No. 3-] THE MUSCULATURE OF CHITON. 599 



case of the median dorsal muscle, some fibres are continuous 

 under the overlapping edges of the shells. 



(4) Between the shells where they overlap, a thin cushion 

 of muscles covering dorsally the apophysis, and having there- 

 fore the same outline. The arrangement of muscles forming 

 this cushion is complicated: at the antero-lateral edge of the 

 apophysis are oblique fibres (Figs. 4, 6, c\) lying in the trans- 

 verse plane, which pass from the ventral shell (apophysis) 

 obliquely toward the median line, to be attached to the dorsal 

 shell. Further relations of these fibres to a large muscle in 

 the mantle will be described in connection with the muscles 

 of that organ. The fibres of this border of muscles at the 

 edge of the apophysis are gradually replaced by fibres in the 

 sagittal plane (Figs. 4, 6, C2), which are attached anteriorly to 

 the ventral shell and pass to a more posterior attachment to 

 the dorsal shell. Between the ventral shell and the posterior 

 ends of these fibres is a region occupied by a set of fibres in 

 the horizontal plane (Figs. 4, 6, c-^ that are also attached by 

 their anterior ends to the ventral shell and pass laterally and 

 posteriorly to be attached by their posterior ends to the dorsal 

 shell. 



I have endeavored by comparing measurements of several 

 specimens killed in a contracted state with those of animals 

 killed while extended, to determine the functions of the various 

 shell muscles, but have been unable to obtain any satisfactory 

 results because of the small size of the animals and the varia- 

 tions in the proportions of different individuals. The following 

 account is a suggestion of the possible mechanism of the shells, 

 based on observations on the positions of the muscles. When 

 the animal is contracted or extended, the relative lengths of 

 the lines on the parts covered by the shells must be very 

 much like those of the finger when bent or straight: the 

 median dorsal line is longest when the finger is bent, and at 

 the same time the median ventral line is shortest, while some 

 line between the two remains constant in length, whatever the 

 position of the finger, i.e., the line, probably, along which 

 there are the fewest wrinkles of the skin. If the finger 

 were covered by strips of paper shaped like the shells of 



