c0flumfiuomcw cwmfino 0cm Hogans coflpmoHMHpcmcH .H magma

 

Outline map of Michigan showing the locations
of the Russ (R), Kellogg (K), and Dunbar (D)
test plantations.

Figure l.

 

 

Figure l

 

of Michigan, except that families #285 through #294 were
not planted at Dunbar. The experiments were replicated
ten times at Kellogg and five times at Dunbar, except
families #295 through #304 which were replicated ten times
at Dunbar.

Additional material included in this study was
bulked stand progenies collected from stands near those
from which the half-sib seedlots originated. These stand
progenies were part of a range-wide provenance test also
started in 1959 (Wright and Bull, 1963). Three of the
provenance outplantings, which were planted in 1961, were
located on the same sites as the half-sib progeny experi-
ments just described. Data from this provenance material
were used to assess between-stand genetic variation.

Fourteen traits plus mortality were measured on
the half-sib families and appropriate provenance trees
in the three Michigan plantations in the summer of 1969
and winter of 1969-70 (Table 2). Measurements were made
in units approximating l/20th of the range in a trait
among individuals. In all analyses, plot means were
used as the basic observation.

An AOV was run for each trait for each site sepa-
rately. This included traits scored in one plantation

only (Table l):

 

Source DF MS F EMS
Total (T) BF-l --— --_ ___

. . M81 _
Families (F) F 1 M81 fi§§ VE+bVFs
Blocks (B) B-l --— —_- _-_
Error (B-l)(F-l) M82 --- VE

Each trait which was scored in two or more plantations

was subjected to analysis of variance of the following

 

form:
Source DF MS F EMS
Total (T) (FEBi)-l --- ---
Families (F) F-l M81 T§l. V +bV +bsV
MS2 E FS F
Sites (S) 8—1 ——- ___
Blocks/sites 2(Bi-l) --— ___
M82
F .X S (F‘l) (S'l) M82 ITS—3 VE+bVFS
Error (F-l)[Z(Bi-l)] M83 VE

where Bi = number of blocks in the ith site

and b = harmonic mean of number of blocks.

The single-site analyses were performed so that
comparisons could be made between single-site and multi-
site results.

Provenance data were analyzed in the same way as
were the half—sib families.

For 2- and 3-side analyses the additional ratios

Family M8

Of Error MS’

 

and, for all analyses, the coefficient of

 

Table 2. Traits evaluated in the study of 140 half-sib
families of Scotch pine from nine stands.

Description

 

Total height at age 2 in the nursery.

Total height at age 5 or 6.

Total height at age 9 or 10.

Total height at age 11.

Five-year height growth, 1965 through 1969.
Diameter at middle of fourth inter node below apex.
Number of branches in the 5th shorl below apex.
Number of trees bearing cones.

Number of cones per tree.

Number of trees attacked by Zimmerman pine moth
(Dioryctria zimmermani).

Number of trees forked.

Number of trees damaged by the pine grosbeak (lateral
buds plucked off).

Number of trees attacked by white pine weevil (Pissodes
strobi).

Number of trees with tops broken out.
Mean branch angle of all branches in 5th whorl below apex.

Foliage color in midwinter on scale from 0 (yellowest) to
6 (bluest).

Needle retention, in years.

 

 

10

W?

= ————-X 100), were calculated.
Mean

variation, in percent (C.V.
Simple correlations between traits were calculated. Dif-
ferences between families within stands were contrasted
with differences between stands, using provenance data
for the latter.

AOV's of age-ll height, height growth the last 5
years, diameter, and Zimmerman moth attack for all 100
East German families together were run, on the hypothesis
that they were representatives of the same population of
Scotch pine and that the sampling of 100 parents from one
stand would have been comparable to the sampling of the
100 parents from five stands. The same statistics were
calculated for these analyses as for the previous ones.

Heritabilities of half-sib family means for traits

measured on several sites were estimated from variance

components using the following formula:

2 VF .

VE/bs + VFS/b + VF

 

Heritabilities for traits measured at one site only were

estimated by the formula:

V

 

h2 = F
VE/b + VF
VF VFS VE
The variance component ratios ——7 ———, and —— were
V V V
T T T
computed, where V = V + V + V

T F FS E'

RESULTS AND DISCUSSION

One or more of the nine groups displayed genetic
variation in all of the traits measured except number of
trees forked, number of trees attacked by the pine gros-
beak, number of trees attacked by the white pine weevil,
and mean branch angle (Table 2). Plantation means, mean
square ratios, coefficients of variation, variance com-
ponent ratios, heritability estimates, and simple correla—
tion coefficients for traits displaying significant
differences are presented in Tables 3 through 24. Family
means for those traits are in Tables 25 through 37,
Appendix. Other traits which were highly correlated with
those whose values are tabulated, are discussed only in

the»text.

Appearance of the Plantations in 1969

 

The nine experiments at Russ Forest were located
on two separated fields. Some of the East German families
were located on the northernmost field, and they formed a
closed stand in which mortality was so low that I found it
difficult to keep located by keying on empty planting
spots. There was very little variation in height in the

stand and no other obvious gradients.

ll

 

 

 

12

 

.mmpHm mo Hogans N m “mxooHQ

m mm m

 

 

 

 

 

 

 

 

> + n\ > + mn\ >
mo Hogans mo cmmE 0HcoEch N n can . m> N N: +
.Ho>oH wH map um “GMOHMHcmHm N *4
mm vm mm +vam£
B>
>.mn o.m> m.mh va WW
>
B
v.o m.o m.o va mm»
>
B>
m.mH v.mH w.mm Rwy MI
>
m.mHH w.HH v.HH va .>.O
. . . m2 mm
mm H mv H ow H m2 muHm x m
. .. . m2 Hm
NaHm v semo m *«mm 0 m2 8mm
m em m m.H m.m m.m m.h m.m w.>
monocH poem
(momma mo w Hmpmfich new: ucmHmc com:
.cho .HHmM mmsm .cho .HHmM mmsm .cho .HHmM mmsm mDHm
mmcoo HH mom HH mom
mcHHmon momma pm HmumEMHo um pcmHom

#l
I

.vw~* ou mnm¢ mmHHHEcm cmHmm3Hoz cH mcHHmwn mcoo can
.HmqucHU .ucmHma Mom mmumEHumm huHHHnmuHHmc can moHpmu #cmcomfioo
OOGMHHM> .coHDMHHm> mo mpcoHonmmoo .mOHpmu mnmovm some .mcmmz .m oanB

l3

 

.mmuHm mo Hmnfisc n m “mxooHQ
m> + n\mm> + mn\m>

 

 

 

 

 

 

mo nomads mo cmoE UHGOEHmc u n can m N mg +
.hHo>HDommmmH .me>oH mH can wm one we DGMOHmwcme N *« can 4
mm L mm om +vamc
B>
v.mm H.mm m.vn Amv WI
>
E
III m.» m.o Amy mm»
>
B>
m.wh m.mm m.Hm Amy MI
>
v.a> m.mH H.mm Rwy .>.U
. . m2 Hm
Ame H *Ams m w: wuflm x m
. . . m2 Mm
*Amm mH NNHH OH 44mm 5 mmlmmm
5H n.H m.H v.H m.m m.m m.m
mwmup mo w mumow mcmnw
mmsm .ocdo .HHoM mmsm .nczo .HHmM mmsm ouHm
xomupm SHOE coHpawpou HOHoo mmMHHom
cmEHoEEHN chooz moucHz
.vmm# ou

mnm¢ mmHHHEmw cMHmwsnoz cH xompum EDOE cmEHmEEHN cam .cOHucmuwn
ochmc .HoHoo How mwumEHumw MDHHHQmuHHmn cam .mOHumu ucmcomfioo
mocMHHm> .GOHDMHHM> mo mucmHOHmmmoo .moHumu onmcvm cmmfi .mcmoz .v oHnme

 

14

 

.hHo>Hpowmmmu .mHm>mH NH can wm ecu um ucmoHMHcmHm N as can *

 

m N Eocwmmm mo mowumwa

 

3 I 0 HH mam
*¥NN. C .III ¥¥Hm O ¥¥Nw O Q.“ .Hwnvmgwflg
. . m can
Ill scan 0 *amm o #6 #nmflwm
. . m mom
«mm o Namh o #m uflmflmm
. munch m pmmH
Nana o Bosonm uanmm
m can m can munch m pmMH HH mum
um unmflmm um nemflmm epzonm peeflmm be uneHom

 

.hmzuoz cum3u50m .chHsm .2 pm wouowHHoo comm Soum czoum vwm¢
op mnm¢ mmHHHEmm mo mDHMHD mHanHm> mcoam mooHDMHoHHoo onEHm .m mHQMB

 

15

Table 6. Means, mean square ratios, coefficients of vari-
ation, variance component ratios, and heritability
estimates for height, branching, and cone bearing
in Belgian families #285 to #294.

 

 

 

 

 

 

Age-6 Branches per 4 Trees with
height whorls cones
Site Russ Russ Kell. Russ Kell.
Feet Number % of trees
4.7 32 28 77 52

M ** 'k **
Er MS 4.32 2.12 5.29
F x Site MS ___
W 0.28 1.45
C.V. (%) 9.2 27.6 53.2
VF

V_ (%) 39.8 12.7 20.4

T
V
VF—S (s) ——- 10.0 4.7
T
VE
—- (%) 60.3 97.2 74.7
V

T
h2 (%)+ 77 7o 77
* and ** = significant at the 5% and 1% levels,
respectively.
+ VF

h2 = and b = harmonic mean

 

VE/bs + VFS/b + VF

of number of blocks; s = number of sites.

