Gene Frequencies in London's Cats

GENE FI~,EQUENC]ES IN LONDON'S OATS

BY A. G.. SEARLE, Unive,rsity Collecye, London

])uring the stunner of ].9~7 I examined 700 cats, which had been destroyed at three of the main animal clinics in London., in order to estimate the frequencies of known genes in the city's cat population. It will be desirable to give a rgsmng of the known :facts of cat genetics. Whiting (1918) di~dded tabby cats into three categories : lined. (usually described by breed ers as Abyssinian), striped and blotched, with dominance in the order given. Sewall Wright (1917) analysed the tabby pattern, pointing out that the banding of individual hairs to form the light stripes is due to a gent which acts only on black, as in the agouti-coloured rodent. He said that the width of these 'agouti' bands varies independently of the character of the black stripes. Action of the' non-agouti' gent gives solid black, which is recessive to tabby (Doneaster, 190,1.). The tabby stripes can still be seen occasionally in black kittens. Soi[id white is due to a dominant gent (Sewa]l Wright, 1917). Doncaster (1912) concluded that yellow was sex-linked, and this is the accepted[ xdew. Bamber (1927) gave a comprehensive rexdew of previous literatm:e.on the yellow and the tortoiseshell cat, normally a female heterozygous for yellow. She pointed out that the rare tortoiseshell male had been accounted[ for by failure of sex-linkage, non-disjunction or sex-reversal. Stwall Wright thought the tortoiseshell male had a pair of X-chromosomes like the normal female, but some other influence had[ over-balanced the sex-determining mechanism. TMs problem is not yet settled. Tjebbes (192~1) worked on Siamese cats, and found that their pattern was due to a gent for 'weakened chromogen' acting on chocolate-black. The latter appeared in full intensity in crosses introducing normal chromogen, and was than epistatic over tabby. The gent for 'weakened chromogen' could be combined with tabby to give a 'striped Siamese'. Keeler & Cobb (1933) showed that this Siamese dilution was recessive to silver, and allelomorphic with it. These alleles are regarded as members of an albino series, other possible members being Bm'mese (included by Castle, 1940) and full albino (reported by Bamber & Herdman, 1931). No adequate analysis of white-spotting has been made, but the genes involved seem on the whole dominant. A number of loci are probably concerned, as in other mammals studied[. The gm~e for dihition is recessive (Doncaster, 1904). Bamber (].927) thought poly- dactylism dominant over normal, and this has been confirmed by Danforth (1947). )'Ianx tail seems to be incompletely dominant, and long hair to be recessive (Banter, 1927), but it seems likely th.at several genes are concerned.

