from warm waters; a relation which was strikingly manifest in the analyses of echinoderms and alcyonarians and which has been amply verified by a considerable number of new analyses made since the original memoir was published. In other groups of organisms the same relation was suggested, but not actually proved to hold, for there were exceptions that needed explanation. In a series of eleven analyses of crustaceans (crabs, lobsters, shrimps etc.), the same variation in magnesia was strongly indicated, but with irregularities which appeared to require further investigation. It was conceivable that different parts of a shell or skeleton might differ in composition, or else that variations might be due to differences in age. It had already been found in the case of two sea urchins that the spines contained much less magnesia than the main body of the shells, but the question relative to age remained to be investigated.

Through the kindness of Dr. H. M. Smith, director of the U. S. Bureau of Fisheries, the large claws of two lobsters (Homarus americanus) from a single locality, Boothbay Harbor, Maine, were obtained, one from a small lobster, the other from a large specimen. The analyses gave the following results, after rejecting organic matter and water and recalculating to 100 %

The difference between these two analyses is very great, the large animal being much more highly magnesian and phosphatic than the small one. Unfortunately, however, the actual sizes of the two lobsters were not given, and more precise data were evidently desirable. Accordingly Dr. Smith had fragments from three lobsters sent to us, all from the same station as the others, with definite figures as to length and weight. The fragments, moreover, in each case represented both the large claw and the carapace, so that variations in the individual as well as variations in age could be determined. The analyses, six in number, were as follows:

1. Small lobster, length 8 inches, weight 10 ounces.

2. Medium lobster, length 11 inches, weight 2 pounds.

3. Large lobster, length 16

cated by a, the carapace by b.

inches, weight 5 pounds. The claw is indi

In each case the claw is richer in magnesium carbonate and calcium phosphate than the carapace. The variations due to age appear more distinctly when the average of each pair of analyses is taken, as follows:

Here the progressive increase in magnesium carbonate and calcium phosphate is clearly shown; and it also appears, in the percentages of calcium sulphate, although the last detail is less significant. The smallest lobster, moreover, differs in the composition of its inorganic portion from that of the two larger animals much more than they do from each other.

From the evidence now at hand it seems clear that some of the departures from regularity in the proportions of magnesium carbonate in the shells or skeletons of marine invertebrates are due to one or both of the two causes which were suggested at the beginning of this paper. It is desirable, therefore, in further investigations of this kind, that in the study of the more highly specialized organisms the analyses should represent the totality of the inorganic portions, and that animals of the same degree of maturity should be taken. With the lower classes of organisms the difficulties are not so great, and regularities are much more easily discovered.

(Published by permission of the Director of the U. S. Geological Survey.)

1 Clarke and Wheeler, Prof. Paper, No. 102, U. S. Geological Survey, Washington.

FLAGELLATE AFFINITIES OF TRICHONYMPHA

BY CHARLES ATWOOD KOFOID AND OLIVE SWEZY
ZOOLOGICAL LABORATORY, UNIVERSITY OF CALIFORNIA

Communicated by W. M. Wheeler, November 13, 1918

The methods of division among the Protozoa are of fundamental significance from an evolutionary standpoint. Unlike the Metazoa which present, as a whole, only minor variations in this process in the different taxonomic groups and in the many different types of cells in the body, the Protozoa have evolved many and widely diverse types of mitotic phenomena, which are characteristic of the groups into which the phylum is divided. Some striking confirmation of the value of this as a clue to relationships has been found in recent work along these lines. The genus Trichonympha has, since its discovery in 1877 by Leidy,1 been placed, on the one hand, in, the ciliates and, on the other, in the flagellates, and of late in an intermediate position between these two classes, by different investigators. Certain points in its structure would seem to justify each of these assignments. A more critical study of its morphology and especially of its methods of division, however, definitely place it in the flagellates near the Polymastigina.

At first glance Trichonympha would undoubtedly be called a ciliate. The body is covered for about two-thirds of its surface with a thick coating of cilia or flagella of varying lengths, which stream out behind the body. It also has a thick, highly differentiated ectoplasm which contains an alveolar layer as well as a complex system of myonemes. The nucleus, however, is that of a typical flagellate. The flagella may equally well be called cilia, since they are arranged in longitudinal rows on the surface of the body, each arising from a minute basal granule imbedded in the ectoplasmic layer. Each basal granule, however, is connected with a fine fibril arising from the myonemes in the ectoplasm. The myonemes form a closely anastomosing network over the body, taking their origin from a complex structure at the anterior end which we call the centroblepharoplast (fig. 1, c). This corresponds to the blepharoplast of Trichomonas (fig. 6, c). The entire group of flagella are thus bound together into one integrated unit, the basis of which is the centroblepharoplast, forming the neuromotor system.

