SELECTIVE MODES ASSOCIATED WITH INVERSION
KARYOTYPES I N DROSOPHILA ANANASSAE.
I. FREQUENCY-DEPENDENT SELECTION1
YOSHIKO N. TOBARI
AND
KEN-ICHI KOJIMA
Department of Biology, Tokyo Metropolitan Uniuersity, Tokyo, Japan,$
and
Department of ZwlGgy, University of Texas, Austin, Texas 78712
Received May 15, 1967
OME Drosophila populations in nature are known to contain two or more
inversion arrangements of a given chromosome under conditions of balanced
polymorphism in nature. WRIGHT
and DOBZHANSKY
(1946) and many others
investigated the causes of maintaining more than one chromosomal inversion
in a given population, using samples of polymorphic populations brought into
the laboratory. One of the common procedures used for such investigations was
to start random mating populations of Drosophila in cages with known proportions of different inversions, and to trace the frequency changes of the inversions
over generations until the frequency ceased to change at an intermediate level.
From such data, the estimates of relative fitness values f o r inversion genotypes
were determined using a statistical method such as least squares, maximum
likelihood or minimum chi-square. In these procedures, the values of fitness
parameters for all genotypes involved were assumed to be constant over generations. In general, it was noted that the conditions of inversion polymorphisms
depended upon many factors such as the geographic origins of inversions tested
together, the kinds of inversions present in a given cage population, the genetic
backgrounds used, the temperatures for cage cultures, and so forth. However,
one conclusion remained in spite of all these modifying factors-when inversion
balanced polymorphisms were observed in cages, the estimates of fitness values
indicated the existence of heterosis among the inversion karyotypes involved.
At the outset of the research to be reported in this series of papers, we planned
a set of experiments designed to evaluate interchromosmal interactions (epistasis)
between the inversions of the second (2) and third ( 3 ) chromosomes in Drosophila ananassae. The first phase of this endeavor was to critically study the
selective values associated with inversion genotypes at Chromosome 2, those at
Chromosome 3 and those jointly at Chromosomes 2 and 3. The cage population
procedure was used for a number of populations which were initiated simultaneously from a set of inversion lines with common genetic backgrounds.
1 This work was calTied out as a LJ.S.-JAF.\Nscientific cooperation prograni with rinancial support from the Japanese
Gakujutso Shinko Kai (The Japan Soriety for the Proniotirm of Srienrr.) and the 1J.S. National Sciencp Foundation
(Grant No. GF-264).
Adihecs of first author.
Genetics 57 : 179-188 September 19G7.
180
Y. N. TOBARI A N D K. K O J I M A
It was expected that the data from cages with either Chromosome 2 segregation
or Chromosome 3 segregation were to serve as a reference for the analysis of data
for joint segregations of both chromosomes. After the collection of data, however,
the analysis of single chromosome segregation data unveiled a mode of selection
which had not been given sufficient attention by population geneticists, namely,
frequency-dependent selection. Thus, the objective of this paper is to present primarily the results of the analysis of changes in selective values over generations
as inversion frequency equilibria are approached.
MATERIALS A N D METHODS
In 1951 PROFESSOR
D. MORIWAKIof Tokyo Metropolitan University received a strain of
Drosophila ar“ssae from the Genetics Foundation at The University of Texas, which was
segregating for Chromosome 2 inversions, L-A and L-B. This stock was partitioned into two
groups: one carrying only L-A and the other only L-B. Each group thus established and kept
in many replicated bottles, was later found also to be segregating with respect to two arrangements of Chromosome 3. In 1963 two new stocks, (In(ZL)-A; Zn(3L)-A) and (Zn(ZL)-B;
Zn(3L)-B), were obtained by making a number of pair-matings among flies of each group. The
two new structurally homozygous stocks will be referred to as the source stocks.
From the source stocks, a few sets of four homozygous lines, (2L=AA; 3L=AA), (2L=AA;
3L=BB), (2L=BB; 3L=AA) and (SL=BB; 3L=BB), were derived by the procedure diagrammed in Figure l . The notations such as “2L=AA’ should be read as “the Left arm inversion
stock
stock
(In 2 L A ;In 3 L A )
(In2 L B ; I n 3 L B )
Pair mating for 10 generations to
produce n (AA;AA)homozygous lines
Pair mating for 10 generations to
produce n (BB;BB) homozygous lines
&
L.
n inbred lines (AA;AA)-1 through
(AA;AA)-n
n inbred lines (BB;BB)-1 through
(BB;BB)-n
I
(AA;AA). (AA;BB), (BB;A+)and
(BB;BB) lines for each series.
