384 Annals of the South African Museum. 



Uropod tapering, inner margin straight, outer evenly convex, setose 

 ranras very small, tipped with a spine and setules ; the presence or 

 absence of a second, concealed, ramus could not be determined owing 

 to the impossibility of removing the dirt. 



Length : <$ 5'5 mm., $ 6'25 mm. ; breadth, 2 - 5 mm., 9 3 mm. 



Colour: In spirit dirty pinkish, eyes reddish. 



Locality : Walker Point (near Knysna) NE. by N. ^ N., distant 

 7 miles. 47 fathoms. <$ $ , ovigerous $ an d juv. ; Knysna Heads 

 N., distant 10 miles. 52 fathoms. $ <$ , 9 ? and juv.; Cape Seal 

 W. by N. i W., distant 7 miles. 39 fathoms. 1 $ ; Agulhas Bank 

 (without more exact locality, depth or date). 1 <^ , 5 $ $. s.s. 

 " Pieter Faure." 11/10/00, 2/7/02 and 20/4/06. (S.A.M., Nos. A3863, 

 A4116, A4188 and A4189 respectively). 



FAMILY ASTACILLIDAE. 



For references see Barnard, Ann. S.A. Mus. vol. 10, pt. 7, p. 206, 

 1914, and add : 



1913. Eichardson, 2me Exp. Autarc. frau9- Isopodes, p. 14. 



It is with great regret that, on the vexed question of the number of 

 marsupial plates in the members of this family, I find myself in 

 opposition to Prof. Koehler. This author (1911, Bull. Inst. oc. 

 Monaco, No. 214) has stated that in all the species examined by him 

 the number of pairs of marsupial plates is constantly three. Amongst 

 these species was Astacilla longicornis (Sow.). 



A. longicornis is a common species at Plymouth, and is found both 

 on Hydroids and among the spines of Echinus esculentus. I have 

 recently examined nearly 50 female specimens of this species in various 

 stages and in every case I have been able to find four pairs of mar- 

 supial plates. In the immature uouovigerous stage the pair on the 1st 

 segment are quite as clearly defined as in the ovigerous stage. This 

 pair does not increase in size so much as the 2nd and 3rd pairs and 

 does not take any part in the formation of the actual brood chamber. 

 The two plates overlie the "vibratory plates " of the maxillipeds, and, 

 like these, evidently help to aerate the brood chamber. 



In the nonovigerous specimen of A. mediterranea mentioned below 

 all 4 pairs of plates are very distinctly seen. They have not reached 

 their full size yet, and the 1st pair embrace the vibratory plates 

 instead of overlapping them. 



In this latter specimen the lst-3rd pairs are very easy to observe 

 (without any dissection), because they are in an early stage of develop- 

 ment. They are more sac-like than the fully developed plates and 