 

 

16

Table 7. Simple correlations among variable traits of
families #285 to #294 grown from seed collected
at Achel, Limburg, Belgium.

 

 

 

Height at Height growth
age 11 last 5 years
Height growth 0 92**
last 5 years '
Height at 0.71* 0.51
age 6
Diameter at 0.64* 0.70*

age 11

Degrees of freedom = 8

 

* and ** = significant at the 5% and 1% levels,
respectively.

 

l7

 

.mwu.._..m WHO .HTQEHHC. H m uwMOOHQ

 

 

 

 

 

_m mm m
mo nonfisc mo cme oHcoEHmc N Q can > + Q\ m> + mb\ > N Na +
.>H0>Hpommmon .mHm>mH NH can wm ecu pm DGMOHmwcmHm N 4* can 4
mm mm mm mm +va N:
.H.>
«.mn m.mv v.¢m N.om Amv m1
>
H.
-L- m.H m.4H- 4.H- Awe mm»
>
B>
m.mm N.mm m.m v.m va MI
>
m.0H m.NH o.m m.wH Hwy .>.U
. . . m2 Hm
ILL mm H mm 0 mm 0 m2 m x m
H
ms.~ Atmm.s .ms.m Nam.m mm awe
v.m m.H h.H m.H m.H o.m o.m m mH MH
Doom mnmom mwcocH Doom
mwsm .Qcco .HHmM mmom .cho .HHoM mmsm .bcco .HHoM mmom mpHm
ucmHon COHucmpmH HH mom HH mom
mlmvfi chomz um HmmemHQ um pcmHom

 

.VOM# OH mmmw meHHEcm cmHmHmm GH GOHDcwpoH mHUmwc can
.HwDoEMHU,.p£@HwQ How momeHumm MDHHHQMDHMmQ can .mOHumg pcocomfioo
moccflnm> LGOHDMHHM> mo mucmHOHmmooo .mOHpcn mumsqm cmoE .mcmmz .m mHQMB

 

18

Table 9. Simple correlations among variable traits of
families #295 to #304 grown from seed collected
at Hechtel,Limburg, Belgium.

 

 

Height at Height growth
age 11 last 5 years
Height growth 0 89**
last 5 years ’
Height at 0 61 0 40
age 6 ' '
Diameter at 0 84** O 93**

age 11

Degrees of freedom = 8.

 

** = significant at the 1% level.

 

 

19

Table 10. Means, mean square ratios, coefficient of
variation, variance component ratios, and
heritability estimate for needle retention
in Belgian families #531 to #540.

 

Needle
retention

 

Site Russ Kell. Dunb.

 

Years
1.9 1.8 2.0

 

 

Fam MS **
—EE—MS 2.84
Fam x Site MS
_——EY_M§—-___ 0.46
C.V. (%) 25.1
v
F
(%) 8 6
vF + VFS + vE
FS (8;) 0 0

———--li—————— (%) 97.1

 

— (%) 74
VE/bs + VFS/b + VF

 

** = significant at the 1% level.

 

20

 

 

 

 

 

Table 11. Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for diameter and cone—
bearing in East German families #321 to #340.

Diameter at Trees
age 11 bearing cones
Site Russ Kell. Dunb. Russ Kell. Dunb.
Inches % of trees
2.9 . 1.3 24 28
Fam MS ** **
Hfifjmg 3.96 2.70
F x Site MS *
m?— 1072 1.46
C.V. (%) 10.6 122.8
V
v—F (%) 10.0 5.9
T

V

V—‘Tls- (%) 9.6 6.7
T

VE

—— (%) 80.4 87.2

V

T
h2 (%)+ 62 50
* and ** = significant at the 5% and 1% levels,
respectively.
V
T h2 F and b = harmonic mean

of number 0

 

VE/bs + VFS/b + V
f blocks;

F
number of sites.

S

 

 

 

21

Table 12. Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for color and needle
retention in East German families #321 to #340.

 

 

 

Winter Needle
foliage color retention
site Russ Kell. Dunb. Russ Kell. Dunb.
Grade Years

4.7 4.2 3.2 1.8 1.7 2.0

 

Fam MS

 

Er MS 3.60** 5.76**
F x Site MS *1.

___—Er MS 1.95 2.18**
c.v. (%) 19.6 12.8

V

VF— (91:) 7.3 14 2

T

v

Vis- (%) 12.7 14.1

T

VE

V‘ (%) 79.8 71.6

T

h2 (%)+ 53 69

 

** = significant at the 1% level.

T 2

h and b = harmonic mean of

 

VE/bs + st/b + VF
number of blocks; 5 = number of sites.

 

 

 

 

 

 

 

 

.mouHm mo menace N m
m mm m
“mxooHn mo gonads mo acme oHcoEMmc N Q can > + Q\ m> + mn\ > N Nc+
>
.H0>0H AH men 06 ueNOHMquHm n .4
mm en Hm mm +va me
8>
m.n> H.mm H.Hm H.Hm Rwy m1
>
E
III III w.m N.m va WNW
>
B>
H.mm o.mm v.m o.NH va MI
>
2 H.mH h.m H.Nm m.HN va .>.o
2
. . m2 Hm
III III *¥Oh H wh 0 m2 Ovflw x m
. . . . m2 Hm
*va N ssmH v «Nah m *shH m meAEMm
¢.m w.m mm om om m.m N.w m.v
Doom Doom umnfisz ocmnw
.cho mmcm .bcso .HHoM mmcm .cho .HHoM mmsm muHm
ucmHmc ucwch mHHOSB v MOHoo mmeHow
alomd mumm< “om monocmum Hmchz

 

.owM#

ou va¢ moHHHEmw cmEHmw ummm cH pcmHoc mlomm can .ucmHmc mummm
.mcHnocmun .HOHoo How mwumEHumm muHHHccanoc can .onDMH ucwcomfioo

wocmHMm> .GOHHMHHM> mo mbcwHonuooo .onumu mumowm some .mcmmz

.MH mHQMB

 

23

Table 14. Simple correlations among variable traits of
families #341 to #360 grown from seed collected
at Neustrelitz, East Germany.

 

Age-11 Height growth
height last 5 years
Height growth 0 55*
last 5 years °
Age-5 *
height 0.39 0.46
Age-9 **
height 0.06 0.63

Degrees of freedom = 18.

 

* and ** = significant at the 5% and 1% levels,
respectively.

 

24

Table 15. Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for cone bearing and
foliage color in East German families #361 to

 

 

 

 

 

 

 

#380.
Trees bearing Winter
cones foliage color
Site Russ Kell. Dunb. Russ Kell. 'Dunb.
% of trees Grade
16 30 0 4.3 4.3 3.4
M ** .1...
Br MS 3.14 5.09
F x Site MS *
Er MS 1.51 1.14
C.V. (%) 134.1 20.2
V
V3 (%) 7.6 17.5
T
V
V1515. (%) 7.2 1 9
T
VE
-—— (%) 85.1 80.5
V
T
h2 (%)+ 57 79

 

* and ** = significant at the 5% and 1% levels,
respectively.
V

T 2 _ F _ .
h — VE/bs + VFS/b + VF and b — harmonic mean of

number of blocks; 5 = number of sites.

 

25

 

 

 

 

 

 

 

 

 

Table 16. Means, mean square ratios, coefficients of
variation, variance component ratios, and
heritability estimates for needle retention
and age-5 height in East German families #361
to #380.

Needle Height at
retention age 5
Site Russ Kell. Dunb. Russ
Years V Feet,
1.8 1.7 1.9 2.2

EEE_H§ ** **
Er MS 5.13 3.43

F x Site MS ** ___
Er MS 2.03

C.V. (%) 12.4 11.4

VF

—— (%) 12.8 35.8

V

T
V
VF-é (%) 12.7 ———
T
VE
V‘ (%) 74.3 64.4
T
h2 (%)+ 57 74
** = significant at the 1% level.
T 2 VF

and b = harmonic mean of

 

vE/bs + VFS/b + vF

number of blocks; 5 = number of sites.

26

.mHe>HDoemmeH .mHe>eH wH use

wm ecu um DceoHMHcmHm

H Ne Use «

 

wo.o HH.o mm.o
v0.0 HH.o *mv.o
III samh.o

emm.o

.wH u

hm.o

mv.o

mv.o

«Nmm.o

«amm.o

Eoceeum mo meenmem

mecoo.oz

.mecoo
QDHB meene

nemame
OHnem<

ngmHeg
muemm

whee» m umeH
cuzonm ucmHem

 

cuHB meeua

 

 

pamHmn pamamg
OHueme mnemm

#I (ll
11' ‘I‘TI

whee» m umeH ucmHec
cuBonm pcmHem HHIemd

.mcefiuew pmmm .Bonbmww um ceuoeHHoo Ueem 80Hm c3oum

omm¢ on Hmm# meHHHEem mo mDHmuu eHneHHe> mcoEm mQOHDeHeHHoo eHmEHm .SH eHneB

 

27

Table 18. Means, mean square-ratios, coefficients of
variation, variance component ratios, and
heritability estimates for height growth,
needle retention, and age-10 height in East
German families #381 to #400.