Inbreed each (AB;AB) to produce 4
(AA;AA), (AA;BB), (BB;AA) and
(BB;BB) lines for each series.
181
FREQUENCY-DEPENDENT SELECTION
TABLE 1
Initial flies put into experimental cages
Inversions segregating (and fixed) *
?L
(3L=AAj
Lines
(AA1 AA)
(BB , A A )
( A A ; BB)
(BB , BB)
LL
(3L=AA)
100
900
900
100
2L
(3L=BBj
21.
(3LzBB)
100
900
9001
100
3L
(LL=4A)
3L
(%=AA)
100
900
900
100
3L
(2L=BB)
3L
(ZL=BB)
100
900
900
100
The figures in the table indicate the numbers of flies.
There were two cages for each of the eight conditlons.
* Segregating inversion is given above; fixed inversion is given below in parentheses
A of the 2nd chromosome is homozygous, AA.” The examination of Figure 1 reveals that four
lines belonging to a given set contain only one 2A o r one 2B, and one 3A or one 3B. Moreover,
the genetic background of four lines in a set is almost identically homozygous. On the other
hand, the lines belonging to different sets differ with respect to their genetic backgrounds and/or
the 2 and 3 chromosomes sampled from the source stocks.
The particular experimental materials used for the study in this paper is only one set of four
lines. They had a n identically inbred background from one of the source stocks, (In(ZL)-A;
In(3L)-A). These lines will be designated as (AA; AA): (AA; BB), (BB; AA) and (BB; BB),
omitting 2L and 3L designations.
To initiate experimental cages, replicated cultures of each of the four lines were built up,
and a large number of virgin females and males were obtained. From these flies, 16 cages were
started within a period of approximately a week. The initial inputs of flies for these cages are
found in Table 1. Equal numbers of both sexes were used for each line. There were eight different
genetic conditions, each condition represented by two cages.
All the cage populations were maintained by supplying two fresh cups of medium twice a
week. The average generation time was estimated to be 15 days under the present experimental
condition. Cytological determination of inversion genotypes was carried out, taking a sample of
25 larvae every day for a 6-day period at each generation from each cage. Thus, there was a
total of 150 larvae (or 300 chromosomes) to estimate inversion frequencies at any given generation of each cage.
RESULTS
The observed frequencies of the A-inversions in all 16 cages are presented in
Figure 2 (2L segregation cages) and Figure 3 (3L segregation cages) for the first
seven generations and some later generations. The frequencies observed for a
pair of replicated cages of a given initial condition are reasonably close in all
cases. The equilibrium frequency of the 2L A-inversion and that of the 3L
A-inversion are evidently in the vicinity of 0.50, regardless of the genetic condiet al. (1956) and
tions. This generally confirms earlier findings by MORIWAKI
MORIWAKI
(1958) on the 2L polymorphism and by TOBARI
(1962,1964) on the
3L polymorphism.
The pattern of frequency changes in Figures 2 and 3 gives an impression that
the underlying mechanism for the observed polymorphisms is likely to be the
superior fitness of inversion heterozygotes (heterosis hypothesis). Thus, a computer program was developed to estimate s and t in the Wrightian fitness model,
182
Y. N. TOBARI A N D K. KOJIMA
,
,
10
,
,
,
I
" " I ,
j
,
I
3 L segregation with ZL-AA
f (AIO
a
t
h
2 L segregation with 3 L' 88
aol
0
FIGURE
Z.-Frequency changes of inversion
2LA over generations (t) in cages. Top: 3L
chromosome was homozygous for A. Bottom:
3L chromosome was homozygous for B.
I
1
I
2
'
1
1
4
I
t
5
I
6
FIGURE
3.-Frequency changes of inversion
3LA over generations (t) in cages. Top: 2L
chromosome was homozygous for A. Bottom:
2L chromosome was homozygous for B.
