SCHIZOPODA. 19 



The rostrum is figured correctly by Sars from a specimen, 13 mm. in length. 

 At that .stage in development the rostrum is an acutely triangular projection, the 

 sides of the triangle very nearly straight and but little concave. As growth proceeds, 

 however, the sides of the triangular plate become deeply concave and the rostrum 

 assumes the form of a long narrow acutely spiniform projection, as shown in Fig. 2. 

 This gives it something of the form as seen in T. gregaria, but it is much longer and 

 more slender than in that species. 



Sars describes and figures the preanal spine in his 13 mm. specimen as armed 

 with only two teeth. This is correctly stated for a specimen that size, but increase 

 in size is also accompanied by an increase in the number of teeth in the preanal spine, 

 till in a specimen 28 mm. in length I found the teeth to number twelve in addition 

 to the strong external one. The teeth are, moreover, fine and comb-like, and the 

 whole spine closely resembles that figured by Sars for T. gregaria. I have noticed 

 all intermediate stages in the present material. This clearly undermines the value of 

 the preanal spine as a specific character, since the numljcr of teeth is dependent first of 

 all upon age, while individual variation must also be taken considerably into account. 



The most considerable changes due to growth are, however, exhibited by the 

 second thoracic limbs. Sars describes them as much smaller than in T. gregaria, 

 with the meral joint scarcely reaching beyond the middle of the antennal scale. 

 But the above description and accompanying figures show that the limb is 

 quite as well developed as in T. gregaria, and that the meral joint in full-grown 

 individuals actually extends beyond the autennular peduncle. Stebbing (1900) 

 has already called attention to the fact that in specimens of this species from 

 the Falkland Islands the elongate limbs were proportionately longer than in Sars' 

 drawing, but he gives no figures. I find that this statement applies generally to the 

 material in the present collection, even in specimens of the same size as Sars'. Only 

 two exceptions were noted, and in these two specimens the elongate limits were 

 proportionally as in Sars' figures. I will deal with these two exceptions and Sars' 

 specimen below, but will proceed first to briefly note the changes which take place in 

 these limbs during growth. In two specimens, 8 and 9 mm. in length, the second 

 thoracic limbs are developed about as fully as in Sars' figures. They are considerably 

 less than one-half of the total body-length of the specimens, the meral joint extends 

 very little beyond the centre of the antennal scale, and is about equal to the terminal 

 three joints combined, while the carpal joint is only about twice as long as the 

 propodal. 



In specimens from 18 mm. to 18 mm. in length, these limbs are about equal in 

 length to the half of the length of the body, the meral joint extends to the tip of the 

 antennular peduncle, and is a little shorter than the terminal three joints combined, 

 while the carpal joint is now three times as long as the propodal. This is the 

 condition noted by Stebbing in his Falkland Lshuids specimen.s. In the full-grown 

 specimen, 28 mm., the elongate limbs measure 21 mm. in length, or three-quarters of 



