218 ORTHOGENETIC VARIATION IN THE SHELLS OF CHELONIA. 



Stage VII. Six neurals and five pairs of costals owing to fusion of the two last 

 pairs of costals into one, or perhaps by suppression of one pair as indicated by numbers 

 26 and 28. This is the normal condition of Thalassochelys, and further reduction of 

 scutes in this species was unknown until Dr Oudemans sent me a drawing of specimen 

 No. 196 (Fig. 26), which but for the vestigial last but one neural would be inter- 

 mediate between stages VII and VIII. 



But in other Genera and families the reduction goes much further, leading to 

 the following stages. 



Stage VIII. Six neurals and only four pairs of costal scutes. This condition is 

 typical of the two species of the genus Chelone, Ch. mydas and Ch. imbricuta. Now 

 behold, the costal scutes fit everywhere into the corners which are formed by two 

 successive median scutes, except between the first and second, which here for the first 

 time come together without any interference of a pair of costals. This condition results 

 from the suppression of the first pair of costals in comparison with other Turtles. 

 It is however a remarkable fact that hitherto no atavistic variations in the young 

 of Chelone imbricata have been observed. This species seems to be, so far as the 

 development of its dorsal scutes is concerned, quite cenogenetic 1 . 



Stage IX. Six neurals and only four pairs of costal scutes, but the first median 

 scute, the so-called neural, has become very small, the resulting gap being filled up 

 by an enlargement of the first pair of marginals. Observe also in Testudo, e.g. T. 

 maaritanica, the lateral angles of the second median scute, which frequently are still 

 persistent, although there is no longer a pair of costal scutes to fit into. This stage 

 or condition is the normal one in most cryptoderous tortoises. Supernumerary scutes 

 occur occasionally, for instance, in Testudo and Chrysemys. 



Stage X. The nuchal shield has disappeared, and there is either left a little gap, 

 or this is closed by the first pair of marginals (X a ). The whole carapace is composed 

 of five neural and four pairs of costal scutes, in all only 13 scutes, omitting of course 

 the marginals. This condition occurs specifically and even individually in the Genus 

 Testudo. It is normal also in pleuroderous tortoises, but in some of these, for instance in 

 Sternothaerus, the number of five separate median scutes seems to be due not to sup- 

 pression but to fusion of the original first or nuchal with the next following scute (X b ). 

 Consequently this condition would represent a side departure or separate stage, some- 

 what intermediate between the Vlllth and last stage. 



Beyond this Xth stage Chelonians have not yet ventured, at least not normally 2 

 and it is not our business to enquire what they perhaps may do in the future, but 

 we can, by the help of comparative anatomy, reconstruct to a certain extent, the ancestral 

 type. 



1 " Supernumerary " scutes seem to be very rare in the genus Chelone. I did not know of a single 

 abnormal specimen until Dr Willey found one of Ch. mydas (in the Manchester Museum) with 7 neurals, 

 5 left and 4 right costals. 



2 The greatest, although quite abnormal, reduction I am acquainted with occurs in a specimen of the 

 now extinct Testudo indica (figured by Perrault, Mem. pour servir a Vhistoire des animaux et des plantes, 

 Amsterdam, 1736; cf. also Gadow, "On the remains of gigantic Land-Tortoises, and of an extinct Lizard, 

 recently discovered in Mauritius," Transact. Zool. Soc. xm. pt. vm. 1894). This specimen has no nuchal 

 and only 4 other median or neural scutes. 



