other periodicals, but such are closed books to most persons. Anyone who looks for information in the popular works on natural history of the day must inevitably be disappointed at the meagerness of the information given. Even in the voluminous German work, so well known as Brehm's Tierleben, the information is meager for almost all fishes, and especially meager for American forms. The sources of knowledge have not been discovered by the compilers of such works, but he who might judge from the paucity of data that no others could be found would be much deceived. To uncover some of the interesting details hidden in comparatively little known journals and other works is the object of the present article, which is devoted to the record of facts about the mating and breeding habits of some among many remarkable species. It is hoped that the information given may indicate points to be observed in the history and economy of other species, as well as of those already noticed. There is, indeed, an urgent call for corroboration and amplification of most of the histories given, as well as for discovery of the natural history of other species.

The species which manifest care for their young are so numerous that the present article must be restricted to those which are inhabitants of fresh water. Such are better known than the marine forms, as they are more easily observed and within the range of observation of a more numerous population. Considerable is known, however, of the habits of many of the dwellers in salt water. Parental care has been especially observed in the marine pipefishes, sea-horses, Pegasids, Solenostomids, Sparids (e. g., Catharus), Labrids (Wrasses), toad fishes, gobies, blennies, sculpins or Cottids, lumpfishes, Gobiesocids, etc. Doubtless analogous care will be found to be exercised by many more when fishes shall have been more thoroughly studied.

Naturally the most common or frequent mode of care is the simplest, consisting of little more than selection of a site for the deposit of the female's eggs and subsequent guardianship of those eggs by the male. The concomitants of such selection are various. In the case of the American sunfishes, black basses, and crappies, the place selected is cleared of stones and weeds, and in the cleared places the eggs are laid. Some of the sunfish-like Cichlids and the North American catfishes, as well as the Grecian glanis, exercise similar means with slight modifications. Another kind of catfish, living in North Australia (Queensland), lays her eggs in the center of a selected area of a river bed, and, after having fertilized them, the fish accumulates stones from the surrounding area and piles them in a heap over the eggs. Other modifications of a general plan appear to be executed by other fishes, but the details remain to be investigated.

Oral gestation or carriage of eggs within the mouth of the parent fish is practiced in a number of unrelated species. In Siluroid catfishes it is associated with enlarged size of the eggs, as in most Tachisurines and one of the Pimelodines (Conorhynchos), and is confined to the males. The Malapteruroid, or electric catfish, is also said to be an oral egg carrier. In other fishes the eggs of the egg carriers are not essentially different from those of normal habits, and many related species do not have the peculiarity. Such egg carriers are Cichlids of America (Geophagus) and of Africa (Tilapia, Tropheus, Ectodus, etc.), as well as species of the marine genus Chilodipterus, one of the Apogonids.

It is especially noteworthy that among the Cichlids are exceptions to the rule that the care taker is a male. In several cases it has been verified that the egg carrier is a female and presumably, of course, the layer of the eggs. From evidence so far accumulated it would seem that the sex of the care taker is coincident with specific characters, and that when the care taker is a female the male is not. At least, in a recent article (L'incubation buccale chez le Tilapia galilaea Artedi, 1904), J. Pellegrin showed that all four individuals of the species examined which had eggs in the mouth were females, and he could find no male egg carriers. Boulenger previously had found eggs in the mouths of females only of Tilapia nilotica and other Cichlids of the genera Ectodus, Tropheus, and Pelmatochromis. On the other hand, the sex of the egg carrier of Tilapia philander was determined by so competent an authority as A. Günther to be male. Further, Lortet named a Cichlid Tilapia paterfamilias, which was declared by Pellegrin to be specifically identical with Tilapia simonis. Lortet gave his name, because he considered the egg carrier to be the male, while Pellegrin confirmed, by dissection, the sex of a specimen of the same species to be female. Evidently, then, there is necessity for further observations as to the sex of the egg carriers of African, as well as the American, Cichlids."

a The eggs of some of the ovigerous Cichlids are very large. According to Boulenger (T. Z. S., XV, 18)," the mouth and pharynx of" a female Tropheus moorii contained "four eggs of very large size, the vitelline sphere measuring 4 millimetres in diameter, with an embryo in an advanced stage of development. The eggs of the fifteen-spined Stickeback, hitherto regarded as the largest Teleostean egg in proportion to the size of the animal, measures only 3 millimetres in diameter." The egg-carrying Tropheus was only about 4 inches long.

The above remarks are left just as they were printed, but on the same day as the proof of this article was received from the printing office a "Fourth Contribution to the Ichthyology of Lake Tanganyika," by G. A. Boulenger, fresh from the press, was also received, containing the much-needed further observations." Doctor Boulenger records that Doctor Cunnington, the latest explorer of the ichthyology of the lake, had been "so fortunate as to considerably extend the list of Cichlid fishes in which the parents protect their

One of the most singular devices for protection of the eggs is the formation of bubbles of air rendered tenacious by a viscous secretion of the mouth. This might seem to be so specialized a method that there would be no independent repetition by representatives of a very different group. In fact, however, the method, or an analogous one, has originated anew time and time again. It is manifest, so far as known, in its simplest form in the fish-of-paradise (Macropodus viridi-auratus), which makes a floating nest of a mere conglomeration of bubbles, but other species of the same family (Osphromenids) evolve nidamental receptacles little more complicated, and among them is the celebrated Gourami. The nests earliest described, in which air bubbles formed an essential element, were those of Hassars (Callichthyids of the genus Hoplosternum), but in them the air or "froth" was used in combination with vegetable material ("fallen leaves or grass "). The large floating nest of the African Gymnarchus, recently described by Budgett, may, perhaps, be partly buoyed up by aeriform secretions. The future investigation of the structures involved in the secretion of such bubbles will undoubtedly yield most interesting results.