CLaSSIFIO~TmN We may take as our standard the short-haired striped tabby, with a fair amount of yellow in the banded hairs, a normal tail and digits. This striped pattern resembles that of Fel'is silvestris. Colour genes can be classified as follows: t 5. Abyssi~'l,,ia'n (Whiting's 'line'). Stripes very fine and close together, except on the legs, tail and nose. The first impression is of a unifomn agouti cat, but it has a dark vertebral line. A. C4. SJ~At~L~ 215 +. Ntril)ed tabby. I found awe major variations of the normal pattern, which may be due to modifiers ag other loci. 17'. Blolched, I,a,bby. Stripes very wide, few in number and much broken up. Prof. Hatdane states gh.at the type of blotched tabby commonest in America is decidedly different front gho English type, so there may be other a.llelomorphs. rb. Nor>e~yo,tt~;. (Jags with a,/c~ are black. The gabby pattern was somegimes seen, bug no ageempt was made to distinguish dift!erenb ~ypes. W. Whig< All white cats seen had yellow eyes. B. Choeolo, l,e-blaclc, I found this only in the Siamese cab, when combined with c s (see below). y. Yellozv. This is a sex-linked gone, y males and y/y females being yellow, +/y females toreoiseshell. Gone a, has no effect on y all({ y/y cats. On the ocher hand, the gabby genes can be distinguished, for the light agougi stripes of the tabby cat are replaced by light yellow stripes composed of uniform hairs. On -I-/y cabs the effect of et is very clear. Bug the different types of tab by ill the gabby-tortoiseshell are sometimes rather hard go distinguish; in three gags (classified +/y t?) the exact pattern could not be definitely determined. d. Dil~tte. Gives Malgese blue (dilute gabby), I~ussian blue (dilute black), and cream (dilute yellow). U'. Nilver. l~emoves yellow tu the banded hairs, bug Dr Spurway points out bhag c~1~ does not affect ~he yellow areas in a tortoiseshell. I found intermediates between the clear-cat silver tabby and normal; possibly there are several slightly different ech, genes, as in the rabbig. c ~. ~9ia'mese dih~tio~'~, e'~B pkenogype is the Siamese cat. This is born white; lager the tips of gho tail, limbs and nose eolour up go a greater or less degree, depending on external temperature (Iljin & Iljin., 1.930). s Wldl, e-sl)ol, ti,j N. Since the amount of white varied continuously between 0 and 100 ~o I classified the cats as follows: s no white-spogging; s less than white; s to g2 white; s over } white, bug nob all white. Other characters noted were: Ha~;r le.~Nth. Long hair is said to be recessive, bug there were several, different grades of hair length, depending perhaps on modifiers of the main gone. I classified for medimn hair and long hair (half-Persia, n and %11 Persian). 51'cd[ le'r~,yth. The ~Vlanx character is dominant, Chore is no definite evidence on whegh.er gho homozygote is ~dable. Intermediate tail lengths were found. I)olydc~otylis,m. Dominant. More than one additional digit may occur; fore- and hind- limbs may be unequally affected. ]~ye-cotours were not noted on the whole sample as pupils were often so dilated by death ghag no aeouraee observations could be made. The significance boundary for X~ tests is taken ag P = 0-05.

~]~SULTS I examined 700 eats in. all, of which 454 were seen a,g dinic 'A', of the AnimM ]R,escue League (R.S.P.C.A.), Islinggon; 202 a~ clinic 'B', of Ore: Dumb Friends' League, Victoria; 44 ag clinic 'C', the Ma,ylmw Home (R.S.P.C.A.), Willesden. Some had. been stray cats, attd ochers unwanted or ill, collected from surrounding districts. 216 Ge'ne fl'effuencies in Londo~?s c~ts There were 356, or 50-9 %, male e~;ts..Five hundred were classiffed for age and mMe castration (J-) in adults. Results were as follows:

Adults ].65 90 87 Kiggens 94 6v'~

[['he:re is thus a significant tendency (Xs=5.06) to abandon :female rather than rome kittens. The frequency of name castrates is here ~19.1. %, and may well be higher in f/he total poptf[ation, as they are probably less likely to stray. The numbers of eats falling into ~he classes of coat pattern are shown in Table 1. Inehded in the striped tabby group (+) are sixteen eats with modified patterns. Nine of these (three males and six fem Mes) had thin broken stripes. The obher seven (five

Table 1. Coat patterns fou,nd, with the degree of white-spotting Phcm)byl)e S,, S~ S,z Sa TeaMs W 3 ] ------1 t -9 5 ------5 6, tL d' [ O 0 0 ll -I- c~ 18 ] 9 5 3 46} 82 -9 16 16 2 2 36 t ~ 3 27 35 11 13 s6} $ 36 27 9. 6 78 166 a c~ 71 73 19 15 178[ -9 69 63 22 9 163, 341 Y 5 6 3 0 i~I -9 1 2 0 0 ~ 17 t~y d' 10 12 4 2 28}32 ? 1 3 0 0 4f +/y 0 1 0 0 ~} ? ] 4 I 1., 8 + /y t" -9 4 6 2 0 12 12 +ly t'./ ? 1 ], 0 1 3 3 -I-/ya 9 ii 16 3 4 32 32 ~13 3 i ------1) 2 ? 1; TogMs 280 282 82 56 700