 

 

 

 

 

 

 

 

Height growth Needle Height at
last 5 years. retention age 10
Site Russ Kell. Dunb. Russ Kell. Dunb. Dunbar
Feet Years Feet
10.4 10.8 6.0 1.9 1.7 2.0 7.0
EEE_M§ ** ** *
Er MS 2.75 4.08 1.87
F x Site MS * ___
Er MS 1.39 0.89
C.V. (%) 10.5 15.7 107.8
VF
V—'(%) 6.6 15.2 15.0
T
V
V§§ (%) 5.7 0.0 ---
T
VE
V—'(%) 87.8 86 2 85 3
T
1'12 (%)+ 53 77 48

 

* and ** = significant at the 5% and 1% levels,
respectively.
+h2 = VF and b = harmonic mean of
VE/bs + st/b + VF
number of blocks; 5 = number of sites.

 

 

 

28

Table 19. Simple correlations among variable traits of
families #381 to #400 grown from seed collected
at Nedlitz, East Germany.

 

L

 

 

 

 

 

Age-11 Height growth
height last 5 years
Height growth 0 84**
last 5 years '
Age-6 ** **
height 0.80 0.66
Age-10 **
height 0.41 0.89

Degrees of freedom = 18.

 

** = significant at the 1% level.

29

Table 20. Means, mean square ratios, coefficient of
variation, variance component ratios, and
heritability estimate for height growth in
East German families #501 to #520.

 

Height growth
last 5 yearS‘

 

Site Russ Kell. Dunb.

 

 

 

Feet
10.7 9.8 7.2

Fam MS **
Er MS 3'58
Fam x Site MS *
4. Er MS 1.50
C.V. (%) 11.1
V
V: (%) 9 2

T
V
VE— ($3) 7 1

T

V

v.4 (%) 83.8

T

h2 (%)+ 61

 

* and ** = significant at the 5% and 1% levels,
respectively.
V

T 2 F °
h = . and b = harmonic mean
VE/bs + VFS/b + VF

of number of blocks; s = number of sites.

 

 

Table 21.

30

Simple correlations among traits of families
#501 to #520 grown from seed collected at
Joachimsthal, East Germany.

 

 

Age-11 Height growth
height last 5 years
Height growth 0 79**
last 5 years °
Age-6 ** *
height 0.73 0.55
Age-10 **
height 0.29 0.59

Degrees of freedom = 18.

 

* and ** = significant at the 5% and 1%
levels, respectively.

 

 

 

 

 

31

 

 

 

 

 

 

 

 

 

Table 22. Variation in height, diameter and cone-bearing
of Scandinavian provenances #543 to #546.
Height at Diameter at Conestper
age 11 age 11 tree
Site Site Site
Prov. Russ Kell. Russ Kell. Russ Kell.
% of mean % of mean % of mean
543 111 108 109 113 167 105
544 117 124 114 118 233 133
545 99 95 98 94 0 77
546 72 73 78 75 0 84
Feet Inches Number
All 7 2 8.6 1.7 2.6 12 57
ngvfigs 18.8** 21.6** 1.3
P gfsMgs 0.54 1.75 0.02
C.V. (%) 21.4 18.2 164.5

 

significant at the 1% level.

 

 

32

Table 23. Simple correlation among traits of Scandinavian
provenances #543 to #546.

 

 

 

Age-ll Height growth I Age-6
height last 5 years height
Height growth 0 99**
last 5 years °
Age-6
height 0.82 0.76
Age'll 0.86 0.86 0 72

diameter

Degrees of freedom = 2.

 

** = significant at the 1% level.

 

 

33

Table 24. Mean, mean square ratio, coefficient of varia-
tion, variance component ratios, and heritabil-
ity estimate for Zimmerman moth attack in East
German families #321 to #400 and #501 to #520,
grown at Russ Forest.

 

 

 

 

% of
trees attacked
23
Fam M8 *
Er MS 1'31
C.V. (%) 99.3
V
v3 (%) 5.93
T
V
VE (%) 96.3
T
V
2 F
h = - (%) 2.40
VE/b + VF

 

* = significant at the 5% level.

 

 

 

34

By contrast, trees in the Norwegian families were
far apart and averaged not much more than a man's height.
They had shorter internodes and needles and yellower
foliage, winter and summer, than the adjacent East German
families. Mortality was low in this stand, too.

The remaining East German and Belgian families were
planted in a large plantation near the south boundary of
Russ Forest. Weed control apparently had not been as ef-
fective on this field as on the northern one, because there
were frequent openings which were choked with multiflora
rose, and the ground here was covered with wild black-
berries. The Belgian trees were so large that the over-
lapping of the branches frequently made it impossible for
me to push through between trees.

At Kellogg Forest all of the half—sib families
except for four replicates of the Norwegian group were
planted together on the tOp of a gentle hill. The first
growing season after planting was a particularly dry one,
and the trees highest on the hill suffered high mortality,
which resulted in large openings.

None of the families formed closed stands at Dun-
bar Forest, although the Belgian families were distinctly
the tallest here, as at Russ and Kellogg. The experiments
were planted on three separate fields, two of them adja—
cent to one another near the northwest corner of the

forest. One edge of these plantings and all of the

 

35

southeastern field were very wet, even in late summer.
Flooding had caused high-mortality in-these areas, and

the variation in heights of surviving trees was striking.
The factors causing the high mortality here.and at Kellogg
Forest also resulted in high error variances among the

surviving trees in the affected areas.

Families #275 to #284

 

The Norwegian families displayed significant dif-
ferences in age-ll height, diameter, cone-bearing, winter
foliage color, needle retention, and Zimmerman moth attack
(Tables 3 and 4). There were significant differences in
height growth the last 5 years, age-5 height, and age—9
height. All height measurements were highly correlated
with one another and with diameter (Table 5). Number of
cones per tree showed significant differences and was
highly correlated with number of fruiting trees (Table 5).

Seedlot #284, which was the 6-tree bulked sample
from a stand near the one providing the other nine Norwe-
gian seedlots, in no case contributed enough to sums of
squares to cause an otherwise low F value to become signi—
ficant. It also did not change variance components enough
to greatly distort heritability estimates. It did, how—
ever, provide the opportunity to compare the magnitude of
error variance of bulked progenies with that of single—

tree progenies, at least indirectly. Error variance

 

 

36

within a plantation was in part a measure of blocks by
seedlots interaction, and so was a function of the range
in plot means within each seedlot. This range for age-11
height and diameter for seedlot #284 never exceeded the
maximum range for any of the other nine seedlots. This
was true in each of the three plantations, indicating that
the error variances for seedlot #284 did not exceed the
error variances for the half—sib progenies. This sug—
gests that for non-genetic experiments, e.g., fertilizer
trials, the experimenter need not be concerned about hav-
ing more error variance if he uses stand progenies than if
he uses single-tree progenies.

Families #279, #282, #280, and #283 were the tall-
est at age 11; three of these also ranked in the top five
in diameter, and three of them were the best families in
number of fruiting trees. The height ranking of the
families did not change appreciably with age. For example,
family #279, which was tallest at age 2 (26% taller than
average) was also tallest at age 11 (12% taller than
average). Family #280 was also among the tallest at age
2 and 11, and #275 was the shortest at both ages. In
other words, the nursery performance was a good predictor
of age-11 height.

Winter foliage color differences were of little
practical significance, because the group as a whole was

quite yellow (Table 4 versus Table 13) and not satisfactory

 

 

 

37

for Christmas trees. The same was true of needle reten-
tion; growers will take interest in this trait only if
three or more years' needles are held.

Zimmerman pine moth attack, present at Russ Forest
only, was easily recognizeable by the presence of foamy-
looking pitch masses full of frass at the union of branch
and stem. The larvae bore into a tree at this point and
mix saliva with the oleoresin to harden it rapidly. This
gives the pitch masses a slightly foamy appearance, and
the attacks are quickly identifiable.

A severely attacked tree might have had attacks
at all branches at a node, with many branches broken off,
or even the stem broken off at the damaged node. For this
reason, I expected a high positive correlation between
number of trees attacked and number of trees with tops
broken out. However, the correlation was very low (non-
significant) and negative. Apparently there were other,
unrecognized, factors accounting for many of the broken
tops.

A 50% thinning, removing the families most at-
tacked by Zimmerman moth (Table 26, Appendix), will leave
trees averaging 82% fewer attacks than the overall aver-
age. Applying a heritability of .93 (Table 4) I estimate
that one generation of such selection could reduce the

incidence of attack by 76%, from the present 17% to 4%.

 

 

38

Such selection for resistance to Zimmerman moth will not
change growth rate appreciably.

However, the Norwegian families are of little com-
mercial interest, anyway, because of their slow growth and
undesirable color, but in the present study they were the
only ones-displaying substantial genetic variability in
Zimmerman moth attack. They should be thinned on the
basis of this trait, even at the sacrifice of high selec-
tion differentials in other traits, so that this apparent
resistance may be available for future breeders.

Ruby (1964) found no significant correlations
among families #275 to #284 and their parents, indicating
that the best-looking parents in the stand in Norway did
not produce the best offspring. This suggests that mass
selection would be largely ineffective as an improvement
regime in this Norwegian population. Because of the strong
age-Z/age-ll correlations in growth rate, it seems that
this ineffectiveness is still true.

Mass selection produced mixed results in Nilsson's
study (1968). The 15 tallest of 30 parent trees providing
material for Trial no. 20 were 5.4% above average. The
offspring of these 15 trees were 1.5% above average at
age 18, so the approximate heritability (comparable to
parent-progeny regression) was l.5/5.4 = .28. Moreover,

the parent-progeny correlation was high; that is, the

 

 

 

 

39

tallest parents produced the tallest offspring and the
shortest parents the shortest offspring.

