the conditions observed must have been due to the failure of a spermatogonial mitosis to complete itself.

Several of the Carabidæ have been studied, and the material, though not especially favorable, is interesting in that some members of the family have an unequal pair of heterochromosomes, others an odd one. Chlænius æstivus (figs. 207-212), Chlænius pennsylvanicus (figs. 213-215), and Galerita bicolor (fig. 216) have the unequal pair, while Anomoglossus emarginatus (figs. 217-223) has an odd heterochromosome (x), which behaves exactly like the larger heterochromosome in other carabs.

In the Elateridæ and Lampyridæ we also have examples of the second type with the odd chromosome. Two Elaters, species not determined (figs. 224-229 and 230-235), have each 19 chromosomes in the spermatogonia (figs. 224 and 230), and in the first spermatocyte division an odd chromosome (x) which is in each case the smallest. In the first of these Elaters, the female somatic number was determined to be 20 (fig. 229). In the second Elater the pairs of second spermatocytes, containing 9 and 10 chromosomes respectively in the two cells, were in nearly every case connected as shown in figure 235, one pair of chromosomes not having separated completely in the first mitosis. Of Ellychnia corrusca (family Lampyridæ) only the spermatogonial equatorial plate, containing 19 chromosomes (x, the odd one) is given, as no material in maturation has yet been obtained, and a comparative study of the germ cells of the Elateridæ and Lampyridæ will be made as soon as suitable material can be secured.

In addition to the species of Coleoptera described here, two others, Coptocycla aurichalcea and Coptocycla guttata have been studied by one of my students and the results published elsewhere (Nowlin, '06). In both an even number of chromosomes (22, 18) was found in the spermatogonia, one being very small and forming with a larger one an unequal pair which remained condensed during the growth stage and separated into its larger and smaller components in the first spermatocyte mitosis. The result of maturation, as in the other species here described and in Tenebrio molitor, is dimorphism of the spermatozoa. The method of synapsis in Coptocycla is like that described for Chelymorpha argus.

HEMIPTERA HOMOPTERA.

Aphrophora quadrangularis.

The abundance of Aphrophora at Harpswell, Maine, in June and July, 1905, suggested that it might be well to examine at least one more of the Hemiptera homoptera for comparison with the many species of Hemiptera heteroptera which have been recently reexamined by Wilson ('05, '05, '06).

The larvæ only were collected, as they gave all the desired stages for a study of the spermatogenesis, and also oögonia and synizesis and synapsis stages of the oöcytes. In the first collections the testes were dissected out, but the many free follicles break apart so easily that the later material was prepared by cutting out the abdominal segments which contained the reproductive organs, and fixing those without dissection. The same methods of fixation and staining were employed as for the Coleoptera. Hermann's safranin-gentian method was especially effective with this material.

In Aphrophora the follicles of each testis are free, forming a dense cluster, each follicle being connected with the vas deferens by a short duct. The very young follicles are spherical, the older ones ovoid in form. The primary spermatogonia (plate XIV, fig. 237)—very clear cells with a lobed nucleus which stains slightly—occupy the tip of the follicle. Next to these comes a layer of cysts of secondary spermatogonia which are conspicuous for their deeper staining quality (fig. 238). There appears to be no plasmosome in either class of spermatogonia. Figure 239 is the equatorial plate of a secondary spermatogonium. There are 23 chromosomes, two of which are conspicuously larger than the others and evidently form a pair. The odd one is one of the three next in size.

Next to the secondary spermatogonia are cysts of young spermatocytes, whose nuclei show a continuous spireme and an elongated deeply staining chromatin rod which is the odd chromosome (fig. 240). This is often more elongated than in the figure and more or less wormlike in appearance. A pair of smaller chromatin masses may sometimes be detected at this stage, and are readily found a little later (fig. 241) when the nucleus has enlarged and the spireme has become looser and stains less deeply. Here the odd chromosome is more condensed, or shortened, and split. There is no synizesis and no polarized or bouquet stage, but the nuclei of all of the spermatocytes contain a continuous spireme throughout the growth stage. Synapsis must occur at the close of the last spermatogonial mitosis before the spireme is formed. Figures 242 and 243 show a slightly later growth stage. The form and connection of the "m-chromosome" pair (Wilson, '05_b) comes out clearly here. Figure 244, from a safranin-gentian preparation, shows both the odd chromosome and the m-chromosomes. Some time before the first mitosis, the spireme splits and the pairs of granules embedded in linin are wonderfully distinct, both in iron-hæmatoxylin and safranin-gentian preparations (fig. 245). The m-chromosomes have here formed a precocious tetrad (m). Figure 246 is a similar stage from a safranin-gentian preparation. Figures 247 and 248 show the condensation of chromatin granules to form tetrads of various sizes, still embedded in the linin spireme. As these tetrads come into the spindle without losing their elongated form, it is evident that each one consists of two longitudinally split chromosomes united end to end in synapsis and separated in the first maturation mitosis, which is therefore reductional. The odd chromosome and the m-chromosomes show no longitudinal split in these figures, but they may appear as in figure 249. Occasionally one of the tetrads takes the form of a cross (fig. 249). In this figure the split "accessory" (x) lies against the nuclear membrane and the archoplasmic material for the spindle is seen along one side of the nucleus. It is certain here that the spindle fibers come from extranuclear material, not from nuclear substance, as Paulmier ('99) describes for Anasa tristis.

Figures 250 and 251 show the first maturation mitosis as it usually appears in sections from mercuro-nitric material stained with iron-hæmatoxylin. The odd chromosome is always more or less eccentric and is attached by a spindle fiber to one pole. In Hermann material, considerably destained, the tetrads and the odd chromosome appear as in figures 252, 253, and 254, the tetrads being in position for a transverse division. The odd chromosome is always so placed that its longitudinal split is at right angles to the axis of the spindle, as though it were to divide in this mitosis. It does not do so, however, but goes to one daughter cell, always lagging behind, as is shown in figures 255 and 256. Figures 257, a and b, are polar plates of the first mitosis with 11 and 12 chromosomes, respectively, and figures 258, a, b, and c, show the polar plates (a and c) each containing 11 chromosomes, and the odd chromosome at a different level (b). The latter is a view of the anaphase which one often gets at three foci in one section. Figures 259, a and b, are equatorial plates of the second mitosis with 11 and 12 chromosomes respectively. Figure 260 shows a side view of the second spindle in metaphase, and figure 261 in anaphase. Figures 262 and 263 are daughter plates from two spindles showing the chromosome content of the two