More specialized than any of the methods of parental care hereinbefore noticed is one manifested by certain American fishes. Those fishes are of a family named Aspredinids, peculiar to the fresh waters of South America and distantly-and very distantly-related to the catfishes of the north. In them it is the female that assumes charge of the eggs, and she does it in a strange and truly characteristic manner. After the eggs have been discharged from the ovaries (and presumably after they have been fertilized), the mother presses her belly and breast over them and they become attached thereto; then the areas of attachment of the skin become elevated into cupules round the eggs, like the cups of acorns round the nuts, and not only so, but strangulation ensues between each cupule and the general skin of the belly, so that the eggs and cupules are borne upon stalks or peduncles, and so they remain till the eggs are hatched and the offspring by giving them shelter in the mouth and pharynx. This mode of nursing is illustrated" by examples of seven additional species of six genera. "The natives" round the lake "say it is always the female, in the cases where one of the parents takes the eggs in the mouth," that is the carrier. This belief has now been confirmed by Boulenger for no less than ten African genera; in fact, whenever he had been able to test the sex of the egg carrier it was "invariably the female who thus carries the eggs. This was in contradiction to statements made by Lortet and by Günther, who ascribed the habit to the male in the species of the same genus with which they had dealt." Of course it is easy enough to tell by dissection whether a fish is a male or a female, and the authors in question probably neglected to take the proper means to ascertain the sex, but took it for granted that the nurse was a male. See for further data the paragraphs on the Cichlids hereafter.-October 25, 1906.

embryos leave, after which the skin returns to its former condition. The only analogue to this occurs in certain anurous batrachians, but in such in a less specialized condition and on the back.

The most specialized of all the care takers are the sticklebacks, or Gasterosteids. These have an important organ (the kidney and its adjuncts) especially modified histologically to yield a thread analogous to that developed by spiders and used for binding the objects selected for a nest. So far as known none of the related fishes has the same structural peculiarity, but it is quite possible, if not probable, that their nearest relatives of the North Pacific-the Aulorhynchids may have a similar history. It is scarcely within the range of possibility that an analogous structure like this should have been independently developed a second time in unrelated fishes.

One remarkable, and to some astonishing, fact is the want of correlation that may sometimes exist within a natural family between structural features and habits. This is strikingly manifest in the typical catfishes or Silurids. Neither of the parents of the wellknown wels of central and western Europe appears to care for eggs or young, but the male of its near relation of Macedonia-the glanis assumes a special charge of his consort's labor. Opposite ways of making their nests are practiced by the North American catfishes on one hand and certain Australian ones on the other. Enlargement of the eggs is manifested in another group and is associated with their reception and carriage by the male in his mouth.

In the last English work on fishes, the Cambridge Natural History, it is declared that, "with the exception of the pelagic Antennarius, which builds its nest in the sargasso weed in mid-ocean, nest building and parental solicitude for the young are confined to fresh water fishes and to marine forms with demersal eggs. Pelagic ova must necessarily be beyond the scope of parental care." The so-called Antennarius is no exception; the species meant is not a true Antennarius, but belongs to a distinct though related genusPterophryne. Its history is a truly remarkable one and it has been more widely noticed as a nest-building fish than any other except the Gasterosteids. In truth, however, it does not build a nest at all. The whole story is the result of a misidentification of the eggs of a fish. In 1872, the celebrated naturalist, Prof. Louis Agassiz, attributed egg-bearing masses of gulf weed (sargassum), which he found in the gulf stream, to the Pterophryne which was abundantly associated with them. His equally able son, Dr. Alexander Agassiz, a decade later (1882), made known the remarkable egg raft which floated the eggs of a relative of the Pterophryne-the angler (Lophius piscatorius). This discovery may have led to thought, but

a Cambridge Natural History, Vol. VII, p. 414 (1904).

did not discredit the earlier identification." In 1905, however, Dr. E. W. Gudger and Dr. Hugh Sraith both caught females of Pterophryne in the act of emission of egg rafts like that of the angler. Of course the parent of such eggs and egg rafts could not be the maker of the nests attributed to the Pterophryne. What fish, then, was the maker of the nests? The only eggs like those found in the nests are those of some flying fishes (Exocœtids). We are forced, therefore, to assume that a flying fish had laid her eggs on a frond of the sargassum, and that they had been fertilized by the male. These eggs have bipolar bunches of very long filamentary tendrils, and such have mechanically grasped and brought together the finely divided branches of the sargassum, with the result that subglobular masses have been formed in which the eggs are protected. They

answer every purpose of a nest, but are they nests? If so considered we must admit that the eggs and not the fish make a nest!

Some of the fishes to be noticed in the following pages have the males gaudily colored and larger than the females. The relations of the sexes to each other with regard to color and size are noteworthy, inasmuch as they have been generally misunderstood. One eminent ichthyologist (Doctor Günther), in an "Introduction to the study of fishes" (p. 656), dogmatically declared that "with regard to size, it appears that in all teleosteous fishes the female is larger than the male," and Darwin was assured by him that he did "not know a sin

a As Professor Agassiz did not notice any filaments on his eggs, I thought it possible that some real eggs of a Pterophryne may have drifted on the outside of an egg mass, but Doctor Agassiz kindly sent me a couple of eggs from the outside and they proved to have bipolar filaments and consequently to be eggs of flying fishes.