MEANS OF DISPERSAL 179 



locomotive, have ceased to pass through a free larval 

 stage. But mere fixity of tenure is not enough, i.e., it 

 will not explain, for instance, the distribution of the 

 same freshwater animals all over the British Isles, or in 

 the case of many freshwater genera pretty well over the 

 whole world. Free-sw T imming larvae provide a means of 

 distribution, and if this be lost some other must be found 

 to take its place. Let us see how Nature has solved this 

 problem. We may turn first to the remarkable case of 

 our freshwater mussels, Unio and Anodon. The ova of 

 these animals develop up to a certain stage within the 



FIG. 52. Larval form (glochidium] of the 

 Freshwater Pond Mussel, (x about 120.) 



gill-pouches of their parent ; they then somewhat re- 

 semble the spat of a sea mussel, in so far as they are 

 provided with a small shell consisting of two valves ; but 

 they differ greatly in other respects, more especially by 

 the production of the free edge of the valves into sharp, 

 recurved hooks, and the possession of a long adhesive 

 filament, technically known as the byssal filament (Fig. 

 52). They remain in the gill-pouch till some passing fish 

 comes near, when they are discharged into the water, and 

 attach themselves first by the byssal filament and then 

 by the sharp teeth of the shell to its fins or tail. An 



