ABOUT thirty years ago, two Fedias with fruits of singular shape were discovered by Mr. Sullivant, near Columbus, Ohio, and published by him as new species under the names of F. umbilicata and F. patellaria. They soon disappeared from their original station, and no botanist seems to have met with either of them again until the Rev. S. W. Knipe of the Delaware Water Gap collected, in the spring of 1870, a few specimens of F. patellaria, in Westmoreland County, Pa., and early in June, 1871, a large supply in the neighborhood of Columbia on the Susquehanna River, where it grew in great profusion along with the F. radiata of Michaux. Specimens of this plant, placed in my hands by the collector, exhibited such diversities in the fruit as to suggest the idea that both it and F. umbilicata might in the end prove to be forms of F. radiata. Dr. Gray, to whom the conjecture was communicated, kindly furnished fruits from Mr. Sullivant's plants, to complete the chain of evidence, and the information that F. umbilicata had also been rediscovered, last summer, on the Hudson River.

The Manual of Dr. Gray contains five species of Fedia; one an introduction from Europe (F. olitoria Vahl.), and four indigenous. All of them are much alike in general appearance, and amongst the latter especially the resemblance is so great that their specific characters are derived from the fruit alone; but how far these char

Entered according to the Act of Congress, in the year 1972, by the PEABODY ACADEMY OF SCIENCE, in the Office of the Librarian of Congress, at Washington.

AMER. NATURALIST, VOL. VI.

25

(385)

acters are constant and reliable, it will now be my endeavor to show, aided by illustrations from the p ncil of Mr. Knipe.

Fedia olitoria Vahl., Fig. 102. Fruit; a, side view; b, cross section with the confluent empty cells shaded. The spongy mass (e) on the back of the fertile cell clearly separates this naturalized foreigner from our native species. It differs also

in its more humble and diffuse habit, and the

pale blue color of its corolla.

Fedia Fagopyrum Torr. and Gray. Fig. 103. Fruit (from West Penn.); a, side view; b, cross section, with the two empty cells shaded. Despite the smaller number of stigmas, the structural plan of the ova

ry, as seen in the five well-de

a

Fig. 103.

C

Fig. 105.

fined dorsal sutures (103 b, 8), is quinary. A Fedia Fagopyrum. single ovule is developed and fills up the cavity of the three posterior confluent cells. The two anterior sterile cells are compressed laterally, until they almost meet in a sharp angle, making the fruit triquetrous like a grain of buckwheat. Between the sharp edges of the angle a narrow groove (103 b, a) runs from base to apex. In a considerable number of matured fruits examFig. 104. ined, from W. Penn. and W. N. York, this groove was found uniformly present. All, too, were more or less downy under a lens, and in no case were the sterile cells confluent. These are variations from the typical plant as characterized in Gray's Manual, and yet the peculiar shape of the fruit and its large size (two lines in length) will probably enable it to hold its place as a distinct species.

b

с

Fedia

a

b

F. radiatu, var.

Fedia radiata Michx. Fig. 104. Fruit; patellaria. radiata. a, side view; b, cross section, with the two empty cells shaded; c, cross section of another fruit, with the two empty cells confluent. The fruit of this species is much smaller, about a line in length, and usually quite downy, but sometimes smooth. The quinary structure of the ovary is not so apparent. As in all these Fedias the bracts are more or less strongly ciliated, or perfectly

naked. In one particular the description should be amended. Under favorable circumstances it often attains the height of thirty inches, and its range of stature is about that of F. Fagopyrum, one to two feet.

Fedia radiata Michx., var. patellaria (F. patellaria Sulliv.). Fig. 105. Fruit (from Columbia, Pa.); a, side view; b, cross section, with the two slightly

a

Fig. 106.

b

divergent empty cells shaded. This small form varies but little in size a and shape from the fruit of genuine F. radiata as seen in Fig. 104, a and b, and appears to have been derived from it by a moderate extension of the F. radiata var. patellaria. walls of the empty cells. Fig. 106. Fruit (from Columbia, Pa.); a, side view; b, cross section, with the two widely di- F. radiata, var. patellaria. vergent empty cells shaded. Here the abnormal lateral extension of the walls of the empty cells is carried to an extreme, and they are so flattened in the centre and curved up on the margins as readily to suggest the image of a miniature platter. This is exactly the form of fruit in Mr. Sullivant's plant in Dr. Gray's herbarium. Fig. 107. Fruit (from Columbia, Pa.); a, side view; b, end view above; c, cross section, with the empty cells shaded. One specimen of

d

a

Fig. 108.