WAA
= 1-s, W A B 1 and W B B
= 1-t, by using a minimum chi-square criterion
for inversion frequency data. The computer program consisted of two stages: the
first stage surveyed a wide range of s and t combinations to locate a set of combinations yielding approximate minimum chi-square values, and was followed by
the second stage, which iteratively minimized chi-squares starting from the firststage values. The results using data from the first seven generations are given in
Table 2, and the degree of fit to the observed data is graphed in Figure 4,which
contains the case of the poorest fit among all 16 cases (the bottom curve) and
TABLE 2
Estimates of Wrightian fitness values computed by minimum chi-square procedures under
the assumption of constant fitness vnlues over generations. W,,=I
Situation
A
us.
2L
2L
2L
2L
3L
3L
3L
3L
B
Fixed
3L=AA
3L=AA
3L=BB
3L=BB
%=AA
%=AA
2L=BB
2L=BB
Replication 9
Replication 1
f(A)*
0.1
0.9
0.1
0.9
0.1
0.9
0.1
0.9
wAA
0.429
0.488
0.204
0.420
0.334
0.257
0.295
0.323
input frequency of A-inversions.
+* fxz( A2 )=the
11.1 is significant at P=0.05.
0.243
0.722
0.221
0.525
0.371
0.136
0.243
0.171
X*t
WAA
w*,
x2t
4.59
6.56
11.06
5.08
7.37
2.71
0.263
0.400
0.428
0.431
0.390
11.44
0.370
0.264
0.172
0.315
0.251
0.623
0.400
0.300
0.300
0.126
7.55
2.90
19.34
3.45
3.76
2.60
8.75
8.71
10.34
0.400
FREQUENCY-DEPENDENT SELECTION
183
FIGURE4.-Examples of minimum chisquare fit of observed inversion frequencies
under the assumption of constant genotypic
fitness values. The data used are those of Replication 2 in the 2L segregation cages with
3LeBB (see Table 2). The lower curve is the
case of the poorest fit among the 16 cages
studied.
another case of an average fit (the top curve). An inspection of Table 2 seems to
give the impression that there is an extremely high degree of heterosis between
the A- and B-inversions at either 2L or 3L. However, the agreement between
the estimates from cages with f ( A ) = 0.1 and those with f ( A ) = 0.90 for otherwise identical genetic conditions is not satisfactory, as may be seen in Table 2.
Such a disagreement suggests that the genetic model on which the analysis is
based may be somewhat in error. For this reason, the assumption of constant W
values for all generations was relaxed, and a new analysis was carried out in the
following manner. Let the frequency of A-inversions at generation t be f ( A )t .
Then, the expected proportions of AA, AB and BB genotypes in generation ( t + l )
under the assumption of no selective value differences among the three genotypes
[ l - f ( A ) t ] andEBB= [1-f(A)tI2,respecare EA,*= [f(A)tI2,,A,=2[f(A),].
tively. The observed proportions of the three genotypes in generation ( t + l ) are
denoted by P A A , P A , and PBBwhere P A A PAB PBB= 1. Now, a set of fitness
values for the transition of generation t to generation (t+l) can be computed by
WAAt = PA.u'EAA,WABt = P A B / E A B and WBBt = PBS/E,B.
(1)
The same set of estimates is obtained by using the expected numbers and observed
numbers instead of the respective proportions in ( 1). The direct estimates of W's
in (1) will probably fluctuate greatly owing to many sources of error, including
sampling errors in f ( A ) t . In order to reduce such fluctuations, the values used
for f(A)t were those read off from the point for generation t on the curves, such
as those in Figure 4, obtained by fitting observed f(A)'s by the minimum x2
procedures. This process would induce some correlation between E's and Ps,but
+ +
184
Y. N . TOBARI AND K. KOJIMA
2L segregotion with 3 L s A A
6.0-
4.0-
-
f ( A ) = O . l at t + O
-
f(A)=0.9 at t = O
-_
-
2.0-
0.0
6.0-
amm
<am
2L segregation with 3 L e BB
--
-F
-
f(A) = 0.1 at t =0
f(A)= 0.9 at t - 0
-
4.0-
-
3 L segregation with
f ( A ) = O . l at t = O
6.0-
-
4.02
n
3L segregationrwith
6.0-
2L
f(Alt0.9 at t =0
1-2
2-3
3-4
Generot ions
-
2 L z BE
f(Al = 0.9 at t = 0
f(Al =0.1 at t = O
0-1
--
AA
4-5
5-6
0-1
1-2
2-3
3-4
4-5
5-6
Gene r ot ions
FIGURE5.-Changes in fitness value estimates over generations. Three consecutive columns
are the estimates for AA, AB and BB karyotypes from left to right. Notations such as 0-1 indicate the transitions of generations for which fitness estimates are applicable.
was considered not to be serious. The important feature here was to reduce
whimsical sampling errors so that a definite pattern of changes in W’s might be
observed.
