466 LESLIE B. AEEY 



tain of the mantle cells exhibit conical elevations, which project 

 far above their expanded bases. Their apices bear several fine, 

 non-motile bristles. These cells have been designated as 'sen- 

 sory,' on the basis of suggestive structure and selective staina- 

 bility with methylene blue (Lillie, '95). .Actual continuity or 

 any definite coordinating intermediary between these cells and 

 the muscle fibers, although generally assumed, has never been 

 demonstrated. There are four pairs of such hair cells, symmetri- 

 cally apposed in the two valves and grouped as outer and inner 

 pairs (figs. 3 and 4). The two outer set sare located about one- 

 third distant from the ventral shell border; they are prominently 

 elevated and can easily be observed in properly illuminated, 

 living glochidia. The two inner pairs are shorter, and are in- 

 conspicuous except in stained preparations; the anterior pair is 

 placed much nearer the hinge line than the posterior set. 



In a few mussels (Anodonta, Unio, Quadrula plicata, Qua- 

 drula heros, Strophitus) there is an interesting structure, the 

 larval thread, formed by an elongated thread gland. Once 

 believed to be a byssus, and still so designated in many texts, 

 it is in fact an organ peculiar to certain glochidia, and is in no 

 way homologous to the later byssus of the juvenile mussel, formed 

 at metamorphosis. Lillie ('95) has interpreted the thread as a 

 filamentous excretory mass. Shierholz ('88) and others have 

 considered it of use in tangling the glochidia into masses by which 

 many glochidia are drawn into contact with a fish after a single 

 one attaches. That this may be an imperfect explanation is 

 suggested by the observation of Lefevre and Curtis ('12) that the 

 threads of Anodonta dissolve within a day or two after the 

 glochidia are free. It is also reported by the same authors that 

 the glochidial thread of Unio complanatus is extruded immedi- 

 ately after the larvae are removed from the marsupium, and, 

 according to Harms ('07), in Margaritana margaritifera it is 

 lost at a premature stage while the glochidium is still within the 

 egg capsule. It would seem to be of no especial importance 

 mechanically in aiding attachment, but may well be indirectly 

 useful in another way (p. 468). 