Table 2. Dist,']but]on of the diluting gene Pheno~ype S o 6'i 5'~ To,Ms + c? 1 0 0 I 1 o 0 2 0 2, 3 t ~ c~ 0 1 0 1 1 ~ d' 0 4 0 ,1:/8 I 2 1 +/y9 o 2 o 2 2 ToteMs 2 11. I 1,1-

names and two females) had a fairly uniform appearance, with fMng stripes, sometimes broken, on ~he ~runk, and darker stripes on head, legs and ~ai[. This was reminiscent of Siamese dilution, but the only undoubted s~riped Siamese which I have seen had no sig~ of stripes oll the trunk, which was Mmost pure white. So i have not classified these seven cats under c s. Included in the y phenotype cats are bwo wibh broken sgripes, and one of bhe +/y phenobype had broken stripes. A. G. S~Ar*L~ 217 Other genes whose effects were seen are: eca. Six male and five female silver tabbies were seen, but seveu others see.med~ inter- mediaee between these and normM. d. Table 2 gives the distribution of the diluting gene. IIai'r-le'ngth. Sixteen males and thirty-three females had medium hMr, seven tomes and six females had long hair. ]~esulbs with classified adults and male castrates (.3~) were: Short Micdium Long TogMs ? :H:(i ]5 (9.1%) ~ (2.~t%) 165 g 84: 3 (3.3%) 3 (3.3%) 90 .~ 7,t. 10 (11.5%) 3 (3.4:%) 87 TerMs 304 28 10 342 There is no significant difference between the results for long hair. But there is a significant tendency (X~=~i'36) for male castration to increase hair length from short to medium. Medium hair length seems commoner in females than in functional males (X~ 2.86). it is well established that castration inhibibs baldness in Man. Tail length. Four mates and one femMe had Manx tails. Six males and three females had short tails, apparently not accidental. PoZydactylism. One rome had an enlarged thumb on the front paws, and an extra digit on the hind paws. One rome and two femaies had an extra digit on each paw. One female had two extra digits on the hind paws and one extra on the front paws.

GENE FREQUENCIES The following is a classification ~th respect to y: + / + or + +/y y/y or y 'l'o~M 277 (269.6) 54 (64.5) 7 (3.9) 338 ~, 311 (315"3) 1? 42 (37'8) 354: White and Siamese ca~s are omitted as unclassifiable for presence or absence of y. These figures can be used both to estimate the gene fl'equency and to check the hypothesis of random mating. If :c be the fl'equency of y, we expect females to be in the ratio (1-x)~+/+ : 2x (1-x)+/y:x~y/y, and males (l-x)+ : xy (omitting the tortoiseshell male). There were 2 x 277 + 5~1 + 311, or 919 + genes, and 5~1 + 1~ + ~2, or 110 y genes; that is, ll0 out of 1029, or 10.69 ~ y genes. The expected values worked out from this fl'equeney are shown in brackets above. There are rather too many yellow and too few tortoiseshell females, suggesting a slight degree of selective mating, but the divergences from expectation were not significant (X~ = ~'90 with 3 ~ of freedom) and the hypothesis can be accepted as reasonably accurate. Table 3. Genefrequency estimations Phenotype Gene Chara, cger No. ch~ssiliable Phcnotypc 11o. fl'equency (%) fl'equency (%) W 700 6 0'9 0"~ t L 296 1 0.3 0.2 + 296 99 33.6 18.5=t=2.3 t b 296 196 66.4 81.4• a 6~i3 373 58.0 76.2=t= 1.7 y 692 -- -- 10.7 ~: 1.2 d 694 14: 2.0 1,i.2~ 1.3 c ca 270 11 4.1 20.2 S 692 420 60.7 -- Table 3 shows other gene frequency estimates. On the assmnption of random mating, recessive gene ii'equencies are taken as the square root of the phenotype frequency. 218 de'he fl'eguencies i~, Londo~'s cats Standard errors have been calculated where" numbers justified this. In Table 3, cats classified for f,abby patterns do nol~ include ~abby-torboiseshells, since the type of tabby pabl, ern could not be determined in three cats of that group.