The raw data for the computation in (1) and the W estimates for different
generations obtained by pooling the replicate cages are presented in the Appendix
Table. The expected standard errors for individual W ’ s are also included. The
standard errors were computed using sampling variance from multinomial distributions with parameters EA*, EAB,E B B and the number of individuals observed.
Such standard errors were not necessarily desirable for some cases, but were considered to give rough yardstick for judging a large difference in W’s.
Figure 5 summarizes the data in the Appendix Table. The eight sections
correspond to the eight genetic situations. The heights of three adjacent columns
represent the W estimates for AA, AB and BB genotypes from left to right. The
generation transitions are marked at the bottom of the figure. These W values
= 1.
are not standardized by setting W A B
Two trends stand out in Figure 5: there are large differences among the three
W’s at the early generations, but they tend to become equal later; and the highest
W among the three at the early generations is that for the minority genotype
under the respective genetic situations.
The pattern of changes in f ( A ) observed in Figures 2 and 3 can be reasonably
explained on the basis of these two trends of change in the W values. When the
FREQUENCY-DEPENDENT SELECTION
185
frequency of an inversion is small, the rate of increase is the greatest for the
homozygote of that inversion, then the second for the heterozygote, and there
will be the corresponding reduction in the number of the majority homozygote.
However, as the minority inversion increases in number, the differences among
the rates of increase for the three genotypes almost disappear. If the heterosis
hypothesis with constant W ualues should hold true, the W estimates given by
(1 ) should be consistent, except for random fluctuations, over generations as
easily shown by examining such a model of selective values.
Thus, it can be concluded that selection on the 2L inversions and 3L inversions
is frequency-dependent, where selective advantage is greater for genotypes with
frequencies far below their equilibrium frequencies, and selective differences
disappear toward equilibrium. Whether the state of complete selective neutrality
is achieved at the point of equilibrium was not answered, since there were not
appropriate data available at a later stage. All that has been definitely shown
is that the degree of selective value differences at or near the point of equilibrium
is not as great as the frequency change data may indicate, and there is no
heterosis as dramatic as indicated in Table 2 in a population at equilibrium.
DISCUSSION
Some evidences that fitness values may not be constant over a wide frequency
range have been reported by several investigators. For example, in a recent paper
PAVLOVSKY
and DOBZHANSKY
(1966) reported a private communication by PROFESSOR WRIGHT
indicating that WRIGHTand DOBZHANSKY’S
estimation method
of adaptive values failed to give satisfactory results when applied to some of
PAVLOVSKY
and DOBZHANSKY’S
new D.pseudoobscura inversion data. It was stated
that this failure was probably due to nonconstancy of adaptive values over a
range of inversion frequency. A more positive evidence was reported by EHRMAN
et al. (1965) and EHRMAN
(1966)’ where they found mating frequencies among
different genotypes of Drosophila to be dependent upon relative frequencies of
genotypes present. EHRMAN’S finding was that a minority genotype in a given
mating population is mated more frequently than expected on the basis of its
finding
relative frequency. Thus, there is a close parallelism between EHRMAN’S
on mating frequencies and the present result on “total” Wrightian fitness values.
As a matter of fact, the type of frequency-dependent mating system reported by
EHRMAN
may have been operating as the principal component of fitness changes
observed in this study.
There are, however, evidences of frequency-dependent selection on components
of fitness other than mating ability. KOJIMAand YARBROUGH
(1967) reported
that electrophoretically detectable F and S alleles at the Esterase-6 locus in D.
melanogaster are subjected to frequency-dependent selection. I n their study, the
mating system was experimentally controlled by placing females of known genotypes fertilized by males of known genotypes into culture bottles according to
fixed proportions. Thus, the factor of frequency-dependent mating was absent in
their specially designed experiment. Their conclusion was that frequency-depen-
186
Y. N. TOBARI A N D K. K O J I M A
dent selection with approximate selective neutrality among genotypes at the
point of equilibrium operated during the time between oviposition and adult
eclosion with respect to the three genotypes at the Esterase-6 locus in their maand KOJIMA,
terial. The allelic frequency changes tested in cages (YARBROUGH
in preparation) 'was what one expected from a system similar to the pattern of
fitness changes described in Figure 5 . The allelic frequency moved fairly rapidly
during early generations from either high or low direction, but slowed down
considerably later. By generation 30, the rate of change was so slow that the
approach to the equilibrium was barely detectable.
