ANIMAL PRODUCTION. 467 



selves. For all ancestral cases the regressions appear to be more stable prop- 

 erties of a particular population than the corresponding correlations ; frequently 

 the regression of offspring on ancestor is the same as for a random sample, 

 though the correlation is changed. . . . 



" The value 0.5 in each case for the fraternal and parental correlations ob- 

 tained when random samples of a general Mendelian population are dealt with 

 does not depend upon the fact that the samples are random ones, but upon the 

 fact that for such a population the frequency of the heterozygote is twice the 

 geometric mean of the frequencies of the dominant and recedent homozygotes. 

 For, if any selected sample of the form pi'iAA)+2piQi{Aa)+q{'(aa) be taken 

 from the general population p''iAA)+2pq{Aa)+q^{aa), the parental and fra- 

 ternal correlations reached when the individuals of the selected sample mate at 

 random within the sample always have the constant value 0.5. . . . 



"Assortative mating within a random sample of the general population, if 

 positive, increases the parental and sibling regressions as well as the correla- 

 tions. The ancestral regressions diminish in geometrical progression, the corre- 

 lations not perfectly, but nearly so. In certain cases the expressions found 

 for the parental and sibling correlations were identical with those reached by 

 the very general methods previously employed by Pearson, and which have no 

 connection whatever with Mendelism, but this can hardly be more than a curious 

 coincidence. 



"[In] assortative mating within a selected sample, the regression of offspring 

 on parent depends upon both the assortative mating and the intensity of selec- 

 tion, and increases as those factors increase. Selection and assortative mating 

 affect the correlations in opposite directions, the decreasing tendency of the 

 former appearing to have the predominating effect in practical cases. The 

 sibling correlation is not raised so much by assortative mating nor reduced so 

 much by selection as is the parental one. 



" Fairly similar qualitative results were found throughout for somatic char- 

 acters, though not so much weight can be given to them as to those for gametic 

 characters. Moreover, it is the latter which agree most closely with observa- 

 tion. It is to the results for gametic characters, therefore, that we must look 

 for theoretical verification for experimental conclusions which, at first sight, 

 appear paradoxical, e. g., the closeness of the resemblance between cousins." 



See also a previous article (E. S. R., 27, p. 175). 



Mendel's principles of heredity, A. H. Marsh (Jour. East Africa and Uganda 

 Nat. Hist. Soc, 2 {1911), No. 3, pp. 52-60, pis. 2).— A popular exposition of 

 Mendel's law, with special reference to applying it to the improvement of 

 domesticated animals in East Africa. 



Concerning' the inheritance and the origin of species, J. Gross (Biol. Centbl., 

 31 (1911), Nos. 6, pp. 161-177; 7, pp. 193-2U; abs. in ZentU. Allg. u. Expt. 

 Biol., 2 (1912), No. 2-'t, pp. eU, 642). — ^A criticism of the extreme views of the 

 Mendelians. 



On the changes in the cranial capacity caused by domestication, B. Klatt 

 (Sitzber. Gesell. Naturf. Freunde Berlin, 1912, No. 3, pp. 153-179, figs. 9). — 

 Measurements of skulls of wild and domesticated animals are given, with a dis- 

 cussion of the changes that have taken place. The cranial capacity of domesti- 

 cated sheep, swine, and dogs was found to be smaller than that of closely 

 related wild species. 



The inheritance of the dun coat color in horses, J. Wilson (Sci. Proc. 

 Roy. Dublin Soc, n. ser., 13 (1912), No. 14, pp. i 83-2 Oi).— Additional data 

 (E. S. R., 23, p. 476) are cited as a further illustration that dun is recessive 

 to gray and roan, and dominant to- brown, bay, black, and chestnut. 



