The data

Ford collated records of female offspring of wild-caught females from various sources (see Ford, 1936 for references), together with approximate percentages of the morphs found in the populations from which the females came.

These appear to be the only records of the offspring from wild-caught females, who had mated in the wild, and therefore represent the only way in which the degree of assortative mating and multiple mating occurring in the wild may be assessed without making further collections.

Throughout this appendix the prefix 'proto-' usually given to imperfectly mimetic forms is dropped, and the categories of 'imperfect' and 'perfect' are merged, as the major mimetic genes carried by both forms are the same.

Offspring of wild-caught females

The data for the more polymorphic races is tabulated below (data from race dardanus, being almost entirely composed of hippocoonides females, is uninformative). The first three columns contain information taken directly from Ford's paper. The final four columns are an interpretation of the raw data in the light of the work done on the genetics of the morphs by Clarke and Sheppard (1959, 1960a, 1960b, 1962, 1963). The letters used to represent the genes are the allelic labels for the locus H (see Appendix 1 for details), with h representing the bottom recessive, hippocoonides. In many cases the second male gene cannot be known with any certainty. Where a gene is likely, it is labelled with a '?', and where one alternative is possible but less likely, it is put in brackets. Where more than two genes are possible, the cell is left blank. The minimum number of matings possible to achieve the observed brood has been assumed in all cases, but where a second mating must have occurred to produce the number of morphs seen, a second line has been inserted for the second male. In these cases, the second gene of one male is often known, but not which male carried it, so it again has a '?' after it.

Proportions of morphs in the races

Ford also published estimates of the proportions of the morphs as assessed by 'random sampling' in various areas (for references, again see Ford, 1936; data also in Appendix 1). For the races above, these are (listed in order of gene dominance):

Exact proportions of the morphs in the mainland population of race tibullus are not recorded.

Assessment of assortative mating (Chapter 5)

Given the proportions of the various phenotypes (morphs) it is possible to calculate the frequency of each gene in each race (since the frequency of the h gene is simply the square root of the proportion of hippocoonides morphs, and the others can be calculated from this):

Therefore it is possible to test whether or not the gene frequencies found in the males mating with a certain morph deviate significantly from these expected values for each race. The fact that dominant genes often cover the effect of others makes this analysis difficult. For race meseres, there is not enough data to analyse statistically, and for race tibullus the gene frequencies are not known, so only the data for races cenea and polytrophus could be analysed. The observed gene frequencies in the males for each morph are shown in Table A4-4.

Unfortunately, most of the data sets are too small to analyse statistically. Two can be analysed using a chi-squared test with some degree of certainty - race polytrophus, cenea and hippocoonides broods. For these, the data for salaami and trophonius need to be combined to make large enough expected values. For both sets the observed values are not significantly different from the expected proportions in the population (cenea: p=0.56, c² =1.18, n=40; hippocoonides: p=0.24, c² =2.87, n=31). It therefore appears that there is no assortative mating.

One further test that can be done involves analysing the genes carried by the females. They are themselves the offspring of wild pairings, and many of their genotypes are fully known - more so than the males. One outcome of assortative mating is that it results in more homozygotes than would otherwise be expected.

The expected ratio of homozygotes to heterozygotes for races polytrophus and cenea can be calculated from the frequencies of the genes, shown in Table A4-3. To do this, it is necessary to calculate the expected frequencies of each possible genotype (the frequency of one gene multiplied by the frequency of the other). These expected frequencies are shown in Table A4-5.

For race polytrophus, the genotypes of 39 females are known with relative certainty, whilst in race cenea, 14 are known (see Table A4-1). This gives expected ratios of heterozygotes:homozygotes in each race to be:

Race polytrophus	27.41 : 21.69
Race cenea	6.51 : 7.47

The observed ratios are:

Race polytrophus	33 : 16
Race cenea	9 : 5

A binomial test on each of these results indicates that neither observed ratio deviates significantly from the expected (polytrophus: p=0.919, n=49; cenea: p=0.152, n=14). Therefore it appears that the males are not choosing female morphs assortatively on the basis of the genes that they themselves carry.

Assessment of the frequency of multiple matings (Chapter 7)

Some wild-caught females had offspring which must have had multiple parentage (see Table A4-1). However, many which had mated more than once would be expected to have offspring which would not show this - particularly when one male had genes which were either exactly the same, or recessive to those of the other.

Considering the females which can be seen to have mated more than once, there is a different pattern in different races and morphs (shown in Table A4-6)

Overall, the two races show a mean frequency of double matings to be 10 %, and as explained previously, this can be taken to be a minimum value. Race cenea shows a minimum double mating frequency of 41%, which is surprisingly high as the real frequency is likely to be much higher. Race polytrophus, on the other hand, shows a minimum frequency of 5.1%, which is quite low. The fact that the bottom recessive hippocoonides gene is so common in race polytrophus (frequency = 0.592) means that many matings will involve hh males. Therefore many second matings will not be recognisable, as the genes from one male will be the same as or dominant over the hh male. By contrast, in race cenea, the more dominant cenea gene is much more frequent (frequency = 0.658), and so the common gene c will show up in broods with double parentage much more frequently. This data therefore suggests that, at least in some races, the frequency of double matings might be quite high.

References

Ford, E.B. 1936. The genetics of Papilio dardanus Brown (Lep.) Trans. R. Ent. Soc. London 85, 435-465.

Clarke, C.A. & Sheppard, P.M. 1959. The genetics of Papilio dardanus, Brown. I race cenea from South Africa. Genetics 44: 1347-1358.

Clarke, C.A. & Sheppard, P.M. 1960a. The genetics of Papilio dardanus, Brown. II races dardanus, polytrophus, meseres and tibullus. Genetics 45: 439-457.

Clarke, C.A. & Sheppard, P.M. 1960b. The genetics of Papilio dardanus, Brown. III race antinorii from Abyssinia and race meriones from Madagascar. Genetics 45: 683-698.

Clarke, C.A. & Sheppard, P.M. 1962. The genetics of Papilio dardanus, Brown. IV data on race ochracea, race flavicornis, and further information on races polytrophus and dardanus. Genetics 47: 909-920.

Clarke, C.A. & Sheppard, P.M. 1963. Interactions between major genes and polygenes in the determination of mimetic patterns of Papilio dardanus. Evolution 17, 404-413.