By contrast, the parent-progeny correlation in
the Lunnaby trial (Nilsson, 1968) was low, even though
the heritability calculated as above was 1.00. Offspring
from superior mother trees were substantially taller in
four of eight groups tested. In the other four, offspring
from the taller parents were shorter than the average.
prgroups were tested individually, there would be as
many showing no gain from mass selection as gain.

The mass selection practiced in the acquisition
of the material used in Ehrenberg's study (1966) did not
produce consistently superior (or inferior) offspring.
Plus- and minus-tree performances were in some cases highly
correlated with offspring performance, and in other cases
were poorly or even negatively correlated with offspring
performances. Mass selection has not been consistently
effective in improving south Scandinavian Scotch pine.

The within-stand differences present in families
#275 to #284 were contrasted with between-stand differ-
ences, using data from provenance stand progenies #543 to
#546 (Table 22). These were collected from Norwegian and
Swedish stands surrounding the stand from which the half-
sib families came (Wright and Bull, 1963). In this
provenance material there were significant between-stand

differences in age-ll height, age-6 height, height growth

 

 

 

40

the last 5 years, age-11 diameter, and number of cones per
tree (Zimmerman moth attack was not scored in the proveé
nance material). F values were from 3 to 18 times larger
between stands than within the parental stand of families
#275 to #284, indicating that between-stand differences
were greater than within. The best of these four prove-
nance stands for all four traits was #544 and the poorest
was-#546. The average height for the four stand progenies
was 7.9 feet (Table 22), as compared to 8.3 feet for the
Norwegian half-sib progenies grown at Russ and Kellogg
(Table 3). This suggests that the Norwegian stand pro-
viding families #275 to #284 was slightly above average
but not among the best stands in the area.

In practical terms, future improvement work in the
south Scandinavian Scotch pine should concentrate selec-
tion in the best stands in the area. This population was
represented by collections from 17 stands in Sweden and
Norway in the provenance test (Wright and Bull, 1963).

The best four or five of these should be used as the
sources for selections in future improvement work in the

population.

Families #285 to #294

 

In the Belgian families #285 to #294, which were
tested at Russ and Kellogg Forests only, the number of

branches and number of trees bearing cones showed

 

 

41

significant differences (Table 6). The number of cones

per tree also showed significant differences-and was highly
correlated with number of trees with cones (r = .90). All
other correlations among traits were low. In none of the
traits was the interaction between seedlot and plantation
significant, indicating that the performance was consistent
in the two plantations.

Family #290 contained 60% more fruiting trees than
the average for other families from this stand. Families
#286, #288, #289, and #292 all were in the range from 12
to 22% above the mean for all families from the stand.

The ratio of family variance to total variance was 20.4%
(Table 6). This means that about 20% of all of the vari—
ance was accounted for by the differences between families.
These differences at age 11 were probably more a measure
of earliness of fruiting than of differences which will
remain throughout the life of the trees.

Family #293 was 17% below average in number of
branches; families #285, #288, and #291 were in the range
from 1 to 10% below the mean. Families #287 and #290
were 20% above the mean. The establishment of a seed
orchard using family #290 because of its fruiting ability
would mean the inclusion of one of the poorest families
in branchiness. However, this relationship for this
family was not part of~a trend; that is, there was not a

high positive correlation between number of fruiting trees

 

 

 

42

and branchiness. Selections based on low number of
branches would not include more low- than high-fruiting
families.

These Belgian families showed significant differ-
ences in winter foliage color at age 1 and age 3 (Wright,
1963). The differences were not discernible at age 11.
In the nursery study the seedlots were grown very close
to one another, which permitted the observer to distin-
guish between color differences which could not be picked
up by an observer walking between plots in a large test
plantation. It is also probable that some of the early
color differences had disappeared by age 11.

Wright (1963) found no other significant differ—
ences among traits which were scored in both studies. He
did find differences in date of bud formation, presence
of primary and secondary leaves, and leaf length. These
were traits which were very difficult to measure precisely
in large trees, and were considered to be of less import-
ance than any of the other traits which are listed in
Table 2.

There were significant differences in total height
at age 6 at Russ Forest (Table 6). The differences were
still significant at age 11 at Russ, but when the data
were included with the Kellogg data, the differences dis-

appeared. This was one of many demonstrations in this

 

 

 

43

study that the results at one site were not consistent

with multi-site results.

Families #295 to #304

Belgian families #295 to #304 showed significant
differences in age-ll height, age-6 height, age-ll dia—
meter, and needle retention (Table 8). Height growth in
the last 5 years also showed significant differences and
was highly correlated with age-ll height (r = .89, Table
9); age-6 height was not highly correlated with either of
the other height measurements (Table 9). This trait was
scored at one plantation only and again shows the incon-
sistency of single- versus multi-site results. Perfor—
mance was consistent throughout all plantations for all
of these traits, as shown by the non-significance of the
seedlot by plantation interaction terms (Table 8).

Family #301 was the tallest at age 11 and was 11%
above the mean; #297, #298, #300, and #302 Were 4% to 5%
above average. Family #298 was 5% above the mean in dia-
meter; #297 and #302 were 3% above the mean. A 50% thin-
ning of the shortest families at age 11 will result in a
selection differential of 4.8% for total height, 2% for
diameter, and 12.5% for needle retention.

The half-sib family heritability of age-ll height
was .76 (Table 8), so the expected gain in height in the

next generation among open-pollinated families coming

 

 

44

from the thinned plantations is 3.6% of the present mean,
or .47 foot.

The families from this Belgian stand were the
second fastest-growing group in the study (statistical
comparisons were not made). The high heritability of
height among these families offers ample opportunity for
improvement in an already outstanding origin, which is of
great practical importance.

Wright (1963) found significant differences in
height among the families at ages 1 and 2. The six tall-
est families at age 2 were still the tOp six families at
age 11, which indicated that nursery growth was a good
predictor of later height growth.

Ruby (1964) found no significant correlations for
any traits between these families and their parents grow—
ing in Belgium. This means that had the best-looking
parent trees been consciously selected they would not
have produced the best offspring. Mass selection in this

stand would have been a waste of time.

Families #531 to #540

 

Belgian families #531 to #540 were the tallest
group at age 11 (15.0 feet, average) of those included in
the study, but displayed practically no within-stand
genetic variability. This absence of genetic variation

is regrettable because of the outstanding commercial

 

 

 

45

potential of the group. It offers no opportunity for
genetic improvement in any trait except needle retention
(Table 10), which is of little practical importance until
needles are retained three years or more. It is doubtful
that Scotch pine anywhere achieves this. Wright (1963),
too, found very little genetic variation among these
families.

Height differences were significant at ages 1 and
2 but not at ages 6, 10, or 11. The ratios of family
variance to total variance were in all cases quite small
(this was true even for needle retention, Table 10), and
error variance to total variance quite large. But in
Wright's data for the Belgian families these ratios were
34% and 66%, respectively for age—l height. This indi-
cates that error variances have increased substantially
since the nursery study, reducing the sensitivity of the
tests. It is not evident what factors are operating in
this particular group of families to increase error vari—
ance.

It is also possible that these families came from
parents (planted trees of unknown origin) which were gene-
tically similar for all traits except early height growth,
which in some populations may be controlled by genes other
than the ones controlling later height growth. If the
parental seed had been collected in such a population, the

offspring might display this kind of variation pattern.

 

 

 

 

46

All Belgian Families

 

The two Belgian stands displaying the least genetic
variability (#285 to #294 and #531 to #540) were the two
planted stands. The seed for the parent stands might have
come from commercial seed dealers who probably collected
fromatwide area and from diverse sources. The offspring
from such a plantation would be expected to display in-
creased genetic variability. The results suggest the op-
posite, that is, the parent plantations originated from a
narrowed genetic base. Nanson (1969) demonstrated that
offspring from plantations originating from a few trees
displayed little genetic variability. This reconfirms
the necessity for establishing seed-orchards from a broad
genetic base, if effective continued selection is to be
practiced.

There is little opportunity for improvement in re—
sistance to Zimmerman pine moth attack among the Belgian
families. The number of Belgian trees attacked by the
pest ranged from 13 to 25% in the Russ plantation, so it
poses a serious threat to the growing of Belgian Scotch
pine in southern Michigan and in many other areas where
it is planted in the U.S. Either resistance will have to
come from other sources (e.g., the Norwegian families) or
it will be necessary to look to chemical or biological

pest control.

 

 

 

47

Families #321 to #340

 

East German families #321 to #340 were signifi-
cantly different in age-ll diameter, number of fruiting
trees, winter foliage color, and needle retention (Tables
11 and 12). For all four traits the seedlot by plantation
interaction was significant; that is, the best families at
Russ were not the best at Kellogg or Dunbar and the worst
at each plantation were not the worst at the others.
Interaction means that individuals or groups of related
individuals perform differently with respect to one an—
other in different environments. Interactions make it
difficult to recommend thinning which will result in a
satisfactorily high selection differential at all sites.
Often, interaction is a measure of the failure of a
majority of the individuals or families to perform con-
sistently, yet several perform well at all sites. For
example, family #323 ranked in the top nine at all three
sites in age-ll diameter; families #326, #327, #330, and
#338 also ranked in the top nine (Table 31, Appendix). A
selection intensity of 75%; saving these five families,
will result in a selection differential of 7%. Using a
heritability of .62 (Table 11), the expected gain in the
next generation is 4.3% of the present mean, or .12 inch

in diameter.

