eggs appear to have hatched and at harvest time the majority of the wheat stems had been bored and many were cut off at the base. Careful harvesting and the use of horse rakes saved a large proportion of what otherwise would have been a total loss. The infestation was much more general than in 1916.

A somewhat hasty reconnaissance was made through north-central North Dakota in August, 1919, that it might be ascertained as definitely as possible just how the Cephus attack was progressing. A number of fields in Bottineau County were examined and found to be heavily infested. Most of these had been raked after harvest and it was consequently impossible to compute accurately the percentage of infestation. The numerous sawfly-inhabited stubs in the drill rows, however, proved the severity of the attack. It was roughly estimated that about 30 per cent of the grain had gone down in most of these fields as the result of Cephus work. This figure is probably very conservative.

It is conceded by many observers in that region that the injury during the year 1919 was greater than during any previous year since the study of this pest was begun. More fields had been seriously invaded and were injured to a larger extent than had before been observed. Even fields of durum wheat, hitherto believed to be nearly free from fly attack, were severely injured in 1919, if the statements of reliable farmers are to be accepted. The question of immunity of durum wheat will be discussed later in this paper.

It may be stated, however, that the farmers are profiting by past experience and have used horserakes in stubble fields to such an extent that the percentage of actual loss of grain has been reduced to a small figure. The quality of grain from the fallen straw is naturally somewhat below the normal, since the work of the larvæ in the stems produces some injury in the heads as they fill.

Cephus was found mining wheat near Hettinger in southwestern North Dakota, July 18, 1917. September 22, 1917, infested wheat was found near Mott, 30 miles north of Hettinger. In October of the same year many wheat fields in Towner and Cavalier Counties, in northeastern North Dakota, showed heavy infestation, although during the previous year it was difficult to discover more than a trace of Cephus presence in the wheat in this region. None was found in the vicinity of Fargo, although it doubtless occurs throughout the entire Red River valley.

A gathering of sods of Elymus canadensis sent to the writer from Charleston, Mo., during the summer of 1917 contained at least one larva of Cephus cinctus that had been boring the stem of this grass in that region. This locality is a little south of the latitude of Pinto, Utah, where this insect abounds.

Roughly speaking, so far as is now known, the sawfly inhabits an area bounded on the north by a line far into Canada; on the east by the Mississippi River, or probably a little east of that; on the south by latitude 36°; and by the Pacific Ocean on the west.

From the foregoing brief summary of its history it will be seen that Cephus cinctus is distributed over an immense territory and that it constitutes a potential menace to the small grains throughout this vast area. As the acreage of native grasses is decreased from year to year by the bringing of wild lands under the plow, pests such as the sawfly will be forced to depend in an increasingly large measure upon the small grains and other products of the farms. On this account the injury caused by these formerly harmless insects bids fair to increase steadily. In the past the numbers of grass-feeding insects such as the one considered in this paper have been governed mainly by the supply of food plants. A dry summer that retarded the growth of long-stemmed grasses would automatically reduce the numbers of the insects that lived within these grass stems and perhaps bring certain species to the point of extinction. It is easy to see how seasonal fluctuations in vegetation would, to a large extent, either multiply or diminish the numbers of these insects.

Then again, the farmer, by introducing fields of grain into a region previously uncultivated, brings in conditions unknown before and invites the attack of these and other formerly harmless insects, making it possible for them to become a menace to his future. Such a study of life history as has been attempted in this paper is urgently necessary in order that control measures may be undertaken successfully when such insects become pests.

FOOD PLANTS

The various species of Agropyron and Elymus, genera both of which are well represented in the West, appear to have been the original hosts of the larva. Since their feeding habits have been modified by changing agricultural conditions, the list of their present host plants, so far as known, stands as follows:

Barley probably should be added to this list.

Since the larva is wholly unable to move from one stem to another, it is very obvious that the host stem must be large enough to afford both shelter and food during its entire growing period. Hence only the larger-stemmed grasses can be mined successfully by the Cephus

larvæ. Occasionally an unusually vigorous plant of a slenderstemmed grass, like Hordeum jubatum, affords stalks with diameter sufficiently great to be attacked by Cephus.

Small grains, such as wheat and rye, readily serve as hosts to this insect, because they are of suitable size and the length of their growing season coincides with the growth of the larva. Even if harvest time should happen to come before the maturity of the larva, the reaping machine probably would sever the stem far enough above ground to leave the larva below the sickle cut, where it could house itself safely before the end of the season.

Judging the future by the recent past, it seems probable that this fly, before another decade is past, will be found attacking practically all of our native and cultivated grasses and most of our grains.

It must be remarked in this connection that up to the present time. this species has confined itself entirely to the West and has been found in only a few localities east of the Mississippi River. Its choice of wheat for food has taken place, so far as known, only in North Dakota and western Canada, although it is probable that Montana wheat fields have been invaded. From present appearances its attacks probably will be confined to vegetation growing within the area where spring wheat is sown.

