246 DISCOVERY REPORTS 



As the Discovery material contains many young stages I can complete the descrip- 

 tion given by me in 1921, for young stages were not then known. 



It is thus possible to show that the external shape of the valves varies much with the 

 age. We must, consequently, know the different stages of development, especially of 

 ScalpeUiim species, if we wish to bring order out of chaos. It is necessary in the future 

 to figure already described species which have been re-identified. 



The first stage (Fig. 9, a) in this collection is a pupa with the first primordial valves 

 developed. Of the other valves a very small upper latus can be traced. Broch, 1912 and 

 1924, states that the upper latus is first developed after the five primordial plates in 

 species of Scalpelliim. No peduncle scales are to be seen as in the corresponding stage 

 of S. gibberum, C. W. Aurivillius, 1892. 



In the second stage here figured (Fig. 9, b) all the valves of the capitulum are formed. 

 As no intermediate stages are represented we cannot indicate the order of the appearance 

 of the plates. We see that the rostrum is originally of a triangular shape with the umbo 

 at the base. On the peduncle the first four large scales are developed. The shape of the 

 inframedian latus is more rounded than in the later, especially mature, stages. The pro- 

 cesses, mentioned by Hoek and other authors in young Scalpelliim at the apex of the 

 capitulum, are very distinct. 



In the third stage figured (Fig. 9, c) the valves of the capitulum are somewhat 

 larger. The new scales of the peduncle appear in the region between the capitulum 

 and the peduncle, according to Broch. The primordial plates as in the previous 

 stage. 



In the fourth stage figured (Fig. 9, d) the inframedian latus is more elongated by the 

 upward growth of the sides which forms a small accessory part above the strong umbo. 

 The scales on the peduncle are more developed in this stage. 



Of the fifth stage (Fig. 9, e) three plates are figured. -The inframedian latus 

 has now reached its definite shape, and in the carina the umbo is removed from 

 the apex. 



Finally, to show the variation in the fully grown stages, I have figured an old specimen 

 (Fig. g,f,g, h), for comparison with that previously figured by me (1921, Fig. 29, h). 

 In that stage all the plates are fully calcified, but separated by chitinous interspaces. 

 The limits in the chitin show the same shape as the original plates, as in the first stages 

 here figured. The calcified portions of the plates are, in some respects, diff'erent from 

 the individual figured by me in 1921. Thus the tergum has the basal margin incom- 

 pletely calcified. The other plates are smaller, and the interspaces wider, than in the 

 young stages here figured. This may be due to the fact that the capitulum and the upper 

 part of the peduncle are much swollen. 



By dissecting the specimens I found a large number of Cypris stages, as the first de- 

 velopment takes place in the capitulum (Nilsson-Cantell, 1921, pp. 101-5). The upper 

 part of the peduncle is much expanded and, in consequence of this, it is without scales. 

 Such an interspace is not to be found in younger specimens. Fig. 9, h from the rostral 

 side shows the degree of deformation. 



