1909.] NATURAL SCIENCES OF PHILADELPHIA. 303 



spinnerets and the endopodites form a "rudimentary cribellum" 

 (more strictly, colulus) that later disappears; on the fifth segment 

 the exopodites become the posterior spinnerets, and the endopodites 

 the median spinnerets. Wallstabe (1908) also found, in Agelena, that 

 the median spinnerets arise by splitting off from the posterior pair. 



I am able on embryos of Loxosceles to essentially confirm the account 

 of Jaworowski and to trace the colulus to its adult condition. Fig. 17, 

 PL XIII, represents that stage before reversion, when the maximum 

 number of abdominal appendages is present, and there are seen to be 

 six pairs of them, on segments 2 to 7 inclusive. The two most posterior 

 pairs are the smallest and are temporary, for in the next stage, figs. 

 18, 19, of the beginning reversion they have disappeared and only the 

 four anterior pairs persist. 3 Later, fig. 20, the third and fourth pairs 

 of these appendages become the largest. The process of reversion 

 brings them close together in the mid-line, and an interesting and 

 decisive condition is exhibited in the stage of fig. 21. The fourth pair 

 of appendages (P. Sp.) are approximated, and on the median side of 

 each is an elongated thickening of the hypodermis, the first indication 

 of the median spinnerets (M. Sp.). Between the third pair of append- 

 ages (A. Sp.) there is an undivided, so unpaired, thickening, the pri- 

 mordium of the colulus (Col.). My observations differ mainly from 

 those of Jaworowski in finding the colulus to be unpaired at the start, 

 and in finding that the colulus and the median spinnerets behind it 

 do not arise as saccular endopodites of the embryonic appendages, 

 but as thickenings immediately mesial to the appendages. This is 

 a rather important difference, for Jaworowski's description would 

 indicate a cleft condition of these appendages, such as is found in 

 Crustacea, while I do not find such a condition. In Loxosceles the 

 colulus and the median spinnerets would seem not to be parts of the 

 appendages, but immediately contiguous to them. Figs. 22-24 show 

 successive later stages, with lengthening of the three pairs of spin- 

 nerets and of the unpaired colulus, as well as the gradual extension 

 forward of the largest pair of spinning glands. 



The colulus, accordingly, arises as a hypodermal thickening between 

 the anterior spinnerets, and the median spinnerets from a pair of thick- 



3 In the stage of figs. 18 and 19 there are eleven well-marked abdominal seg- 

 ments exclusive of the caudal lobe. The largest number yet described for 

 araneads is twelve, which with the seven of the cephalothorax makes a total 

 of nineteen. The three most posterior of them together with the caudal lobe 

 constitute a projecting postabdomen, such as has been described for Pholcus 

 bv Claparede (1862) and Schimkewitsch (1887); this may be compared with 

 that of Evagrus, fig. 28, PI. IV. 



