26G DR. W. A. CUNNINGTOK OK 



the whole of the then known species of A?'gid'us, but without 

 achieving a great success. He was unfortunate in choosing for 

 the purpose certain characters which clearly depend upon the 

 state of contraction of the specimen and, in addition, are difficult 

 to distinguish with exactness. It is not easy, for example, to 

 determine how far the swimming-legs are covered by the 

 carapace, as this is influenced by the relative states of con- 

 traction, as well as by any chance displacement. Again, while 

 it is easy in extreme cases to point out forms in which the 

 carapace lobes overlap, or do not reach the abdomen, it is 

 virtually impossible to distinguish accurately those with " Cara- 

 pace lobes just reaching the base of the abdomen," Most keys 

 are open to the objection that they are purely artificial, and 

 Wilson, it is true, fully admits the artificiality in this case. 

 Nevertheless, a key in which in so many cases male and female 

 specimens of the same species are separated under merely 

 arbitrary standards, falls very short of what is to be desired. 



It is certainly a matter of great difficulty in this genus to 

 discover characters which satisfactorily lend themselves to the 

 purpose of key -making. Where male and female difier markedly 

 in external shape, as is sometimes the case, it seems perhaps 

 inevitable that the sexes should be ai-tificially separated ; but 

 this might be avoided to a greater degree than has been done by 

 Wilson. While objecting to the form of the key above men- 

 tioned, we feel that Wilson is to be commended for attempting 

 to tabulate the species ; and since he included all the then-known 

 forms, it at first seemed wise to incorporate in his key the seven 

 new species described in this paper. It was on going into the 

 matter with this end in view that the unsatisfactory features 

 already mentioned revealed themselves, and, in addition, certain 

 regrettable inaccuracies * which only added to the difficulties of 

 incorporation. Under these circumstances it was finally decided 

 to prepare a key to include only the African species ; and this is 

 the more desirable, since the one species of this genus hitherto 

 known from the continent is now increased to eight. 



Before proceeding to give the key and the descriptions of new 

 species, a few words of caution may be offered. The characters 

 employed in diagnosing the forms belonging to this group, while 

 perhaps the most obvious, are in some cases profoundly influenced 

 by artificial contractions and contortions, and in addition by 

 the age of the individual. Proportions of length to breadth 

 are seldom quite constant, size of suckers may vary with degree 

 of expansion, and certain spines and processes may be more 



* Op. cit. p. 702. On this page, for example, under the heading 6., A. niger is 

 said to have an " abdomen narrow, ovate." Under the heading 7., A. pwgeitensis is 

 contrasted with this as having "abdomen wide, elliptical." Actual measurements 

 of both these species are given in the body of the work, and may be written l)readth 



over length as follows : — niger — ^-, pugettensis — . Expressing these in the form of 



percentage of breadth to length, they become -.—niger 81"2, ptigettensis 75"4. Thus 

 it is evident that the abdomen of niger, described as "narrow," is iu reality broader 

 than that oi' puyettensis, described as " wide.' 