 

48

However, present and anticipated crown closure
necessitates a thinning of 50% rather than 75%, so fami-
lies #321, #328, #329, #335, and #336 will be saved also.
This reduces the selection differential to 4% and the
gain to 2.5%

The alternative would be to thin each plantation
separately. But since I wish to recommend families for
planting over a broad area of the U.S., it will be neces—
sary to either (1) produce material which is superior at
all sites where Scotch pine may be planted, or (2) iden-
tify the environmental factors accounting for the inter-
action at the three test sites, grow material suited to
those factors, identify the factors at potential planting
sites, and plant the appropriate material. Those environ—
mental factors are not yet identifiable, so approach (1)
is the only feasible one. This means the same families
will be thinned at all three test plantations and low
selection differentials will have to be accepted.

Wright (1963) found significant height differences
at ages 1, 2, and 3, and significant differences in winter
foliage color in all three years in the nursery. The
environmental uniformity in the nursery made it possible
to distinguish small differences in heights. This is a
trait which is sensitive to environmental influences and
for this reason small height differences in the test

plantations could not be detected. It is also possible

 

 

 

 

49

that height differences disappeared with age, as a result
of genetic influences, as discussed in a preceeding sec—

tion.

Families #341 to #360

 

Height differences were significant at ages 1, 2,
and 3 in the nursery (Wright, 1963) and still significant
at age 5 at Russ Forest and at age 9 at Dunbar Forest in
the East German families #341 to #360 (Table 13). However,
the differences had disappeared by age 11 at each of these
plantations, and were not present in the combined analysis.
The test at Dunbar had become much more insensitive be-
cause flooding of the planting site had caused high vari—
ability and mortality among these families, as reflected
by an unusually high error variance. At Russ Forest the
age-ll family mean heights ranged from 11.8 to 15.1 feet,
which means that differences of approximately 9% were not
detectable, even in the most precise of the three tests.

There was no correlation between nursery height
performance and later height performance at either Russ
or Dunbar plantations. Likewise, there was no correlation
between height ranking at Russ at age 5 and Dunbar at age
9 (Table 32, Appendix); the best agreement was in family
#348, which ranked among the shortest three in the nursery,

at Russ (age 5), and at Dunbar (age 9). Also family #352

 

 

50

was among the tallest five both in the nursery and at age
5 at Russ Forest (not included in the Dunbar data).

These data suggest that planting or other environ—
mental influences caused differences in height among these
families which were unrelated to their genetic potential
in this trait. These influences were not outgrown by age
9, as shown by the Dunbar analyses and by the non-
significance of height-growth differences in the last 5
years in the combined analysis. This stand should not be
thinned on the basis of height data presntly available,
but should be thinned on the basis of height growth after
age 9, if significant differences appear.

These East German families offer opportunity for
improvement in branchiness and in winter foliage color
(Table 13), which was greener in this group than in any
of the others included in this study. When thinning on
the basis of height growth after age 9, the selection dif—
ferential which is produced for winter foliage color should
also be considered. Color is of little importance to pulp
and lumber growers, but is of great interest to the Christ—
mas tree industry. The East German origins are faster-
growing than Spanish origins, but not as blue—green

(Wright and Bull, 1963).

 

 

 

 

51

Families #361 to #380

 

In these East German families there were signifi-
cant height differences at age 5 at Russ Forest (Table 15),
and these differences were significantly correlated with
nursery performance (r = .52, p > .05 r'= .444); the height
differences at ages 1 and 2 were significant (Wright, 1963).
However, the low correlation did not permit accurate pre-
diction, as the tallest and shortest families in the nur-
sery were not the tallest and shortest, respectively, at
age 5 at Russ Forest. Height differences were still
significant at age 11 at Russ, but the correlation be-
tween age-S and age-ll, although significant, was not high
(r = .69, Table 17), again suggesting that environmental
influences accounted for much of the height difference.

There was little consistency in fruiting between
Kellogg and Russ, the only two plantations which had high
fruiting. This was measured by the significant seedlot
by plantation interaction term (Table 16). Of all the
groups showing significant differences in the number of
trees with cones in this study, these families were the
only ones showing significant interaction. This means
that selection based on 3-site average family performance
may not produce high selection differentials at all three

sites.

 

52

None of the traits showing significant differences
is important enough commercially to be used as a selection
criterion. It is recommended that, as in families #341 to
#360, height growth measurements after age 9 be acquired
and, if significant differences appear, selection be based

on this character.

Families #381 to #400

 

These East German families showed significant dif-
ferences in age-10 height at Dunbar Forest (F = 1.87,
Table 18); the significance of the differences had de-
creased by age ll (F = 1.61, significant at the 10% level
only). Furthermore, the correlation between age-10 and
age-ll heights was only r = .41 (Table 19). However,
these families grew an average of 1.9 feet in that one
year, and this was 27% of their average age-10 height.
Obviously, differences among the families in height in-
crement in that one year could drastically change prior
relationships. These increment differences might be re-
lated to genetic potential, but might also be conditioned
by environmental influences present in a particular year.

There was low correlation (r = .34) between height
in the nursery and height growth the last 5 years for the
combined data (Table 35, Appendix), and between nursery
heights and age-10 heights at Dunbar (Table 35, Appendix).

The significant interaction term for height growth

 

 

53

supports the contention that the performance in this
trait was unstable in changing environments.

These families came from the only planted stand
among the five East German groups. There was no dis—
tinctly different pattern of genetic variation in this
group than in any of the other four. Apparently the
parents did not originate from either widely diverse

sources or from a very narrow genetic base.

Families #501 to #520

 

There was very little genetic variability among
these East German families; the only trait showing signi-
ficant differences was height growth in the last 5 years
(Table 20). Because of high seedlot by plantation inter-
action, selection for height growth on the basis of
average height over all test sites would not produce a
high selection differential at all three plantations. How-
ever, selection on this basis is not recommended, because
this trait, as in other East German stands, was not highly
correlated with nursery performance (r = .39) or with any
other height measurements, except height at age 11 (r = .79,
Table 31). The latter did not show significant differences,
so the correlation is probably chance.

There is no apparent explanation for this lack of
variability in this one of the five East German groups,

except that by chance trees were picked which grouped

 

 

 

54

near the mean. A 10- or 20-tree sample of a forest stand
is probably not likely to provide a good estimate of the
genetic variability of that stand. It was for this reason
that the 100 East German families were analyzed together,

as though sampled out of one stand.

The All-East German Analysis

 

LaFarge (1971) tested between—stand differences in
age-11 height, height growth the last 5 years, and age-ll

diameter among the five East German parental stands repre—

 

sented in the present study. He treated the Russ, Kellogg,
and Dunbar plantations as three replications of stands and
found no differences between stands in any of these three
traits. Likewise, I found no significant differences in
any of nine traits measured in stand progenies collected
from East German Scotch pine stands for the provenance
test (Wright and Bull, 1963). I concluded that the East
German population of Scotch pine was fairly uniform and
that the 100 families from five stands might be considered
as though sampled from one stand.

Including all East German families in one analysis
showed significant differences in Zimmerman moth attack
(Table 24). The 51 least-attacked families averaged 13.3%
attack, which represented a selection differential of 42.5%.
However, the heritability estimate for this trait in this

group was very low (h2 = .024, Table 24), so in the next

 

 

55

generation I expect 22.78% instead of 23.00% of the trees
to be attacked. The low predicted gain results from the
very low heritability estimate. It demonstrates that the
practical limits for producing high selection differen-
tials inevitably cause commercially meaningless gains when
heritabilities are low.

Age-ll height, height growth, and age-11 diameter
all showed no significant differences in the all-East
German analysis. Wright (1963) found significant height
differences among these families in a similar analysis
performed on the nursery material at ages 1 and 2. This
supports the contention that height is a trait sensitive
to the relatively larger environmental influences present
in test plantations than present in a nursery; that is,
the influences of the three different test sites created
non-genetic differences, but they cancelled each other
out when the data were subjected to combined analysis.

The decision not to thin any of the East German
groups on the basis of present height data was supported
by the results of the all-East German analyses; the im—
portant commercial traits so far measured (except Zimmer-
man moth attack) displayed no genetic variation. The
absence of genetic variation in East German Scotch pine
was further evidenced by the lack of significant between-

stand differences found by LaFarge (1971) and myself.

 

 

 

 

56

The East German Scotch pine is in the variety

hercynica (Ruby, 1964; Wright and Bull, 1963), in which

 

substantial genetic variation has been demonstrated (Wright
and Bull, 1963; Wright, et_al., 1966). The apparent ab-
sence of substantial genetic variation within and between
stands in East Germany suggeststhat varietal boundaries

may have been too broadly drawn in the earlier studies.

All East German Families

 

Genetic variation was minimal among the five East
German groups in the commercially important traits height
and diameter. This was confirmed by the all-East German
analyses. The putative desirability of thinning families
#381 to #400 and #501 to #520 on the basis of height
growth in the last 5 years was not supported by the all-
East German analysis, which showed those differences to
be non-significant for all groups.

Zimmerman moth attack differences became signifi—
cant in the all-East German analysis, whereas they were
not in any one of the groups individually. This makes
East German Scotch pine of more commercial interest to
the tree breeder.

The all-East German versus the individual groups
analyses suggest that 20-tree samples may not permit pre-
cise estimates of within-stand genetic variation, and may

lead to incorrect selection recommendations. In the

 

 

 

57

present study, I would have been tempted, on the basis of
single-group analyses, to recommend thinning families
#381 to #400 and #501 to #520 on the basis of height
growth, and to ignore Zimmerman moth attack differences
altogether.