The mode of selection reported in this paper provides an efficient mechanism
for maintaining polymorphisms. Under this mode, there is tremendous selection
to push back the frequency of alleles or inversions when the frequency is deviating from its equilibrium value. However, there are very small selective differences among genotypes when populations are at or near equilibrium. This implies
that there is very little segregational genetic load in a nearly equilibrated population under this mode of selection.
SUMMARY
The mode of selection associated with a pair of inversions of the second chromosome and another inversion pair of the third chromosome was investigated, using
random mating populations in cages. These inversions were known to form
balanced polymorphisms. Initial flies put into cages were obtained from stocks
prepared to have a set of two particular second-chromosome inversions and two
third-chromosome inversions in an identical genetic background. The mode of
selection was shown to be frequency-dependent for either chromosome system.
When the frequency of an inversion homozygote in a population was smaller
than its equilibrium value, that homozygote was found to have a selective advantage over the other two inversion karyotypes. Such a selective advantage was
always associated with the minority karyotype, regardless of the kinds of inversions studied. When the frequencies of the inversions approached their equilibrium values, selective differences among different karyotypes diminished
toward selective neutrality.
LITERATURE CITED
EHRMAN,
L., 1966 Mating success and genotype frequency in Drosophila. Animal Behavior
14: 332-339.
L., B. SPASSXY,
0. PAVLOVSKY,
and TH. DOBZHANSKY,
1965 Sexual selection, geotaxis,
EHRMAN,
and chromosomal polymorphism in experimental populations of Drosophila pseudoobscura.
Evolution 19 : 337-3443,
KOJIMA,K., and K. M. YARBROUGH,
1967 Frequency dependent selection at the Esterae 6 locus
in Drosophila melanogaster. Proc. Natl. Acad. Sci. U.S. 57: 645694.
MORIWAKI,
D., 1958 Genetic studies in Drosophila anamsae. Japan. J. Genet. 33: 364-377.
MORIWAKI,
D., M. SHINAI,Y. YOSHIDA,
and M. TSUSUE,
1956 Frequency changes of two inversion arrangements in artificial populations of D. amnassae. pp. 95-97. Syudan Idengaku.
Edited by T. KOMAI
and K. SAKAI.
Baifukan, Tokyo.
187
FREQUENCY-DEPENDENT SELECTION
PAVLOVSKY,
O., and TH. DOBZHANSKY,
1966 Genetics of natural populations. XXXVII. The
coadapted system of chromosomal variants in a population of DrosophiIa pseudoobscura.
Genetics 53 : 843-854.
TOBARI,Y. N., 1962 Heterosis relating to a terminal inversion in artificial population of
Drosophila ananassae. Japan. J. Genet. 37: 302-309. __ 1964 Relation between
adaptive values and composition of the population in Drosophila ananassae. Evolution 18:
343-34.8.
WRIGHT,
S., and TH. DOBZHANSKY,
1946 Genetics of natural populations. XII. Experimental
reproduction of some of the changes caused by natural selection in certain populations of
Drosophila pseudoobscura. Genetics 31 : 125-156.