T:~STS FOI~, ]:[OMO61.:BNEJTY AND INDlgPI,3NDEN(~I,] TesLs for homogeneiey were carried out between the ehrec clinics. Sex ra~ios were as follows : A B C 9 246 79 ] 9 d' 208 123 25

The d_ilI'ercnees between A and CO are not significant, but there is a significantly greater proporl)ion of females at A ~han at ]3 (X2 = 12"6). This may be due to h.uman l?references. The nmnbers of cats falling into the main colour groups were also compared. These are shown in Table 4. T,~bte 4. Comparisons oJ' 2)henotyl~e n,tmbe,rs Clinic fL PlmnoLype A B C + 49 25 9 t o 106 4:2 16 a 233 94 I4: y aud left 26 19 4 Table 5. 2 • 2 tests Ch~rge~crs compared X" Sexes • A, a 0"1.7 Sexes xs s 1'1.9 +, t b xA, y 0.02 A', sxA, a, 0.12 The difl'ermmes between A and ]3 a.re not signifi.caut (X'-'=,I-14- with 3 ~ of freedom). The differences between A and C are significant (XP= 9.47), but the population sample at C was statistically inadequate. 2 x 2 tests for independence were made on the whole sample, to reveal associations between characters, which would imply lack of randomness in mating, or human selection. Table 5 shows characters compared, with values of X2 fom~d. None was significant, ,so no associations Were found. Age. For the last~ 500 ca~s, which had been classified for age, a comparison was made bebween t~he numbers of ldttens and adults in the main coal>eotour classes. I:~esults were : + t ~ a All y All +/y To,at ]~t~ens 18 4:4: 73 7 15 157 Perceu~uge 11.47 28.03 46.50 4.46 9.55 Adult)s 40 67 178 26 25 336 t?er eent,~ge 11-91 19-94 52.97 7.7,1: 7',I.4 Comparing figures for t o and a, X~= 3.92 which is significant for P = 0.05. The higher percentage of blotched, tabby and the lower percentage of black in the kittmr population, compared wiLh adults, seem ~o indicate the working of human selection. For this is exercised n~ainly on the young progeny, the ldtkens here beit~g t~hose which have bee~t selected against, ned destxoyed, wlfile gh.e adnlts are those which ihave esc~pcd this early selection. So bla.ek kittens appear to be favoured, and therefore less liable to destruetiou ~han blo~ched gabby. Similarly, yellow kittens seem to be preferred fo tortoiseshell

(X2 = 2.28). A. G. SEAI~LI~ 219 Ua,st,ratgon. A coral)arisen was made between castrated-rod fm~etionat adult males in the main coat-colour classes. [.P,.esults were: + t ~ a. All y '];'ol~a,l (Jastra,~od 1.]. 17 45 ] 2 85 ]i'tltlctlonal ] 2 ] 8 ,51 8 89 There are no signifi.cant differences between these :figm:es. So there is no evidence that any type of cat is more lfkely to be made ster.ile than any other.

