20 BULLETIN OF THE BUREAU OF FISHERIES. 



this sponge we find a conspicuous system of subdermal cavities, an oscular tube (near 

 one end), and radially projecting spicules. Embedded in the sponge is a considerable 

 fragment of the skeleton of the parent (near pointed end). 



Skeletal fragments derived from the parent were present in several but not in all of 

 the masses of tissue which metamorphosed completely. The presence of such a frag- 

 ment is then not a necessary condition to complete metamorphosis. Nevertheless, the 

 impression from numerous experiments is made on my mind that where the mass of 

 tissue is of some size and does not flatten out into an incrustation the chance of metamor- 

 phosis is increased if there is present a piece of the original skeletal framework. This 

 may possibly act as a mechanical support. 



Experiment record 5, July ig, 1907. — Result: Tissue teased out of fresh normal 

 sponge quickly attaches to substratum and spreads out as thin sheet. Will such sheets 

 metamorphose, without passing through condition of compact gemmule-like mass? 

 Indications are that they will. 



In the preceding experiment when the bulk of the tissue was gathered into center 

 of watch glass, a large number of small masses remained adherent to the bottom. These 

 soon flattened out into thin, irregular plasmodia which were watched for an hour, and 

 were observed to change shape and establish connection with one another. One such 

 Plasmodium is shown in figure 25. 



These plasmodia were kept under observation and it was found that they did not 

 contract into compact masses but spread as thin sheets over the bottom of the glass. 

 In this condition they were removed to the live box in the hope that they would metamor- 

 phose. The plasmodia on July 21 exhibited a slight but significant change. They 

 were no longer homogeneous in appearance, for in many places a distinct surface mem- 

 brane existed which was separated from the opaque general mass by a space filled with 

 branched cells and colorless transparent matrix. In other words collenchyma had 

 made its appearance. In a day or two some of the plasmodia had what appeared to be 

 flagellated chambers and the beginnings of canals. The conditions in the live box were 

 not good. Sediment was deposited in the watch glasses and the plasmodia did not 

 develop further, eventually dying. 



Before disappearing they diminished in size, and became once more homogeneous 

 in appearance (a regressive series of changes). Some of them sent up solid massive 

 processes into the water, as if growing away from the sediment. Others contracted 

 again into minute compact rounded masses. Some of these lose their compact character 

 and break up into separate cells, for the most part spheroidal, that are loosely held 

 together. 



Experiment record 6, July 20, 1907. — Question involved: Will smooth rounded 

 gemmulelike masses formed by teased-out sponge tissue transform in laboratory aquaria ? 

 They did not. 



In the preceding two experiments the contrast in behavior between tissue masses 

 which while small quickly made firm attachment to the glass, and such as did not attach 

 to the glass or were prevented from doing so, was marked. The former spread over 



