874 



SCIENCE. 



[N. fc). Vol. XXII. No. 574=. 



of mine that had ' misrepresented ' his hy- 

 pothesis, and so needed ' correction.' In fact, 

 I did not suppose I could have misrepresented 

 it, because I had given it merely by quoting 

 his own words at length. Two of the para- 

 graphs quoted by me Gadow quotes in his 

 recent paper, following them with the words 

 " I think I had stated the case fairly. It left 

 no doubt about the definition of at least one 

 kind of orthogenetic variation" (p. 638). But 

 we are indebted to his recent paper for the 

 very concise statement that ' cases of ortho- 

 genetic variation are simply ontogenetic 

 stages, passing reminiscences of earlier phylo- 

 genetic conditions.' The basis for his assump- 

 tion that the abnormalities in number and 

 arrangement of the horny shields of turtles 

 are ' orthogentic variations ' is a table of per- 

 centages of abnormalities, made from 76 speci- 

 mens (4Y new-born, and 29 of various sizes, 

 from three inches to 'large'). This series of 

 percentages of abnormalities, the percentage 

 decreasing with age, is supposed to indicate 

 that turtles ' amend their scutes ' and ' grow 

 out of these irregularities by the reduction or 

 squeezing out of certain scutes.' 



Gadow states (p. 639) that with the addi- 

 tion of my embryos to his 76 specimens, ' the 

 percentage still decreases with age ' ; and gives 

 the following revised table, including both 

 sets of turtles: 



Per 

 fi Cent. 



Of 73^ embryos or newborn, 53 are abnormal. . 7* 



Of 9 specimens from 3 to 8 inches, 3 are ab- 

 normal 33 



Of 19 specimens from 8 to 24 inches, 5 are ab- 

 normal 22= 



Of 9 specimens from 24 inches to ' large,' 2 are 



abnormal 24° 



Of 7 large specimens, only 1 abnormal 12^ 



The table requires a comment. The last 

 four groups are based, not on (9 + 19 + 9 + '*^) 

 44 specimens, but on only 29, for 15 turtles 

 were counted twice : the six 8-inch specimens 



^ Presumably typographical error. Gadow's 47 

 newborn plus my 28 embryos ^75. 



* Presumably typographical error. 70 intended. 



= Presumably typographical error. 26 intended. 



° Presumably typographical error. 22 intended. 



' Presumably typographical error. 14 intended. 



in both second and third groups; the two 24- 

 inch specimens in both third and fourth 

 groups ; and the seven ' large ' specimens in 

 both fourth and fifth groups. Twenty-nine 

 specimens divided into four groups, from 

 which a series of per cents, is computed to, 

 be compared with a per cent, based on a com- 

 paratively few newborn turtles — and this the 

 sole hasis for an elaborate hypothesis, given 

 to the scientific world with the supporting ( ?) 

 evidence and, subsequently, given to the gen- 

 eral puhlic, without the evidence, in a com- 

 prehensive monograph on ' Amphibia and 

 Reptiles.' This I regard as the ' sole basis,' 

 for though his comparison of the abnormali- 

 ties with supposed phylogenetic stages is in- 

 teresting and suggestive, and may support an 

 interpretation of the abnormalities as ata- 

 visms, it does not in the least imply that the 

 individual recapitulates these stages, and if 

 the latter assumption has other basis than the 

 table of percentages, what is it? 



The writer is now pursuing anatomical and 

 embryological studies, the results of which 

 may have some bearing on the interpretation 

 of the abnormalities in question, and these 

 results will be given out in due time. But 

 the question at present is not — Can Gadow's 

 assumption be disproved? but — Have there 

 been in hand facts to justify its promulga- 

 tion? Being promulgated, should it be in- 

 cluded without qualification in a compre- 

 hensive monograph intended for the general 

 public, who will not refer to the original 

 paper to find that it is merely a hypothetical 

 assumption frora a very small number of 

 facts ? The reference is to the ' Cambridge 

 Natural History,' Vol. VIII., 'Amphibia and 

 Eeptiles,' by Hans Gadow (London, 1901), 

 where the following unqualified statements 

 occur (the italics are mine) : 



It is absolutely certain that the number of 

 transverse rows also was originally much greater 

 than it is now. The mode of reduction of the 

 numbers of the neural and costal shields has been 

 studied in Thalassochelys caretta (cf. p. 388.) 

 The accompanying illustration (Fig. 68) [This is 

 a reproduction of text figure, from Willey's Zoo- 

 logical Results. R. E. C] shows some of the many 

 stages actually observed in the reduction of the 