THE EGG

The egg of Cephus cinctus is, when newly laid, decidedly crescentshaped, glassy in appearance, milky-white in color, usually quite symmetrical, the ends of the crescent tapering and rounded. It is marked by very faint, short, longitudinal lines or wrinkles, placed without regard to order or pattern.

The size of the egg varies with the size of the female that produced it and measures from 1 mm. to 1.25 mm. in length. The greatest breadth is about one-third the length.

The covering membrane is hyaline and transparent. Although very thin and delicate it is sufficiently strong so that the egg may be safely lifted and moved by the aid of a fine brush. The egg always lies free within the stem of the host plant, either in the stem cavity or in a hollow excavated by the ovipositor of the female that placed it. This cell is always a little larger than the egg, so that it is comparatively an easy matter to remove the egg to a moist cell or elsewhere for study.

The number of eggs distributed by each female appears to vary but little. Dissections of a number of adults taken in the field and of others reared in captivity agree in most cases in giving a count of about 50 eggs in the ovaries, these eggs being, as a rule, equal in size and apparent maturity.

150056°-20-Bull. 841- -2

DEVELOPMENT OF THE EGG

After a number of trials it was found to be impracticable to rear the egg in situ, since it was next to impossible to maintain the proper moisture conditions within the stem. The method that finally was adopted, and that gave excellent results, was to remove the egg from the stem and place it in a minute drop of water within a small thin watch glass which was then immediately inverted on a glass slip and sealed with a ring of water to prevent undue evaporation. This form of moist cell proved quite satisfactory and permitted continuous examination of the egg with a moderately high-power lens during the entire period of incubation. It was found necessary, in order to continue the requisite moisture supply during a period of several days, to invert over the sealed cell a larger watch glass and over this in turn a tumbler. In this manner evaporation was reduced to a minimum. It is altogether probable that the amount of moisture in such a protected cell exceeded that normally present within the grass stem, but in every egg treated in this way the incubation appeared to proceed naturally.

Temperature and moisture are, without any doubt, the prime factors that hasten or retard the egg development. The temperature maintained within the laboratory during the course of these investigations was much more equable than that in the field, where, as in Utah, the heat of the sun through the daytime, followed by a chilly night, must alternately hasten and check development. The data given below, therefore, may only approximate what actually takes place under field conditions.

A few hours after the egg leaves the oviduct the milky-white contents of the egg which at first completely filled the envelope shrink a little from each end leaving a transparent space or vacuole. Gradually the interior mass of exceedingly minute particles coalesces until about the second day when a series of faintly discernible cells arranging themselves along a central axis begins to appear. Early on the third day the form of the larva can be dimly seen, the head being almost transparent and filling one end of the egg sac. The body is looped on itself, the cauda folded beneath the abdomen and extending forward nearly to the head. By the close of the third day the abdominal segments are usually well defined.

During the fourth day, in most cases, a spasmodic and intermittent heart beat may be noticed. These pulsations become more and more regular as the hours pass and during the fifth and sixth days the heart beats with much regularity at the rate of about 120 impulses per minute. At intervals, for some unknown reason, it may slow down to 75 beats, but soon resumes its former rate.

The head appears abnormally large at this time, but although its general outlines are well defined the brown jaws and eye spots are

not yet visible. Over night, at the close of the fifth day, the jaws turn brown and the eye spots appear and darken. Usually, after the fourth day, the muscular system of the larva is in almost constant motion, shifting and adjusting, with the heart pulsating and the muscles moving, all clearly to be seen through the transparent membrane that serves as the shell.

The activity of the larva within the sac increases during the sixth day, and either on this day or the seventh it escapes from its confinement by a series of convulsive movements that rupture the delicate shell and set it free.

After the first day the egg changes shape, becomes intumescent, generally loses its crescentic shape entirely, and grows oval or reniform in outline.

THE LARVA

When it escapes from the egg the larva (fig. 6) possesses a very large head armed with a pair of powerful biting jaws, a weak, slender body, and a most vigorous appetite. It is very active from the start and begins almost at once to feed upon the living parenchymatous tissue by which it is surrounded in the interior of the stem, excavating for itself a threadlike gallery both above and below the spot where the egg formerly lay. The larva is at first nearly transparent and colorless until it becomes filled with the tissue on which it exists.

The body segments are strongly and clearly marked from the time the larva leaves the

Newly-hatched larva. Greatly enlarged.

egg. The jaws are brown, three or four FIG. 6.-Western grass-stem sawfly: pointed, the points chisel-shaped, beveled on the inside edge. The brown face plate is filled with crossed bands of striated muscular fiber that actuate the powerful jaws which form the most important item of the domestic economy of the young Cephus. The caudal horn, by means of which the larva moves up and down in its gallery, is also brown and is armed, even in the first instar, with a series of stout bristles at the base of its cylindrical and squarely truncate extremity. The larva is footless, the position of the legs being marked by minute, rounded tubercles terminating in a few short bristles.

Although the primary excavation made by the larva may extend for a short distance above the egg cell, the general course of the progress is invariably downward. In its earlier stages of existence, at least, the larva traverses its gallery several times, swallowing repeatedly the same fragments of tissue that have already been devoured during the first excavation of the stem. Young larvæ are