sent.

"Correct path," as opposed to the incorrect, logically implies in these experiments "aquarium”; and its selection, as shown by the ratio of improvement from day to day may, though not necessarily, imply the representation of the construct "aquarium," but it does demand the admission that acts of recognition and discrimination, or even of what LLOYD MORGAN calls "perceptional inference" take place. These in turn presuppose necessarily, as is well known, retention and production.

Carefully excluding the possibility of the crab's merely following a path by smell, taste, or touch (although if it did only this one could not account for a correct after an incorrect choice had once been made) YERKES found in one case that after 40, in another that after 250, experiences no mistakes in choosing were made. In a number of cases the subject turned from, before it reached, the partition which blocked the passage, thus showing the important part played by vision in directing the animal in the absence of smell, taste, and touch. All of these, however, together with muscular sensations, YERKES concludes normally play a part in the formation of labyrinth habits. These experiments therefore seem to show that upon the basis of the "constructs" which one sense alone, viz., vision, give the crab, a consistent selection of the correct path is possible; but this is explainable it seems, even if it is considered that only a recognition of each successive part of that path and consequently a discrimination between it and the incorrect is made, and yet that no representation or "reconstruction" of "aquarium" takes place, although of course this latter interpretation is not excluded.

A method of experimentation, however, which shows that in the formation of a habit, or in the learning of a motor reaction involving two sense fields, e. g., taste and vision, it is necessary to overcome an instinct or tropism in the opposite direction, such a method, we think, would at least give more cogent grounds for accepting the presence of representation than one not doing this, although even here conservatism in making this claim would be the safer course.

Some Characteristics of the Hermit Crab.

The genus Eupagurus is easily found in the shallower waters about Woods Hole and is represented by four species, longicarpus, annulipes, acadianus, pollicaris. E. longicarpus was selected for the present investigation on account of its convenient size (34-11⁄2 inches in length) for aquarium purposes, and because of a manifestly greater brightness. Supplementary experiments show that E. pollicaris, e. g., learns with greater difficulty.

Members of the entire genus inhabit, under normal conditions, the shells of gastropods, by which the abdomen is completely protected, the cephalothorax alone protruding. This peculiar mode of life is correlated with a dextral asymmetry, which extends to almost all the organs of the entire body, and which shows a very nice adaptation. This favors the view that the asymmetry is a result of life in dextrally spiral shells, exemplifying at the same time degeneration.

The establishment of the fact that these Hermits learn is not surprising in view of the complexity and fineness of their physiological sense apparatus, which is essentially the same as that of all the Crustacea, so that it is very probable that any denial of this ability to any species of the group, even upon the basis of experiment, is due to incomplete or faulty methods of investigation.

Sense Organs.

The crab has only two general kinds of sense organs, viz., eyes and sense hairs, the latter of which are, however, differentiated as to their function. These hairs, which are found in all the extremities, are epithelial in nature, and are not penetrated by a nerve, but rather this latter spreads out underneath each epithelial group and gives to each cell a fibril. These epithelial sense cells lie in a support of "Matrixzellen,"1 and according to variations in their structure and especially position are respectively gustatory, tactile, and auditory or "equilibra

1 VOM RATH, Orro. Zoologischer Anzeiger, No. 386, 1892.

tory." The gustatory hairs, lie as two patches of "minute curiously flat organs" on the under surface of the outer filament of the antennules, the innermost appendages, which observation shows are kept moving constantly. There are no sense organs in the mouth.1

At the basis of each antennule is a little sac formed by an infolding of the chitinous integument, communicating freely with the water, and containing little sand grains or otoliths. Here are present a second kind of sensory hairs, connected with the central nervous system by branches of the antennulary nerve, and whose function is either that of audition or of equilibrium. The third class of hairs are tactile in function, and are especially numerous on the antennae, i. e., the second pair of appendages, although there are some on the antennules, the other appendages, and the remaining integument.

2

The two eyes of the crab are compound or facetted and are seated on movable pedestals. They are covered by a transparent chitinous cuticle, forming a cornea; this is divided into facets, beneath each of which there is an ommatidium with two segments (a) an outer, which is vitreous and refractive, and an inner, a short retinula, which is sensitive, thus giving a structure analogous to rods and cones. These cones are surrounded and so separated from each other by a pigment; their apex is embraced by elongated cells in the midst of which is a fibril of the optic nerve. Each facet functions as a single eye and therefore like the vertebrate eye gives no sensitive continuum but, rather, "mosaic vision," i. e., various images in juxtaposition. The eye as a whole is supposed to give a vision of distinct objects. and space relations.

