DECAPOD CRUSTACEA FROM MADAGASCAR. 915 



Mountains leading to the Mangoro Terrace. 



Ambatonharanana, near Ampasimpotsy. July, 1911. 



Mangoro Terrace. 



Imerimandrosa, nortbevn end of Lake Alaotra, altitude 

 800-900 metres. July 1911. 



Mountains leading directly to the Plateau. 

 Ambohidratrimo. June, 1911, 



South-ivestern Region. 



Andranolabo and Tongoroby, Onilaby River, Dist. of 

 Betsioky. October, 1911. 



Part I. — Family Potamonid^ (River-Crabs). 

 1. Description of the Material. 



Tbe River-Crabs of Mr. Methuen's collection are referred to 

 tbe following species : — 



Potamon [Potamon) madagascariense A. Milne-Edwards. 



,, ,, goudoti H. Milne-Edwards. 



,, ,, methueni, sp. n. 



Hydrothelphiisa agilis A. Milne- Edwards. 



As will be explained below, the specimens included under the 

 first of these names differ considerably among themselves and 

 represent at least three distinct forms which may possibly deserve 

 to i-ank as varieties or even as species. 



In the specific descriptions which follow, regard has been 

 given chiefly to the characters of the carapace, to the exclusion of 

 those of the appendages, since the latter seem to afford no useful 

 points for separating the species here dealt with. Attention may 

 perhaps be called to the difficulty of deciding whether any single 

 .specimen is to be I'egai-ded as adult. In the case of females the 

 broadening of the abdomen seems to afford a fairly definite 

 indication, and it is usually easy to sort the specimens into 

 " adults " and "immatui'e." In the males, however, the genital 

 appendages of the first and second abdominal somites are often 

 found apparently well developed in very small individuals, which 

 differ considerably in the outline of the carapace from larger 

 males. 



In the lists of specimens the numbers preceded by the letters 

 " B.M." refer to the British Museum Register of Crustacea. The 

 m.easurements given in the tables are expressed in millimetres 

 and were taken by means of a scale graduated to "5 mm. No 

 very high degree of accuracy can be claimed for them, but I 

 believe that they are fairly comparable among themselves ; it is 

 not in all cases certain that they can be directly compared with 