APPENDIX TABLE
The observed and expected numbers of A A , AB and BB karyotypes, and their W estimates with
standard errors for each transition of generations. The numbers are based on pooled
data from two replicated cages for each initial condition
f(R)=O.l
Transition
AA
at
2L segregation with 3L=AA
7/ 2.06*
59/31 .OO
1
0-1
2
66/27.00
12,’ 1.50
W
2.1 7 & 0.12
5.70i0.53
1-2
I
74/59.27
25/11.01
2
88/61.52
16/12.44
W
1.78 0.20 1.34t0.07
2-3
1
46/25.46
61/72.68
2
41/28.65
76/73.82
1.62i0.12 0.93 t0.06
W
77/74.97
3-4
1
47/36.17
2
80/75.00
45/36.90
W
1.26+ 0.10
1.05 t0.18
-1-5
1
44/42.14
73/74.73
90/74.91
2
29/40.10
W
0.88t 0.09 1.09 10.06
35/45.21
5-6
1
90/74.28
2
30/41.34
93/74.82
0.75 & 0.09 1.23C 0.06
W
2L segregation with 3L=BB
@1
1
4/ 1.82
59/29.37
2
2/ 0.99
39/22.34
W
2.11k0.62 1.88?~0.13
1
22/10.86
82/59.00
1-2
2
16,’ 6.74
76/50.12
W
2.20 ? 0.24 1.45t0.08
2-3
1
43/24.24
70/72.12
2
39/20.10
80/69.62
W
1.86t0.14 1.06t0.19
*
f ( A ) ~ 0 . a9t t=O
1x0
AB
BB
AA
AB
BB
84/116.96
72/121.50
0 . 6 6 i 0.03
51,’ 79.71
46/ 76.04
0.62? 0.06
43/ 51.86
33/ 47.55
0.7610.08
26/ 38.87
25/ 38.10
0.66 t 0.10
33/ 33.14
31/ 35.00
0.94-t-0.1 1
25/ 30.51
27/ 33.84
0.8120.11
107/128.90
99/118.82
0.83 10.03
86/113.80
64/ 96.72
0.71 10.04
86/ 95.04
62/ 75.41
0.86 +- 0.05
56/ 56.00
49/ 60.68
0.77 0.06
43/ 59.16
57/ 52.22
0.91 10.08
31/ 47.21
4Q/ 47.71
0.75 10.09
41/20.30
46/29.37
1.79It 0.13
56/33.71
76/47.4.6
1.63i0.10
54/48.72
70/70.01
1.1 2 t 0.08
74/61.77
78/69.46
1.16t 0 . 0 7
85/70.08
63/72.57
1.04t0.06
90/73.89
79/73.77
1.14C0.06
2/ 0.80
5/ 1.82
2 . 6 3 t 0.67
8/ 2.49
lo/ 5.82
2.47f0.37
10/ 6.24
18/12.7 1
1.51 t 0 . 2 4
20/12.62
23/19.88
1.37C0.17
22/20.76
30/25.22
1.13 ? 0.14
29/28.9 1
31/28.52
1.05 t 0.12
87/118.82
109/126.69
0.80 t0.03
46,’ 80.16
58/ 93.14
0.60 t0.05
37/ 53.64
31,’ 60.30
0.fjOt 0.07
95/122.31
100/118.82
0.81 iO.03
83/101.36
61/ 97.44
0.72 C 0.04
64,’ 79.07
60/ 72.69
0.82 t0.06
48/26.28
44/29.3 7
1.6610.12
60/43.89
81/4.6.91
1.55& 0.09
63/59.67
65/59.96
1.071-0.07
7/ 1.41
6/ 1.82
4.14k0.56
7/ 4.76
8 / 5.64
1.45 ? 0.3 1
23/11.27
20/12.3 7
1.83i0.20
*
188
Y. N. TOBARI AND K. KOJIMA
f ( A ) ~ 0 . at
1 t=O
f ( A ) ~ 0 . at
9 t=O
-
Transition
3-4
1
2
W
1
2
W
1
2
AA
AB
BB
-
39/22.92
44/52.75
23/ 30.34
1.70t0.13 0.83t0.07
0.76k0.11
4-5
34/35.00
81/74.91
35/ 4.0.10
9/ 7.96
13/14.99
8 / 7.06
1.05t0.16 0.97 t0.10 1.00+-0.12
5-6
20/24.01
46/49.98
34/ 26.01
33/36.45
59/62.10
33/ 26.45
W
0.87&0.12 0.94t0.07
1.28t0.12
3L segregation with 2L=AA
0-1
1
9/ 2.04
63/30.78
77/116.18
2
11/ 2.82
49/35.48
90/111.72
W
4.15t0.46
1.71f0.11 0.73 -+ 0.03
