ANATOMY AND GENERIC CHAEACTERS OF SCORPIONS. 377 



The genera adopted by Peters appear to me to be, in most cases, unnecessary, often 

 not even justitiable as subgenera. He has at the same time rendered great service by 

 pointing out the confusions which have arisen in the use of the generic terms of one 

 author by another, in new and unjustifiable senses. 



Thorcll has added a number of genera to the akeady superfluous list, and has 

 modified Peters's classification in what appears to me to be a retrograde spirit. He 

 recognizes four families of Scorpions, viz.: — (1) the Androctonidae, corresponding to 

 Peters's Androctonini and Centrurini combined ; (2) the Telegonidse, identical with 

 Peters's Telegonini ; (3) the Vejovidse; and (4) the Pandinoidffi, the last two resulting 

 from the breaking up of Peters's Scorpioniui on no assigned grounds. 



Both Peters and Thorell make use of the presence or absence of a keel on the 6 th 

 metasomatic segment as a means of generic distinction, and of other characters even 

 more trivial. The small value of such characters is shown by the fact that the common 

 American Scorpion, the Scorjno americauus of De Geer, is sometimes provided with a 

 spine below the sting, and sometimes has none. 



In order to appreciate more clearly Peters's four groups of Scorpions, we may refer 

 to four types which are figured in the Plates accompanying this memoir, viz. for the 

 Telegonini the Telegonus of Tasmania, PI. LXXXII. figs. 5, 8, and PI. LXXXIII. 

 figs. 5, 6 ; for the Scorpionini the Scorpio cyaneus and S. Kocld of Ceylon, PI. LXXXII. 

 figs. 1, 10, 19, and PI. LXXXIII. figs. 9, 10 ; for the Centrurini the Androctonus 

 americanus, PI. LXXXII. figs. 6, 13, 17, and PI. LXXXIII. figs. 3, 4; for the 

 Androctonini the Androctonus funestus of North Africa, PI. LXXXII. figs. 2, 14, 15, 

 and PI. LXXXIII. figs. 1,2. 



My observations, which relate not only to the characters made use of by Peters, but 

 also to two points of internal structure, viz. [a) the disposition of the segmental ganglia 

 and their great nerves and (Jj) the sculpturing of the lamellae of the lung-books, have 

 led me to the conclusion that the existing species of Scorpions should be grouped in 

 two and not in four primary divisions ; the first group, or Scorpionini, corresponding to 

 Peters's Telegonini and Scorpionini combined, whilst the second group, the Androcto- 

 nini, correspond to his Centrurini and Androctonini combined. 



It appears that the linear compressed sternum of the Telegonini may be regarded as 

 only an extreme form of the broad pentagonal sternum of the Scorpionini. In both 

 series there is but a single row of teeth in each ramus of the chelicera, except a single 

 tooth of a second row on the movable ramus in some species of Telegomts (PL LXXXIII. 

 fig. 7). But what is of far more importance is that in both Telegonini and Scorpionini 

 the ganglia of the nerve-cord and their oflf-springing nerves are arranged as shown in 

 the woodcut, fig. 2, B, whereas in the Scorpions of Peters's groups Centrurini and 

 Androctonini these structures have the arrangement shown in drawing, fig. 2, A, C, D 

 (p. 378). This difference may be described by saying that in the Scorpionini (inch 

 Telegonini) only the region of the first pair of lung-books is innervated from the 



3k2 



