49 



moving of stones by heavy seas. The crabs must therefore be 

 constantly exposed to the risk of having their Umbs crushed, 

 and so of losing much blood. If autotqmy at the breaking plane 

 occurs at once, practically no haemorrhage follows, for a very 

 effective venous valve comes into operation at the moment of 

 amputation. This has been described by the writer f in another 

 place. Self -amputation of Hmbs in the Brachj^ura is therefore 

 more a device for saving blood than for evading an enemy. 



Fredericq's explanation of autotomy has been given above, 

 but it has been found that it does not hold for all positions of the 

 limb. For example, if the leg be forcibly extended, it cannot be 

 cast off, and then, according to Fredericq's theory, it ought to be 

 most easily autotomised. 



A few words on structure are necessary in advancing the 

 present explanation of self -amputation in the Brachyura. The 

 form of the basal limb segments is almost uniform for the various 

 " true crabs." In all cases the second segment or basi-ischium 

 is divided by a furrow (the breaking-plane) into two parts. It is 

 connected by muscles to the first segment and to the body w^all, 

 the short extensor and the short flexor moving it on the first 

 segment, and the long extensor and long flexor joining it to the 

 body. The bodies of these last-named muscles pass through 

 the cavity of the first segment. Figs. 11 and 12 show this arrange- 

 ment diagrammatically from the lateral aspect, and looking through 

 the Hmb cavities, respectively. 



The tendon attachments of the extensor muscles are of great 

 importance in this description of the autotomy process, and they 

 must be described in some detail. In the upper division of 

 arthrodial membrane of the articulation between the first and second 

 segments, the tendons may be seen shining through in the region 

 of their attachment to the lower edge of the basi-ischium (fig. 13). 

 The long extensor is atttached anteriorly and the short posteriorly. 

 Between the attachments a knife can be passed upwards to the 

 breaking-plane through tough material, which is not brittle and 

 calcareous like the rest of the basium. This passage made by the 

 knife is really through a slanting joint in the hard ring of integu- 

 ment forming the basium. It is shown in section in fig. 14. 



The action of the short and long extensors is in directions 

 shown in fig. 15, and when the muscles contract together, they 



