Aquarium tanks are of all sizes and shapes, from the small fish-globe to the plate-glass tank, whose dimensions are measured by yards and whose contents are hundreds of gallons. In such tanks veritable whales have been kept in good health. On the other hand we have formed a microscopic aquarium out of a homeopathic phial, and in it have kept minute plants and animals for months in good condition.

Probably the most pleasant and useful size for an aquarium is about thirty inches for the length and fifteen each for width and depth. Such a tank is easily manageable, while at the same time it admits of a fine display of plants and rockwork, and allows abundant room for the fish as well as nice resting-places for the amphibious animals. For ourselves, we confess that we have a liking for large tanks, - larger, even, than that just described. In such tanks we are able to watch the natural growth and development of most ordinary fish; the plants that are introduced need not be mere dwarfs; and the large body of water which they contain is not subject to such sudden and violent changes of temperature, unless exposed to the direct rays of the sun, - a condition which should never be allowed. But when the main tank is of a large size, it will in general be found necessary to have a few small ones for the more minute specimens, which would otherwise be difficult to find in the large tank.

Avoid globes and all tanks with curved surfaces, as they give a distorted view of the animals, and when large are easily broken by any tap or increase of pressure from within. It is true that for scientific purposes bottles of all kinds, and even test-tubes, may be used; and on one occasion, where we required a large number of vessels, we made good use of a lot of two-quart fruit-jars. But these are makeshifts, and not very good ones at that. Even hexagonal and octagonal vessels, although they are peculiar and somewhat pretty, we dislike, because the field of view (if we may so express it) is very limited. The fish, in moving about quickly, get behind another plane, and then the distortion is horrible. Now, what we want to secure is a clear and unobstructed view of all parts of the tank, so that the movements of the fish and the relations of the plants may be clearly and constantly visible when we wish them to be so. Nothing meets this requirement so thoroughly as a four-sided tank made of good plate-glass.

It is probable that most fish and other animals would prefer opaque sides, as more closely imitating a natural pond; and acting on this idea, some makers have constructed their tanks with backs made of slate. The idea is a good one for some purposes, such as experiments in fish-breeding, but the plan is unsuited to the wants of the naturalist. If an opaque back is thought to be advantageous, just hang a black cloth behind the glass: this can be removed when a view from that side is needed.

Most tanks are made with cast-iron frames, into which the glass is cemented; and when the work is well done, so that the metal is nowhere exposed to the action of the water, this plan answers very well. The bottom, as well as the sides, should be of glass, however, - a plate of common window-glass, cemented to the cast-iron bottom, answering every purpose. A very excellent aquarium may be made with slate for the bottom; and for the corners four cast-iron pillars, into which the glass is cemented. The slab of slate should be considerably larger than the space inclosed by the glass, so as to secure abundant strength; and as slate is as easily cut and planed as wood, the edge may be molded so as to have a very handsome finish. A slab of marble is sometimes used, but it is entirely unsuitable, unless when covered with glass, firmly cemented to it. The reason of this is that marble is soluble in water containing carbonic acid, and it forms a deposit on the sides of the tank, besides injuring the fish.