1-2
1
14/ 6.93
59/50.64
77/ 92.43
2
14/ 8.01
75/53.30
61/ 88.71
W
1.88t0.25 1.29f0.08 0.76 1- 0.05
2-3
1
14/15.08
88/64.95
48/ 69.98
2
18/15.75
73/65.70
59/ 68.55
W
1.04t0.17 1.23t0.07 0.77 -C 0.06
3-4
1
24/22.94
82/71.43
44/ 55.64
2
16/23.16
94/71.57
40/ 55.26
W
0.871-0.14 1.23t0.06 0.76 t0.08
4-5
1
27/28.38
79/73.74
44/ 47.88
2
29/28.52
75/73.77
46,’ 47.72
W
0.981.0.12
1.04f0.06 0.94t 0.08
5-6
1
39/31.61
69/74.49
42/ 43.91
2
34/31.88
85/74.55
31,’ 43.58
W
1.15+0.11
1.033E0.06 0.83 t0.09
3L segregation with ~ L E B B
0-1
1
7/ 1.22
69/24.57
74/124.22
2
5/ 1.04
62/22.83
83/126.14
W
5.30t0.67 2.76k0.13 0.63 t0.03
7/ 7.86
77/52.97
1-2
1
66/ 89.16
2
7/ 5.58
73/46.73
70/ 97.68
W
1.07f0.27
1.51k0.04 0.73&0.05
2-3
1
27/20.99
91/70.23
32/ 58.79
2
13/15.84
93/65.82
44/ 68.34
W
1.05f0.15 1.35t_0.06 0.59t0.07
3-4
1
28/31.05
90/74.40
32/ 44.55
2
28/22.15
72/60.94
25/ 41.90
W
1.08t 0.13
1.20t 0.06 0.66 3EO.09
1
30/36.17
4-5
90/74.97
30/ 38.87
2
27/33.56
91/74.78
32/ 41.66
W
0.82t0.11
1.21F0.06 0.77 k 0.10
5-6
1
32/38.4-0
92/74.99
26/ 36.60
2
36/37.35
84/75.00
30/ 37.65
W
0.9OtO.10 1.17a0.06 0.75 k 0.10
AA
AB
BB
~~
52/ 61.05
46/ 57.48
0.83 t0.07
38,’ 49.25
23/ 45.54
0.64&0.09
30/ 41.97
31/ 37.95
0.77 -C 0.10
74/69.29
82/70.76
1.I 1 t0.06
89/73.40
99/74.22
1.27 t0.06
81/74.75
78/75.00
1.06k0.06
24/19.65
22/21.78
1.12t0.14
23/27.35
28/30.24
0.88t0.12
39/33.27
41/37.50
1.14-+ 0.11
77/118.82
107/120.69
0.77t0.03
55/ 84.60
80/ 98.90
0.73-+0.05
53/ 59.54
52/ 77.33
0.78+-0.06
42/ 48.90
44,’ 62.21
0.78t0.08
38/ 45.21
48/ 53.28
0.87 t 0.08
35/ 43.91
39/ 48.57
0.80f 0.09
64/29.3 7
32/27.72
1.68t0.12
81/56.10
63/4+5.80
1.41f0.08
75/69.93
82/60.75
1.21f0.07
76/73.49
85/68.78
1.I4 t0.06
84/74.28
80/72.24
1.12C0.06
92/74.40
89/73.58
1.22+0.06
9/ 1.82
11,’ 1.59
5.94t 0.54
14/ 9.30
7/ 5.30
1.41t0.27
22/20.54
16/11.93
1.21f0.17
32/27.60
21/19.01
1.13f0.14
28/30.51
22/24.48
0.91 10.12
23/31.61
22/27.87
0.76f 0.12
102/119.61
93/118.W
0.82-CO.03
69,’ 91.97
55/ 84.60
0.70t0.05
44/ 68.15
38/ 60.11
0.6410.07
41,’ 55.08
35/ 49.26
0.73+- 0.08
37/ 49.43
32/ 45.73
0.82k 0.09
46/ 47.04
43/ 39.95
1.03 & 0.09
48/28.67
50/30.08
1.67 t0.12
67/50.97
80/56.10
1.37 1- 0.08
78/65.91
81/69.69
1.17 +- 0.06
86/71.63
93/72.83
1.24t 0.06
63/73.35
95/73.64
1.18f 0.06
81/73.92
74/66.44
1.11f0.06
O/ 1.71
7/ 1.92
1.82?0.52
14/ 7.07
15,’ 9.30
1.80k0.24
28/15.95
31/20.21
1.65t0.16
23/W.28
21/26.91
0.88 t0.13
20/27.23
22/29.64
0.83 +- 0.13
23/29.00
17/27.62
0.70t0.12
* All entries such as 7/2.06 mean that seven individuals were observed while 2.06 individuals
were expected