ATLANT. DEEP-SEA EXPED. 1910. VOL. Ill] 



PTEROPODA. 



33 



as well as the main course of the intestine, which is 

 however considerably longer than that of a typical Lima- 

 cina; it turns round the right edge of the liver, forming 

 a long ventral loop, and then twice crossing the dorsal 

 side, first in a direction from right to left, and then back 

 again to the right side, where the anus is found. During 

 its last dorsal passage the intestine crosses the stomach. 



Taking the intestine of Peraclis as a point of depar- 

 ture the digestive tract seems to have followed two 

 diverging lines of development, both resulting in a 

 considerable shortening of the intestine. 



Such a gradual reduction in the length of the intestine 

 is seen in the genus Limacina. In Limacina helicoides 

 (textfig. 27, 2) we still find the same course of the intestine 

 as in Peraclis, with two ventral loops and a double crossing 

 of the dorsal side, but all these parts of the intestine are 

 shorter than in Peraclis, and this gradually leads to the 

 conditions found in the typical species of Limacina (textfig. 

 27, 3), where the intestine is reduced to a short dorsal loop. 



A corresponding reduction in the length of the intestine 

 is found also in the C a v o 1 i n i i d a e (textfig. 27, II — IV), 

 but here it is combined with other changes, which result 

 in a state of the digestive tract directly the reverse of 

 that found in a typical Limacina. 



In Clio falcata (II), where the pallial cavity, unlike 

 that of Peraclis, occupies the right and ventral side of 

 the body, while the retractor muscle passes along its 

 dorsal side, we also find a change in the position of the 

 digestive tract which most probably should be viewed in 

 connection with that of the pallial cavity. Through a spiral 

 twisting of the oesophagus the stomach is turned 

 round, so that its more convex (originally dorsal) side 

 carrying the unpaired tooth is turned towards the (ventral) 

 pallial cavity, while the opposite, shorter (originally ventral) 

 wall of the stomach lies beneath the retractor muscle. At 

 the same time the whole stomach is displaced from the 

 median line of the body towards the left side. 



The long intestine has still kept essentially the course 

 characteristic of Peraclis. From the stomach it sweeps 

 towards the right but does not reach the right edge of 

 the broad and flattened body, and after having made a 

 long loop all over the ventral side it crosses, just as in 

 Peraclis, the dorsal side of the stomach in a direction 

 from left to right. Instead, however, of opening as in 

 Peraclis at the right side of the stomach, which would 

 in Clio falcata be the median dorsal line of the body, 

 the intestine once more turns back to the left side, where 

 it reaches the edge of the pallial cavity. 



The condition of the digestive tract in Clio falcata 

 is of great general interest, as showing the old inherited 

 structures in process of adaptation to the new conditions 

 called forth by the displacement of the pallial cavity. 



A further step in the development of the intestine in 

 the Cavoliniidae is represented in Clio cuspidata (textfig. 

 27, III). The oesophagus is shorter than that of Clio falcata, 

 but in its course traces of spiral twisting may still some- 

 times be seen. The unpaired tooth occupies the same 

 position within the stomach, and the whole change of the 

 intestine consists of a reduction of all superfluous loops. 

 Instead of making the first turn towards the right side, 

 the intestine now leaves the stomach towards the left, 

 the ventral loop is shorter, so that it does not reach the 

 right margin of the body, and the last dorsal loop (across 

 the stomach) is omitted, the rectum running directly 

 along the left side of the stomach up to the anus. The 

 stomach in Clio cuspidata occupies a more median and 

 more ventral position than in Clio falcata. 



The digestive tract as found typically in the Cavo- 

 liniidae is seen in Clio pyramidata (textfig. 27, IV), the 

 stomach being found ventrally in the median line, the 

 whole intestine consisting of a short loop like that in 

 Clio cuspidata and exactly the reverse of that in a typical 

 Limacina. Without knowing the series of stages connect- 

 ing Limacina balea and Clio pyramidata with the archaic 

 genus Peraclis, a development from one type to the other 

 through a rotation of the whole digestive tract would 

 seem to be, and has been accepted as, the most plausible 

 hypothesis regarding their phylogeny. 



Nervous system. The visceral and abdominal ganglia 

 have proved to be of special interest from the point of 

 view of comparative anatomy, in so far as they are 

 mutually combined in different ways in each of the three 

 great families of thecosomatous pteropods. While in the 

 Cymbuliidae three distinct ganglia are found (see 

 textfig. 28 B), in the typical representatives of the two other 

 families the visceral mass is asymmetrically developed, the 

 right side in Limacina (C, D'), and the left side in the 

 Cavoliniidae (E", F", D'"), being the largest. 



It has been shown also (Pelseneer, Tesch, Meisen- 

 heimer) that with regard to the nervous system the genus 

 Peraclis (textfig. 28 A) agrees more closely with the 

 Cymbuliidae than with the typical species of Limacina. 

 The asymmetry of the visceral mass in Limacina and in 

 the Cavoliniidae has been regarded as the results of a 

 coalescence of the abdominal ganglion with one or other 

 of the visceral ganglia, but no plausible explanation of 

 this asymmetry, or of the difference between the two 

 groups, has yet been given. 



On this point also the two species Limacina helicoi- 

 des (B ) and Clio falcata (B") form links connecting 

 Peraclis on the one hand with the typical species of 

 Limacina and the Cavoliniidae on the other (as shown 

 in textfig. 28). 



The visceral mass of Peraclis (A) consists of three 



