270 



count on several series of shells the v umber of ridges per millimeter or an arbitrarily 

 chosen part, in order to determine whether or not there is actually a greater number of 

 ridges per unit area on the shells from the middle bay. The portion of the shell chosen 

 was the first millimeter of the anterior median denticulated area, approximately in a 

 line with the lower edge of the anterior lobe, but definitely at right angles to the ridges. 

 The following curves (fig. 106) are based on 50 shells from each of three localities. 



o 



1- 



Mean Ltode 

 Crockett - 14. 5 14 

 Dumbarton- 15. 4 14 

 Goat Is. - l6. 6 15 



_L 



J_ 



_L 



_L 



Fig, 



n 13 It IS It n le if zo n zz 



Number of ridges per mm. 

 106. Graph showing number of ridges per mm. on shell of Teredo from three stations. 



These curves appear to indicate a very definite correlation between environ- 

 mental conditions and the number of ridges per unit area on the shell, the greatest 

 number of ridges being found on shells from the middle bay, and the least number on 

 shells from the upper bay, while those from the lower bay stand intermediate between 

 them. The exact retison for this correlation is not entirely clear. We would suggest 

 that it lies both in the absolute differences in temperature and salinity and in the range 

 of such differences. 



That environmental conditions play a prominent part in determining the de- 

 position of ridges there can be no doubt. We are often able to read in the sculpture 

 of a series of shells from a given locality the history of some definite alteration in the 

 life conditions, which has left its mark on a number of shells. As an example of this, 

 note in fig. 107 the unusually wide interspace by which the newest ridge is separated 



