234 NV M- s - MARSHALL AND C. T. VORH1ES. 



not the brownish sand from the springs in which the larvre lived, 

 but crushed red sandstone. When a larva rebuilt the part of 

 the shell we had removed it was easy to see just how much new 

 material had been added. 



The construction of a new case is easily observed by removing 

 larvre from their cases and placing them in aquaria at the bottom 

 of which is some sand. The larva.' appear at first much dis- 

 tressed and wander around in the aquaria attacking each other, 

 if a number are in the same dish, but never, so far as we have 

 observed, injuring one another. From the fact that the caseless 

 larva- always attack each other it is much better to place but a 

 single one in an aquarium. Many will remain for one or two 

 days without attempting the construction of a new case while 

 others will begin building within an hour, or in a few instances, 

 in even less time. When a larva begins the construction of a 

 new case it first, with its mandibles, gathers a few of the larger 

 sand grains and cements them together with the secretion from 

 the spinning glands. This done, the larva gathers other grains 

 from a distance carrying them in its mandibles to the pile already 

 formed, and, as each is brought, it is cemented to the others. 

 In this way, in the course of from two to four hours, a large pile 

 of sand grains is made, each grain of which is cemented to the 

 others, forming all together a loose mass nearly as large as the 

 larva itself. It is now somewhat surprising to notice that the 

 larva, in beginning to construct its shell, does not, apparently, 

 use this pile at all but begins its case at one side and adjacent to 

 it, and there, from smaller grains than those in the pile, the new 

 case is built. The first process in the actual construction we 

 did not see but the larva soon has a narrow band of case around 

 its body and to this, at the anterior end, new grains are continu- 

 ally added. As the new case increases in length, the diameter 

 is slightly enlarged although the new cases do not show as much 

 difference in width at the two ends as did the old shell. When 

 the case has reached the necessarv length the larva turns in it 



* o 



and begins to pull down the posterior margin, finally making the 

 opening at this end smaller than the diameter of the shell. The 

 larva again assumes its original position and finally builds a 

 slight dorsal hood at the anterior end. During the construction 