All analyses of these groups showed that nursery
performances were not good predictors of later height
growth. This trait in the East German material appeared
to be much more sensitive to environmental influences than

the same character in either the Belgian or Norwegian popu-

 

lations. By contrast, the winter foliage color differences
were less sensitive, for they remained consistent from age

1 and 2 to age 11.

HERITABILITY ESTIMATES IN FORESTRY

Sewell Wright (1921) demonstrated that the measur-
able genetic variance among half-sibs was only 1/4 of the
additive genetic variance present in the parents. Thus,
heritability estimates applicable to the present genera-
tion should correctly contain 4VF in the numerator. But
the next generation of offspring, because they will be
half-sibs, will contain only 1/4 of the genetic variance
present in the present generation, so it will be necessary
to divide by 4 again. For convenience, the 4 in the
numerator was omitted in the formulae on page 10. To
make the heritabilities applicable to selection based on
family rather than plot means it was necessary to divide
the error variance by the number of sites and number of
replicates per site, and to divide the family by site
interaction by number of replicates.

Heritability estimates vary because, in the ab-
scence of estimates verified by realized gain in forestry,
tree breeders are not in agreement about the components
to be included in heritability estimates. Namkoong,
Snyder, and Stonecypher (1966) published a summary of all
forest tree heritability estimates they considered to be

reliable. The report demonstrated that family

58

 

 

 

 

59

heritabilities had been calculated in various ways, result-
ing in a range of estimates. None was verified by realized
gain, so heritability estimation cannot be regarded as an
absolute.

Heritability estimates for a particular trait in
a particular species vary from study to study, too, be-
cause most improvement programs in forestry to date involve
too few samples of a population. In the present study, no

one of the five East German groups provided the same esti-

mate of the genetic variability in East German Scotch pine

 

as did the five groups combined. Examples of variation in
heritability estimates appear in studies of the inheritance

of specific gravity in slash and loblolly pine:

 

 

 

h2
Species Age Nugger hgig- fgii- clonal Author
parents
STE? >14 8 ~73 Efiiiia‘fiéez)
3:13:11 >14 6 56 23:???962)
:12? 7 gig-Tats...
:12? v 13 SSTSaT‘Egé‘ET
3.2251” 2 23:88:38
3.22:9” 3 agree,
3.21251” . assassin“
1

Broad sense heritability.

 

 

 

60

The only estimates which were consistent were the

two involving lOO-tree samples, and much of that consis-
tency was because the two reports were on the same material.
With small sample sizes, heritability estimates must be
only within those boundaries.

regarded as parameters of the individual tests, and useful

to be parameters of larger populations.

They cannot be considered

 

. ;

 

 

 

 

THE EFFICACY OF MASS SELECTION

The parent-progeny correlations reported for the
Norwegian and Belgian populations included in this study
showed that mass selection would not have been effective
for growth traits in those parental stands. Comparable
work in Scandinavia and Belgium was cited which also
showed that mass selection in Scotch pine was not consis-
tently effective. Obviously, mass selection would not
have been effective in the East German material included
in the present study, as there was no detectable genetic
variation present in growth traits.

In a study of black wattle, Moffett and Nixon
(1963) found significant differences in only one of four
traits between offspring from parents selected for (1)

high values and (2) low values in each of the traits:

Diameter . . . . . . . . . not significant

Bark thickness . . . . . . not significant

Percent tannin . . . . . . significant at the 1%
level

Percent gummosis . . . . . not significant.

The authors concluded, "For diameter and bark
thickness, phenotypic selection has had a negligible or

only slight effect, . . ." (p. 5). Webb and Barber (1965)

61

 

62

found non-significant correlations between selected slash

pine mother trees and their open-pollinated offspring in
test plantations in height, diameter and volume at age 8.
Barber (1964) provided data from a progeny test in slash
pine which permitted the calculation of parent-progeny
correlations for age-7 heights and diameters. The height
correlation was r = .25 and the diameter correlation was

r = .15, which means that the parents selected for superior
height and diameters did not necessarily produce the tall—
est or thickest offspring. This was further supported by
the observation that the four tallest of fifteen parents
produced offspring which were only .2 foot above the over-
all progeny mean of 20.1 feet. In another test involving
some of the same parents, the offspring from the tallest
three parents averaged 2.6 feet taller than the overall
progeny mean of 21.0 feet. However, this information was
of limited use because the differences among the progenies
were not significant.

In a study of specific gravity in slash pine,
Goddard and Cole (1966) calculated a parent-progeny cor-
relation of r = .488, so relative parental performance
was a poor predictor of relative progeny performance;
the same was true for the specific gravity study conducted
by Squillace, Echols, and Dorman (1962). Stonecypher
(1966) found no correlations in height between randomly-

selected loblolly pine parents and their open- or

 

 

 

 

63

control-pollinated l- and 2-year-old offspring. Canavera
(1969), who practiced intense selection in even-aged
stands of jack pine in Michigan, found that offspring from
phenotypically superior mother trees were no better than
those from average trees. Also, phenotypically inferior
mother trees did not produce relatively poorer offspring.
Six other studies reviewed did not permit a pre-
cise estimate of the effectiveness of mass selection be-
cause (1) control material against which the selected-tree
progenies were compared originated from other populations
that those represented by the selected parents, and (2) no
parental data were given, so that parent-progeny correla-
tions might be calculated. Only by testing selections
against average material from the same population is it
provable that gain comes from the genetic improvement re—
coverable by mass selection and not from other sources of
genetic differences or from improved cultural treatment.
Genetic gains in forest tree species are being pro—
vided by sources of genetic variation other than those
recoverable by mass selection. The two major sources have
been provenance selection and family selection. This means
that the initial phase of an improvement program can be
cheaper and more efficient because (1) random selection
of parent trees is always much faster than phenotypic
selection, (2) provenance selection permits bulking of

seedlots, (3) the time saved in (l) and (2) permits a much

 

 

 

64

larger number of parent trees to be sampled, and (4)
provenance selection continues to be cheaper, because in-
dividual tree progenies are kept separate in testing.

It must be recognized, however, that mass selection
is the only form of selection practiced by nature, and the
genetic variation present within species (e.g., provenance
differences) attests to the effectiveness of mass selec-
tion in nature. There must be genetic differences present
within stands or there would not be genetic differences

present between stands or between populations. But natural

 

selection can operate on very subtle differences and has
available to it immense periods of time to make changes;
wood users are less patient with tree breeders. Never-
theless, since single-tree genetic differences are surely
present within stands, a judicious amount of time might
be devoted to single-tree selection within the framework
of an improvement program whose main emphasis is on pro—

venance or family selection.

 

 

 

INCREASING PRECISION OF GENETIC TESTS

Genetic gain in growth traits in the East German
population will not be possible in the present test. Gene-
tic differences between families were masked by uncontrol-
led variation. The breeder interested in East German
Scotch pine will be faced with these alternatives: (1)
forget about getting gain in growth traits, (2) assume
that a lOO-tree sample was insufficient to assess the
genetic variation present in a population, and include
more selections in the test, recognizing that there will
be increased costs, (3) increase the precision of the test
by improving nursery practice, providing better weed con-
trol, supplying irrigation, or increasing the number of
replications or otherwise modifying the experimental de-
sign. At Russ Forest, the present test permitted distinc-
tion of 18% differences in family mean heights. To
distinguish 9% difference, the number of replications
would have to be increased from 5 to 20. The costs of
increasing precision in this way will usually be greater
than the costs of improving cultural practices.

Those portions of the Kellogg and Dunbar planta-
tions which, when I was working in them, gave me the im-

pression of being unusually variable did in fact produce

65

 

66

the highest error variances. Mortality was high in those
portions and the factors causing the mortality caused much
variation among the surviving trees. The high mortality
at Dunbar was caused by flooding which could have been
avoided by choice of a more suitable planting site, had
one been available. The high mortality at Kellogg was
caused by insufficient water in the first growing season,
which might have been avoided by irrigating once or twice
during the summer.

Population samples of 20 trees will not provide
the same picture of genetic variability of a population
as will lOO-tree (or larger) samples. Assessment based
on small samples may lead to erroneous management decisions
even within the confines of a particular study. Heritabil-
ity estimates calculated for a test involving a small
sample size apply only to that test. The range in herit-
ability estimates for a particular trait in the present
study are more likely due to inadequate sample sizes than
to great differences in the frequencies of the genes con-

trolling the trait.

 

 

 

LIST OF REFERENCES

 

LIST OF REFERENCES

Barber, John C. 1964. Inherent variation among slash
pine progenies at the Ida Cason Callaway Founda-
tion, U.S. Forest Serv. Res. Pap. SE-lO. 90pp.

Canavera, David S. 1969. Geographic and stand variation
in jack pine (Pinus banksiana Lamb.). Ph.D. the-
sis, Michigan State UniversIty. 100 pp.

 

Ehrenberg, Carin Eklundh. 1966. Parent-progeny relation-
ship in Scots pine (Pinus silvestris L.)---results
from three progeny tests with plus and minus tree
progenies in southern Sweden. Studia Forstalia
Suecica. Nr. 40. Skogshégskolan. Stockholm.

54 pp.

 

Goddard, R. E. and D. E. Cole. 1966. Variation in wood
production of six-year-old progenies of select
slash pines. Tappi 49: 359-362.

LaFarge, Timothy. 1971. Genetic variation in the height-
diameter ratios and correlations of three pine
species. Ph.D. thesis. Michigan State University.

Langlet, Olof. 1937. Om mi1j6 och arftlighet samt om
farut-sattningarna f6r vaxtfaradling av skogstrad.
Sartryck ur Norrlands Skogsvardsfarbunds Tidskrift
Hafte I. Stockholm. 99pp.