F. radiata, var. umbilicata.

b

Mr. Knipe's last collection has this remarkable form of fruit throughout. It seems to have been produced by the doubling of that represented in Fig. 106. Two fruits have coalesced by the union of their anterior empty cells, and the dissepiments vanishing have left a single large cell in the middle. On one side the usually fertile cell is empty; on the other, it contains a seed but in some cases all the cells are sterile.

Fedia radiata Michx., var. umbilicata (F. umbilicata Sulliv.). Fig. 108. Fruit (from Columbus, Ohio); a, side view; b, another side view, showing the cruciform opening caused by the tendency of the cell in the abnormal expansion of its walls to split along the sutures; c, cross section of the same; d, side view of a more mature fruit, showing a further enlargement of the opening into the empty cell; e, another side view. As the fruit of the former variety came probably from that of F. radiata, with two empty cells, as seen in Fig. 104 b, so this may have been derived, by the operation of the same cause, from that of Fig. 104 c, with the empty cells confluent.

In view of the decided disposition toward monstrosity evident in Fig. 107, and the differences of the fruits in size and shape, it is questionable whether F. patellaria and umbilicata are worthy to stand even as varieties of F. radiata; but, since no typical fruits of the latter have been observed intermingled with the aberrant forms on the same stalk, they may for the present be recognized as such.

MIMICRY IN THE COLORS OF INSECTS.

BY DR. H. HAGEN.

HAVING observed that in treating of the interesting phenomena of mimicry, writers have used indiscriminately, very different factors, I shall try to give some preliminary ideas which I do not find published, and which I believe will be useful in explaining this interesting subject.

It will be best to consider the color and pattern separately. There are three different kinds of colors: viz., colors produced by interference of light, colors of the epidermis, and colors of the hypodermis. All three may either be wanting, or all three, or two of them may occur together in the same place.

Colors produced by interference are produced in two different ways; first by thin superposed lamellæ, as in the wings of Diptera, Neuroptera, etc., without any other color, as in hyaline wings, or connected with other colors as in the scales of Entimus and others. There must be at least two superposed lamella to bring out

colors by interference, and there cannot be more than four, as both wings and scales consist only of four layers, two internal belonging to the hypodermis, two external belonging to the epidermis. In fact, if scales taken from dry specimens of Entimus are observed under the microscope, many partly injured can be found, which give different colors according to the layers of the lamella which remain.

Secondly, colors by interference are produced by many very fine lines or striæ in very near juxtaposition, as in Apatura and other color changing insects. Colors by interference may perhaps be sometimes also produced in the same way as in the feathers of the dove's neck by very small impressions situated near together.

The colors produced by the interference of light are only optical phenomena, differing in this respect from the other colors of the body, the epidermal and hypodermal colors.

The epidermal colors belong to the pigment deposited in the cells of the chitinized external skin, the epidermis. These colors are mostly metallic blue, green, bronze, golden, silver, black, brown, and perhaps more rarely red. The epidermal colors are very easily recognized, because they are persistent, never becoming obliterated or changed after death.

The hypodermal colors are situated in the non-chitinized and soft layer, called hypodermis by Weismann. They are mostly brighter and lighter, light blue or green, yellow, milk white, orange and all the shades between. The hypodermal colors in the body of the insect fade or change, or are obliterated after the death of the A fresh or living insect when opened may easily be deprived of the hypodermal colors simply by the action of a little brush. I said hypodermal colors in the body, because there are hypodermal colors which are better protected, being encased nearly air-tight, and therefore are more easily preserved even after the death of the insect. I refer to the colors in the elytra and wings, and in their appendages, the scales. The elytra and the wings are, as is well known, at first open sacs in communication with the body, of which they are only the extension: of course they are formed of the epidermis and hypodermis which become so strongly glued together after the transformation into the imago state that a maceration of years tried by me showed no effect at all on such wings. This fact is very interesting as it explains how wings, and even colored wings, can be found in palæontological