])ISCUSSION We have been dea]ing here ~dth a sample from a polymorphic population, maintained thus by h.mnan selection. Yet it is very seldom that man. has any direct influence on the cat's choice of a mate, although, by castration, he reduces the male breeding population by about half, and the female by an unknown amount. ])arwin (1859) remarked on the difficulty of pairing cats, ' o~qng to their noctm:nal rambling habits', and in this matter eats are certainly the most independent of all. domestic breeds of animal. Calculations made on the present sample, based on numbers of yellow and tortoiseshell, and on comparisons between different combinations of characters, show fairly conclusively the randomness of mating. Human selection is exercised mainly on the progeny, demon- strated here by the large number of ldttens destroyed, the greater proportion of females among them, and other divm'gences from the make-up of the adult population. ]But the fact of random mating indicates that a city's cat population is a less artificial assembly than a similar population of other domestic breeds. Indixdduals are less isolated and less protected from the hazards of the en\dronment. Ph.enotypic differences in fitness can therefore sh.ow thhemselves more easily, keeping some types rare despite human preferences. Wholly white eats are uncommon, for instance, although some degree of white- spotting is very comm.on. Th.e former might we].l be favoured, by hmnan selection, but it is already known that nearly a!l white cats with blue eyes are handicapped by deafness. Blotched tabby and black are both very common, with gene fi:equencies markedly above that for the wild type pattern. We have seen that black "kittens appear to be favoured by hmnan selection, having the advantage of being considered lucky by common superstition. They, and blotched tabby, may well have other special advantages in industrial areas. The bio]ogical significance of the different gene frequencies cannot be assessed, however, until comparisons can be made ~dth results from elsewhere. It is certain that other populations have a very different composition; until data are available fl'om, them no important conclusions can be drawn from this investigation.

SUB'IB'IARY 1. The results are given of a survey of inherited characters in 700 London cats. 2. The hypothesis of random mating was tested by using the results for yellow and tortoiseshell, and was found satisfactory. 3. Oene frequencies were estimated on this hypoth.esis, and are given in Table 3. 4:. There was a significant tendency for male castration to increase hair length. I should like to thank Prof. J. ]3. S. I-Ialdane, who suggested this investigation, for the great interest he has shown in it, and for his constant help and encouragement. I also thank Major G. Hancock, Chief Veterinary Officer of the I~.S.P.O.A., and Mr W. E. Murts, Veterinary OtKcer of Our Dumb Friends' League, ['or permitting me to work at clinics und.er tlheir charge. 220 Gene ,frequen, cies in London's cats

~EI~EI~ENCES ]~A~m~, ]~. C. (1927). Genel~ics of domestic c~s. Bibliogr. genet. 3, ]-83. BA~fU:,m~, I~. C. & :[-I~:al)~.,~N, E. C. (1931). Two new co~-colour t~yl?es in c,%~s. Natu~.e, Lond., 127, 558. C~S:rL:E, W. E. (1940). 2Iammalian Genetics. Ox:[brd Univ. Press. D~trvo~T~r, C. H. (]947). Herediby of :polydact~yly in bhe c~%1~. J. Here& 38, 107. :DAi~w:l:~, C. (1859). (/~he Origin of Sl~ecies. Do:~cAs~':m~, L. (1904). On t)he i~dmrib~nce of ~or~oiseshe][ and rel,%~ed colours i~l c.%~s. P~'oc. Camb. Phil. Soc. 13, 35. Doi~oas'e~, 7L. (1912). Sex-limited inheritamee in c~%ts. Science, N.S. 36, 14:~L Ir~,~, N. A. & I,:,J[N, V. N. (1930). Tempe:l:~ure effeobs on bhe colour ofbhe Siamese c~fi. J. Here& 2i, 309. K~c:~r~ & Con~ (]933). Alldomorphism of silver and ,Si~meso co~1, varia.~ions in the domestic c~. J. Ho'ed. 24, I81. [[?J:~m~:~s, K. (1924:). Crosses with Si~mmse ca,ts. J. Genet. i4, 355-66. W~-~:]:'.r:[~o, :P. W. (1918). Iuheritance of coi~t-colour in c~l,s. Y. EXp. ZooL 25, 539-69. Wt~cm'r, S. (1917). Colour inheritance in m~umnds. X. The c~g;. J. Here& 9, :139-~H:.