Moffett, A. A. and Kathleen M. Nixon. 1963. One parent
progeny testing with black wattle (Acacia mearnsii

 

 

De Wild). World Consult. Forest Genet. and Tree
Imp. FAO/FORGEN 63 — 2a/5.

Namkoong, Gene, E. B. Snyder, and R. W. Stonecypher. 1966.
Heritability and gain concepts for evaluating
breeding systems such as seedling orchards.
Silvae Genet. 15: 61-100.

Nanson, A. 1968. Perspectives d'amélioration en premiere
génération par sélection des provenances. Silvae
Genet. 17: 121-156.

67

 

68

Nanson, Alphonse. 1969. Tests de descendances de pin
sylvestre. Ministere de l'Agriculture. Adminis—
tration des Eaux et Foréts. Belgique. Travaux---
Série E, No. 3. 51 pp.

Nilsson, Bo. 1968. Studier av nagra kvalitetsegenskapers
genetiska variation hos tall (Pinus silvestris L.).
Institutionen f6r Skogsgenetik. Skogshégskolan.
Rapporter och Uppsatser. Nr. 3. Stockholm. 117
pp. plus 22-page English summary.

 

Ruby, John Lindley. 1964. The correspondence between
genetic, morphological and climatic variation
patterns in Scotch pine. Ph.D. thesis. Michigan
State University. 227pp.

Squillace, A. E., R. M. Echols, and Keith W. Dorman. 1962.
Heritability of specific gravity and summerwood
per cent and relation to other factors in slash
pine. Tappi 45: 599-601.

Stonecypher, Roy W. 1966. The loblolly pine heritability
study. International Paper Co. Tech. Bull. No. 5.
Bainbridge, Georgia. 128 pp.

Stonecypher, Roy, Franklin C. Cech, and Bruce J. Zobel.
1964. Inheritance of specific gravity in two and
three-year-old seedlings of loblolly pine. Tappi
47: 405-407.

Stonecypher, R. W. and B. J. Zobel. 1966. Inheritance of
specific gravity in five-year-old seedlings of
loblolly pine. Tappi 49: 303-305.

Webb, Charles D. and John C. Barber. 1965. Selection in
slash pine brings marked improvement in diameter
and height growth plus rust resistance. Eighth
South. Conf. Forest Tree Impr. Proc: 67+72.

Wright, Jonathan W. 1963. Genetic variation among 140
half-sib Scotch pine families derived from 9
stands. Silvae Genet. 12: 73-104.

Wright, Jonathan W. and W. Ira Bull. 1963. Geographic
variation in Scotch pine--results of a 3-year
Michigan study. Silvae Genet. 12: 1-40.

 

 

Wright,

Wright,

69

Jonathan W., Scott S. Pauley, R. Brooks Polk,
Jalmer J. Jokela, and Ralph A. Read. 1966.
Performance of Scotch pine varieties in the North
Central Region. Silvae Genet. 15: 101-140.

Sewell. 1921. Correlation and causation. J.
Agri. Res. XX (7): 557-586.

 

 

 

 

APPENDIX

   

APPENDIX

Table 25. Variation in height, diameter and cone-bearing
in families #275 to #284 grown from seed col-
lected in N. H¢1and, southern Norway.

 

Height at Diameter at Trees bearing
age 11 age 11 cones

 

Family Russ Kell. Dunb. Russ Kell. Dunb. Russ Kell. Dunb.

 

% of mean % of mean % of trees
275 92 91 92 88 93 96 0 10 0
276 98 101 92 102 107 97 0 18 0
277 99 102 83 100 113 87 3 38 0
278 100 97 93 96 93 88 5 15 0
279 105 110 121 107 107 119 3 25 0
280 109 104 106 114 107 107 10 38 10
281 99 95 98 99 93 102 3 25 0
282 100 110 116 103 106 108 15 45 20
283 109 97 106 99 91 107 13 10 0
284 92 93 93 90 97 95 3 18 0

 

70

71

 

 

 

 

 

 

Table 26. Variation in color, needle retention and
Zimmerman pine moth attack in families #275
to #284 grown from seed collected in N. H¢land,
southern Norway.
Winter Needle Zimmerman
foliage color g retention moth attack
Family Russ Kell. Dunb. Russ Kell. Dunb. Russ
Grade Years % of trees
275 2.8 1.7 2.4 1.4 1.7 1.6 0
276 4.2 2.8 3.2 1.5 1.8 1.6 5
277 2.8 2.0 1.6 1.4 1.6 1.9 0
278 3.8 1.5 1.0 1.1 1.4 1.5 15
279 3.8 3.0 3.0 1.6 1.7 1.7 25
280 3.2 2.2 2.4 1.2 1.6 1.6 10
281 3.2 1.9 2.2 1.6 2.0 1.8 20
282 3.8 3.3 2.8 1.5 2.1 1.9 80
283 4.0 1.2 2.2 1.6 2.2 1.9 0
284 4.0 l 9 2.6 1.6 l 9 l 8 15

 

 

72

Table 27. Variation in height, branchiness.and cone-
bearing in families #285 to #294 grown from
seed collected at Achel, Limburg, Belgium.

 

 

 

 

Age-6 Branches per Trees with

 

 

 

 

height 4 whorls cones
Families Russ Russ Kell. Russ Kell.
% of an. Number % of trees
285 90 29 29 65 64
286 107 33 29 80 78
287 115 38 34 70 64
288 87 30 24 99 58
289 105 32 29 90 61
290 97 36 -- 100 ——
291 90 29 27 75 50
292 115 34 31 99 58
293 92 27 23 45 0

294 98 32 29 50 36

 

73

Table 28. Variation in height, diameter, and needle reten-
tion in families #295 to #304 grown from seed
collected at Hechtel, Limburg, Belgium.

 

-—- *-
‘—“—-

Height at Diameter at Needle Age-6
age 11 age 11 retention ht.

 

 

Fam. Russ Kell. Dunb. Russ Kell. Dunb. Russ Kell. Dunb. Russ

 

% of mean % of mean Years % X
295 96 96 96 97 98 102 1.7 1.6 1.9 94
296 91 96 90 95 97 95 1.8 1.8 1.9 88
297 103 106 99 103 102 103 2.0 1.9 1.9 102
298 99 104 105 103 104 108 1.6 1.7 1.8 107
299 99 98 -- 103 103 -- 1.9 1.9 -- 90

300 101 102 102 92 99 105 2.0 1.8 1.9 100
301 102 105 126 101 103 99 1.9 1.8 2.1 103
302 111 103 102 106 100 103 2.0 1.9 1.8 114
303 94 90 86 99 93 91 1.6 1.6 1.7 96

304 104 100 94 103 100 96 1.5 1.5 1.8 103

 

74

Table 29. Variation in needle retention in families #531
to #540 grown from seed collected at
Campine, Belgium.

 

 

 

 

Needle
retention
Family Russ Kell. Dunb.
Years
531 1.9 1.6 2.0
532 1.6 1.5 1.8
533 1.9 1.8 2.0
534 2.1 1.7 1.9
535 2.1 1.9 2.0
536 1.6 2.6 1.9
537 1.9 1.8 2.2
538 1.6 1.7 1.8
539 2.0 1.7 2.2

540 2.0 2.1 1.7

 

 

 

 

75

Table 30. Variation in diameter and cone-bearing in
families #321 to #340 grown from seed collected
in Ravershagen, East Germany.

_—
__7 J

 

 

 

 

Diameter at Trees
age 11 bearing cones
Family Russ Kell. Dunb. Russ, Kell. Dunb.
% of mean % of trees
321 102 99 103 35 65 0
322 100 99 93 15 32 0
323 111 112 102 25 28 0
324 100 101 92 25 45 0
325 96 96 84 15 12 0
326 109 102 111 5 2 0
327 103 100 127 20 35 0
328 103 100 102 25 22 0
329 98 100 95 35 25 0
330 101 101 106 5 4 0
331 97 94 92 15 30 0
332 89 105 98 20 38 0
333 89 105 112 40 15 0
334 104 104 86 35 35 0
335 103 101 95 35 38 0
336 101 100 112 35 22 0
337 92 88 103 10 22 0
338 109 105 108 25 15 0
339 99 97 93 35 40 0

340 97 91 87 15 22 0

 

 

 

 

76

Table 31. Variation in color and needle retention in
families #321 to #340 grown from seed collected
at Révershagen, East Germany.

 

 

 

 

 

 

Winter Needle
foliage color retention
Family Russ Kell. Dunb. Russ Kell. Dunb.
Grade Years
321 5.4 4.0 4.2 1.7 1.6 2.0
322 3.8 3.5 2.6 2.0 1.7 2.0
323 5.2 4.9 3.6 1.7 1.9 1.9
324 5.4 4.6 2.6 1.7 1.6 1.7
325 4.8 4.7 3.6 1.6 1.5 1.8
326 5.0 4.1 3.4 1.7 1.7 2.1
327 5.0 4.2 4.2 2.0 2.0 2.3
328 4.6 4.0 3.6 2.0 1.7 1.9
329 4.8 3.9 3.2 2.2 2.0 2.3
330 5.0 4.7 4.0 1.7 1.8 1.9
331 4.4 4.3 2.2 1.9 1.5 1.9
332 4.8 4.1 3.0 1.6 1.8 1.9
333 4.2 5.2 2.6 2.0 1.7 2.1
334 4.8 4.1 3.2 1.8 1.8 1.8
335 4.2 4.2 3.8 2.0 1.9 1.9
336 4.8 4.5 2.2 1.5 1.8 2.1
337 4.2 3.6 2.4 1.7 1.5 2.0
338 5.0 4.2 4.0 1.8 1.6 1.7
339 4.6 4.2 2.6 1.9 1.6 1.9

340 5.0 4.0 2.4 1.9 1.9 1.7

 

 

Table 32.