This integrated organelle system is found in a simple form in Trichomonas (fig. 6) where it consists of a centrosome-blepharoplast (c) imbedded in the anterior end of the axostyle (ax). From it arise the flagella and the undulating membrane with its marginal filament and parabasal body.

A more complex form of the same system and one which in some features leads toward the stage attained by Trichonympha with its much greater multiplicity of flagella, is shown in the motor organelle complex of Giardia.2

Here we find eight distinct flagella arising from different points on the surface of the body, but connected internally by a number of fibers which all take their origin from the centroblepharoplast complex attached to the anterior ends of the axostyles. The condition in Giardia differs from that in Trichonympha mainly in the small number of parts of its motor organelle complex. Correlated with the very great increase in the number of flagella in the latter species, is the vast increase in the number of distinct myonemes which have been developed. This has also necessitated an enlargement and increased complexity in the centroblepharoplast complex at the anterior end of the body. These structures still remain, however, distinctly flagellate in their structural relationships, though superficially resembling the ciliary coating found in the holotrichous Ciliata.

The mode of division found in Trichonympha offers still more striking proof of its flagellate affinities, and also emphasizes the wide divergence existing between it and the ciliates. As the name signifies the centroblepharoplast complex takes the rôle of centrosome in mitosis. In the prophase this entire structure divides by longitudinal splitting, followed by a division of the ectoplasmic part of the body, including a separation of the myonemes and flagella into two parts (fig. 2). As the centroblepharoplast divides the two moieties spin out a darkly staining band between them, the paradesmose (par).

The nucleus migrates from its submedian position to the anterior part of the body. A precocious splitting of the chromosomes has taken place during the vegetative phase and the fifty-two chromosomes thus formed are reunited, forming twenty-six V-shaped chromosomes composed of distinct chromomeres (fig. 2). As the nucleus reaches the paradesmose it elongates until its length coincides with that of the paradesmose. The nuclear membrane remains intact throughout the entire process of division. Spindle fibers are formed from the ends of the paradesmose, or the centroblepharaplasts, pass through the nuclear membrane and become attached to the chromosomes, a fiber from one pole to one end while the other is connected with the opposite pole. With the shortening of the spindle fibers the chromosomes are straightened out in the equatorial plate stage (fig. 3).

In all these mitotic figures the divided centroblepharoplast functions as the centrosome, while the myonemes, surface ridges and flagella are found attached to each half and streaming out from it like the astral rays in mitosis, in a metazoan nucleus. The paradesmose lies above the nucleus, that is, towards the surface of the body, outside the nuclear membrane but partly enfolded within it.

In the telophase the constriction of the nuclear membrane takes place with the chromosomes still showing their attachment to the spindle fibers. The chromosomes are never drawn close to the poles as in mitosis in Trichomonas, but the fibers fade out before reconstruction begins in the nucleus

(fig. 4). As the daughter nuclei round up they become free from the paradesmose which still connects the two daughter blepharoplasts (fig. 4). Later the paradesmose loses its staining reactions and fades out or is absorbed, either by the cytoplasm or by the centroblepharoplasts. This is followed by division of the body and separation of the two daughter flagellates.

A comparison of various points of this process with certain ones in Trichomonas reveals a striking similarity in the two types. In the latter the

FIGS. 1-5. Trichonympha campanula Kofoid and Swezy. 1. Anterior portion of trophozoite showing three groups of flagella, centroblepharoplast, myonemes and uncleus. X 225. 2. Late prophase of division. Ectoplasmic structures and flagella divided; paradesmose connecting the centroblepharoplasts; 26 split chromosomes. X225. 3. Metaphase of nucleus, X 600. 4. Telophase. Centroblepharoplasts still connected by paradesmose. X 600. 5. Mitosis completed, cytoplasmic division approaching. X 225.

division of the centroblepharoplast also produces a connecting paradesmose (fig. 7, par.). Attached to each new centrosome or centroblepharoplast are the motor organelles, which are produced partly by division of the old structures and partly by new outgrowths. The condition here is similar to that in Trichonympha (fig. 2). A precocious splitting of the chromosomes produces ten moieties which become reunited at the time of spindle formation, giving five as the number of chromosomes going to each daughter cell. The same process also occurs in Trichonympha.