The brain is formed from the first three pairs of embryonic ganglia, and is therefore a "syn-cerebrum"; it supplies the eyes, the antennules and the antennae with nerves. It is connected

1 VOM RATH, OTTO. Zur Kenntniss der Hautsinnesorgane der Crustaceen, Zoologischer Anzeiger, 365, 1891.

'HENSEN (Studien über das Gehörorgan, Zeitsch. f. wiss. Zoolog., Bd. 13, 1863) says they are auditory, while DELAGE, Archiv. d. Zool. Expér., 1887, (2), T. 5, says they are for position and qeuilibrium. (ited by voм RATH.)

by two oesophageal branches with the central nerve cord, which is represented principally by a single large thoracic ganglion or concrescence of ganglia. The thoracic portion of the nervous system is, however, symmetrical. The mandibles, or the third pair of appendages, crushing jaws, the right of which is larger, the maxillae and the maxillipeds all receive nerves. The nervous system is therefore in general so constructed that it would seem at least reasonable to expect that associations might be formed.

Experimental

The systematic experiments by which the association between the "constructs" of two sense fields, taste and vision, was established, and a "reconstruction" or reproduction subsequently shown possibly to take place, were preceded by various preparatory observations, some of which were made in the summer of 1902. For instance, it was then shown that the Hermit is remarkably thigmotactic, for when a shell inhabited by a crab is suspended at the distance of about twice the diameter of the shell from the floor of the aquarium, the animal is thereby made decidedly uncomfortable, protrudes nearly its entire body, feels about, and usually leaves its shell, especially if there is a vacant shell near by. Suspended at the height of from eight to ten inches the crab will remain in the shell until it dies. They are also somewhat rheotactic. New shells thrown into the aquarium are soon examined and accepted at what would sometimes seem to be a disadvantage. This constant "desire" for change, together with both a great natural rapacity and pugnacity, are indeed indications of a strenuous life even among Hermits.

Both that series of observations upon which special emphasis is placed in this paper, and that preliminary one which showed that the method adopted would probably lead to satisfactory results, were made with a very simple apparatus and in a very simple way. A number of crabs which had been kept in an ordinary laboratory glass jar aquarium about twenty inches in diameter, were made to go into a darkened portion of this that they might get their food, which consisted of a freshly

cleaned Fundulus held in place on a wire. The same portion of the aquarium was darkened each time just before feeding by setting down into the constantly running water a screen consisting of two thin boards fastened at right angles, leaving only an opening at each end of the vertical board wide enough for the crabs to go through one by one; around the outside of the aquarium from end to end of this portion set off by the screen was kept a piece of heavy brown paper to shut out the light coming from the other direction. The only light which could enter came therefore through the openings at each end. Sand was placed in the bottom of the aquarium, and all the conditions such as its position and that of the tap, with the exception of the putting in and taking out of the wooden screen, were kept constant during the entire series of experiments.

These were conducted in detail as follows: to establish, first, an association and, second, present an occasion for possible "reproduction" from the "after-effect" of one by the external stimulus of the other "construct.'

Thirty-six crabs of the species Eupagurus longicarpus were placed in the aquarium on July 30th, and, first, allowed until August 6th to become accustomed to aquarium conditions; during this period they were simply fed each day with a fresh Fundulus, no screen being used; they seized their food most eagerly, oftentimes fighting and driving each other away from it. The death of six selected the thirty most fit individuals. The crabs were, furthermore, observed to remain in the lightest part of the aquarium practically all of the time, i. e., they were positively heliotropic. The positive heliotropism was confirmed by a number of control experiments with other lots and by that of Aug. 6th. The screen was inserted and all the crabs, 30 in number, placed behind it. In 10 minutes 28 had gone through the opening into the light and 27 of these were near the point of its maximum intensity.

Each day following this, the screen was inserted, a fish on a wire placed in the darkened portion, and the number of crabs going into the dark through either one of the entrances at the end within a given time, which was constantly shortened during