77

Variation in color, branchiness, height at age 5,
and height at age 9 in families #341 to #360
grown from seed collected at Neustrelitz, East

 

 

 

 

 

 

Germany.

Winter Branches per Age—5 Age-9

foliage color 4 whorls ht. ht.

Family Russ Kell. Dunb. Russ Kell. Dunb. Russ Dunb.
Grade Number % 7? %7K

341 5.0 4.6 4.6 27 29 25 104 91
342 5.2 4.0 4.2 28 37 29 105 91
343 5.0 4.4 3.4 33 36 26 109 108
344 4.8 4.7 3.4 28 32 25 105 99
345 4.2 3.9 3.2 28 33 29 94 104
346 5.4 —-— 4.6 29 —— 28 98 102
347 5.2 4.0 4.0 26 26 24 110 97
348 4.6 4.2 4.0 26 26 22 91 89
349 5.6 4.9 4.6 32 31 25 109 101
350 4.6 4.0 3.4 31 27 24 83 77
351 5.0 3.9 3.2 30 27 26 102 85
352 5.0 4.3 -—— 26 28 —— 107 --
353 5.4 —-— -—- 25 -— -- 104 ——
354 4.4 3.8 3.8 28 28 23 104 91
355 5.2 4.0 4.2 33 27 33 93 104
356 5.0 4.0 4.2 39 30 34 109 112
357 4.4 4.5 4.0 31 40 33 87 89
358 3.8 3.8 2.8 33 29 28 93 119
359 5.0 4.4 3.6 31 28 31 97 124
360 4.8 —-— 5.2 33 -- 33 97 115

 

 

78

Table 33. Variation in cone-bearing and color in families
#361 to #380 grown from seed collected at
Gfistrow, East Germany.

 

 

 

 

Trees bearing Winter
cones foliage color
Family Russ Kell. Dunb. Russ Kell. Dunb.
% of trees Grade
361 30 36 -— 4.4 4.1 -—-
362 5 36 0 4.0 3.8 3.4
363 20 18 5 4.0 4.3 3.4
364 5 25 0 3.8 4.1 2.8
365 30 43 0 4.2 4.3 2.8 _-
366 25 43 0 4.4 4.6 4.8 L4
367 0 4 0 4.2 3.8 3.2
368 15 —- 0 3.6 --- 3.4
369 15 18 0 3.6 3.6 2.4
370 25 36 0 4.4 4.6 3.0
371 15 61 0 4.2 4.6 4.0
372 0 l4 0 4.6 4.7 4.2
373 10 18 0 4.8 4.6 2.6
374 20 25 0 4.4 4.8 3.4
375 5 4 0 4.4 3.6 2.6
376 10 14 0 4.6 4.3 4.0
377 20 43 0 4.4 4.1 3.4
378 5 29 0 4.6 4.0 3.4
379 40 54 0 4.8 4.8 4.4

380 10 54 0 4.6 4.6 3.2

 

 

79

Table 34. Variation in needle retention and age-5 height
in families #361 to #380 grown from seed
collected at Gfistrow, East Germany.

-' j
r—T

 

 

 

 

 

Needle Height at.
retention age 5
Family Russ Kell. Dunb. Russ
Years % of mean
361 1.7 1.7 --- 104
362 1.8 1.8 1.9 96
363 1.5 2.0 2.1 99
364 1.9 1.7 1.9 93
365 1.5 1.7 1.9 103
366 1.6 1.2 1.7 101 ‘
367 2.0 1.9 1.8 99
368 2.1 --- 1.8 89
369 1.7 1.6 1.7 85
370 2.0 1.7 2.1 100
371 1.8 1.8 1.9 108
372 1.8 1.7 1.9 120
373 1.8 2.0 2.1 96
374 1.9 1.8 2.3 97
375 1.8 1.7 1.8 106
376 1.6 1.8 2.0 106
377 1.7 1.5 1.8 112
378 2.0 1.8 2.0 103
379 1.9 1.9 2.0 103

380 1.7 1.6 1.9 81

 

Table 35.

80

Variation in height growth, needle retention, and
age-10 height in families #381 to #400 grown from
seed collected at Nedlitz, East Germany.

 

 

 

 

 

Height growth Needle Height at
last 5 years retention age 10
Family Russ Kell. Dunb. Russ Kell. Dunb. Dunbar
% of mean Years % of mean
381 108 95 114 2.0 1.8 1.9 184
382 98 100 94 1.8 1.7 2.0 45
383 104 103 90 2.1 1.6 2.0 20
384 110 109 105 1.9 1.9 2.1 120
385 92 102 95 2.0 1.8 2.0 85
386 95 101 98 2.1 1.8 2.3 30
387 96 93 98 1.8 1.5 2.1 80
388 96 101 93 1.5 1.1 1.8 90
389 108 109 97 1.7 1.7 2.0 65
390 103 106 113 1.8 1.5 2.0 150
391 102 101 113 1.7 1.6 2.1 180
392 103 104 92 2.0 1.8 2.1 35
393 102 95 104 1.7 1.5 1.9 115
394 97 98 94 1.7 1.4 1.8 45
395 104 104 111 1.8 1.6 2.1 190
396 107 94 120 2.1 1.7 2.1 254
397 97 96 90 2.0 1.9 2.1 50
398 91 100 101 1.8 1.9 2.0 160
399 102 94 83 1.8 1.7 1.9 60
400 94 98 96 2.0 1.6 1.9 45

 

 

 

81

Table 36. Variation in height growth in families #501
to #520 grown from seed collected at
Joachimsthal, East Germany.

_—

 

 

 

Height growth
last 5 years

 

Family Russ Kell. Dunb.

 

% of mean

501 94 98 94
502 113 108 109
503 104 --- 101
504 99 106 96
505 105 102 109
506 103 94 100
507 99 106 104
508 108 101 94
509 99 107 104
510 101 103 95
511 96 100 98
512 97 89 93
513 99 --- 110
514 110 92 101
515 88 98 92
516 102 97 97
517 98 96 107
518 105 105 108
519 97 104 108

520 86 93 83

 

 

82

Table 37. Variation in Zimmerman moth attack among East
German families #321 to #400 and #501 to #520,
grown at Russ Forest.

 

 

 

 

Family Family Family Family
and trees and trees and trees and trees
attacked attacked attacked attacked
Fam. % Fam. % Fam. % Fam. %
321 35 346 35 371 10 396 25
322 30 347 20 372 20 397 5
323 35 348 20 373 30 398 33
324 20 349 35 374 20 399 0
325 20 350 5 375 45 400 40
326 30 351 40 376 20 501 15
327 50 352 15 377 45 502 15
328 5 353 25 378 20 503 15
329 25 354 30 379 30 504 15
330 30 355 10 380 0 505 50
331 30 356 20 381 5 506 35
332 20 357 15 382 20 507 35
333 25 358 10 383 25 508 40
334 5 359 10 384 45 509 5
335 30 360 20 385 15 510 30
336 35 361 35 386 23 511 40
337 30 362 15 387 15 512 15
338 25 363 5 388 40 513 20
339 30 364 15 389 15 514 15
340 35 365 10 390 8 515 35
341 25 366 20 391 25 516 30
342 25 367 10 392 25 517 20
343 15 368 10 393 38 518 15
344 5 369 15 394 15 519 40

345 5 370 25 395 15 520 30

 

 

 

VITA

Name:
Geroge Edward Howe
Biographical Items:

Place and date of birth: Indianapolis, Indiana.
Aug. 18, 1941.
Home town: Indianapolis, Indiana.

Education:

Undergraduate: Purdue University, 1959-1963.
B.8. - Forestry, June, 1963.

Graduate: University of Washington (Seattle), 1963—1965.
M.S.F. - Forest Genetics, August, 1965.
Thesis: "Colchiploidy in Populus trichocarpa
T. & G. ex Hook."
Advisor: Reinhard F. Stettler
Ph.D. Michigan State University, 1971

 

 

Experience:

 

Forestry Aid; Bureau of Land Management; Baker, Oregon;
summer, 1962.

Teaching Assistant (Surveying); University of Washington
Forestry summer camp; LaGrande, Washington; 1963.

Research Assistant; University of Washington College of
Forestry; 1964-1965.

Associate Plant Geneticist (sugar maple research); U.S.
Forest Service; Northeastern Forest Experiment Station;
Burlington, Vermont; September, 1965 - August, 1968.

Publications:

Gabriel, Wm. J. and G. E. Howe. 1968. Practical prob-
lems of a sugar maple selection program. N.E.
Forest Tree Impr. Conf. Proc. 15: 72-74.

Howe, George E. 1969. A vibrator for collecting pollen.
Forest Sci. (in press).

Howe, George E. Early results of a sugar maple prove—
nance study. N.E. Forest Tree Impr. Conf. Proc.
16 (in press).

Stettler, R. F. and G. E. Howe. 1966. The production
of homozygous tree material. 2nd Genet. Wkshop
and 7th Lake St. Forest Tree Impr. Conf. Proc.
1965: 67-69.

 

83

 

 

 

HICHIGQN STQTE UNIV. LIBRARIES
”HIHWIHHIITHIHWIWIHIIHHIHWHIHHIHHIHll
31293106154812