582 BULLETIN OF THE BUREAU OF FISHERIES. 
vicinity of the mouths of rivers must be avoided. The water should at all times 
be of a saltness not much below that of the open sea, and no locality where it 
frequently falls below a specific gravity of 1.019 or 1.020, reduced to a standard 
temperature of 60° F., can be regarded as safe. Disastrous experience on the 
plantations of the Bureau and of Messrs. Cheyney and Bigelow at Anclote Key, 
where the density, especially near the surface, fell below 1.018 for a consider- 
able period and killed the majority of the sponges, emphasizes this considera- 
tion. Along the keys below Lower Matecumbe or between the upper end of 
Key Largo and Cape Florida this trouble is not likely to be experienced, but 
there are few places near the mainland opposite the keys or in the Gulf of 
Mexico, where localities close to shore can be occupied without apprehension 
from this cause. Several years may intervene between successive periods of 
fresh water, and the fact that it does not occur for one or two years does not 
guarantee that a plantation will not experience its effects just when the planter 
is beginning to hope to reap the results of his labor. With this consideration 
firmly in view, there is no great difficulty in avoiding most of the disasters 
which have affected the experimental work. 
The experiments with wires and lines for the suspension of the cuttings 
above the bottoms having not reached favorable termination, the use of soft 
muddy or densely grassy bottom can not be recommended, and the sponge 
grower is, therefore, practically restricted in his choice to such bottom as is 
more or less capable of supporting a natural growth of sponges. Disks and 
triangles tend to sink on mud marl or shifting sand, or are overgrown when 
deposited among dense marine vegetation, in any of which contingencies the 
cuttings are either killed or injured. Even when the sponge is but partially 
buried in sand the basal portion will die from suffocation, and the still living 
upper parts will eventually lose attachment and be carried away. But a sparse 
growth of grass may prove highly beneficial in stimulating a more rapid growth 
of the sponges. This appears to be indicated by very recent experiments, and 
if it should prove true a large area of virgin bottom will be opened to produc- 
tiveness. Otherwise, the selection of a sponge farm is practically limited to 
rocky bottom, though there are occasional localities where a mud bottom com- 
paratively free from vegetation is sufficiently firm to support disks or triangles. 
Care must be exercised also to avoid places where there is much sand car- 
ried in suspension in the water, as this will be deposited in eddies created by 
the sponges and their supports, and will gradually build up around the plants. 
This difficulty can be avoided on otherwise suitable bottom by raising the cut- 
tings on spindles, so as to allow the currents to have a scouring action below 
the growing sponges. For attachments it is recommended that the disks 
described in a previous section of this paper be employed. The plain disk per- 
forated by two holes is the simplest and cheapest form of attachment, but its 
