November 17, 1905.] 



SCIENCE. 



639 



15 costals and neurals. These two specimens 

 are really very interesting, since they fill the 

 gap in my table, between totals of 16 and 14. 

 This reduction is due to the falling out of a 

 pair of costals, as is shown by comparison of 

 Thalassochelys with Ghelone. This shortage 

 of one costal scute in Coker's turtlets is, how- 

 ever, not ' precocious,' since 16 is the normal 

 number in this species, but it is rather 

 prophetic so far as Cheloniidas are concerned. 

 Therefore, no need for him to say that it 

 would be difficult to explain this deficiency 

 as precocity. But I am at a loss to under- 

 stand what bearing his table on page 20 has 

 upon my hypothesis, considering that all his 

 turtlets were embryos, most of them still 

 unripe. I trust that he does not impute to 

 me the belief that the turtlets should begin to 

 mend their ways before they are born. 



The fact remains that, with the addition of 

 Coker's turtlets to the 76 specimens enumer- 

 ated in my table, the percentage of specimens 

 with supernumerary scutes still decreases 

 steadily with age. Of course, such calcula- 

 tions are always precarious unless they are 

 based upon very great numbers, and it is not 

 unlikely that the amount of variation in the 

 members of one brood differs much according 

 to clans, or regions, not to mention the pa- 

 rents about whom we know nothing. 



Per cent. 

 Of 73 embryos or new born, 53 are abnormal = 7 



Of 9 specimens from 3 to 8 inches inch, 3 abnormal =33 

 Of 19 specimens from 8 to 24 inches incL, 5 abnormal = 22 

 Of 9 specimens from 24 inches to "large," 2 abnormal =24 

 Of 7 large specimens, only 1 abnormal =12 



Still more surprising are Mr. Coker's con- 

 clusions as drawn from the examination of 

 250 specimens of the diamond-back terrapin, 

 Malacoclemmys centrata Latr. Examination 

 of 250 individuals of allages and sizes in the 

 condition just as they were caught. He 

 found, however, a few cases which indicate 

 that partial fusion, or division, of neighboring 

 scutes has been going on during the creature's 

 life. The occurrence of reduction by fusion, 

 of increase by division during the individual's 

 growth is, therefore, demonstrated. Would it 

 not go a long way in support of my hypothesis 

 if actual observation showed that the total 

 number can be reduced during life by one 



single scute? I ought to feel grateful for 

 such a help, willing to waive the experimental 

 proof, and be glad to accept fusion instead of 

 my suggested squeezing out. Lastly, I ought 

 to feel crushed by Mr. Coker's table, which 

 shows, if anything at all, that in the diamond- 

 back the scutes seem to increase instead of 

 decreasing with age. The author himself is 

 confirmed in the ' belief that in Malacoclem- 

 mys we do not have orthogenetic variation in 

 the sense that there is normally in the indi- 

 vidual life history a progressive reduction in 

 the number of scutes.' 



No, of course not. The whole investigation 

 has ended in a farce. A beast which fre- 

 quently goes in for the longitudinal splitting 

 of its true neural scutes is not a fit subject 

 for the study of orthogenetic variation. This 

 is malformation pure and simple ; it is as little 

 atavistic, or in line with anything, as are six 

 fingers in man, or five toes in fowls. 



I do not profess to know what the diamond- 

 backs are after. Perhaps this crazy splitting 

 may in time establish a new type of chelonian 

 with a completely double series of neural 

 scutes! However, Coker himself did not fail 

 to see that in these creatures ' the tendency is 

 not toward, but away from the type.' He 

 should, therefore, have ruled them out of 

 court. But there surely is also a teyond the 

 type, and this seems to have puzzled him 

 sorely, as is shown by his footnote 3 on page 

 18 of his paper. For instance, if (whether 

 congenital, or due to fusion, division or sup- 

 pression) a loggerhead or other marine turtle 

 has only five neurals, that would be something 

 beyond the type of Cheloniidse, and in this 

 respect very abnormal, but still orthogenetic, 

 since reduction to five neurals has been real- 

 ized and has become the normal condition in 

 Pleurodires and even in certain species of 

 Testudo. Such a case in a marine turtle 

 would be, so to speak, prophetic, and quite 

 reasonable, and a trained morphologist would 

 recognize that such things are 'quite in the 

 line ' of tortoises. 



My appeal to comparative anatomy and 

 common sense has been of little avail. I 

 venture to make a practical suggestion. The 

 eastern states are the home of many kinds of 



