Charles Robertson

The Structure of the Flowers and the Mode of Pollination of the Primitive Angiosperms

THE purpose of this paper is to show how some questions involved in the phylogeny of angiosperms might be regarded from the point of view of the anthoecologist. The proposition to be offered is that the primitive angiosperms were entomophilous, and that the anemophilous ones are metamorphosed entomophilous flowers whose seemingly simple structures are degraded, not primitive. lIhe characteristic anemophilous flower has a pistil with a one-celled ovary containing a single ovule and ripening into a one-seeded indehiscent fruit. The characteristic entomophilous flower has a pistil with several ovules which ripens into a several-seeded dehiscent fruit. l1he explanation of the monospermy of the Anemophilae and the polyspermy of the Entomophilae is obvious when one considers the difference in the conditions of pollination. Even with a large plumose stigma the Anemophilae have enough to do to secure enough pollen for a single ovule. The Entomophilae with a small stigma can readily secure enough pollen for many ovules. The relation of many-seeded fruits to entomophilous pollination is emphasized in the orchids, where the high degree of polyspermy is correlated with the disposal of the pollen in masses. There are conditions under which entomophilous flowers become monospermous, and anemophilous flowers derived from them may show monospermy as an inherited character. But polyspermy is an entomophilous character. Syncarpy involves polyspermy and is also an entomophilous character, the union of oligospermous fruits producing polyspermy. It is hard to imagine conditions under which an anemophilous plant would be likely to become syncarpous. Accordingly, in the absence of other entomophilous characters, and in spite of the presence of anemophily, plants may still retain the entomophilous characters of polyspermy and syncarpy. Syncarpy is important in diagnosing advanced stages of anemophilous metamorphosis because indications of syncarpy may be the only remaining evidence of a previous polyspermous condition. I also regard highly specialized inflorescences as entomophilous. At any

Insect Relations of Certain Asclepiads. II

On the ground of usefulness to Asclepias, butterflies may be thrown into three divisions. The Papilios are the most useful, and pollinia were found on all of our species except Ajax. They suck with their wings in motion. Using their legs to offset the motion of the wings, they rapidly repeat those movements which are necessary to draw the pollinia from the anther cells and insert them into the stigmatic chambers. Other large butterflies, like Danais, hold their wings still in sucking, spending moretime on an umbel, but generally carrying pollinia. Small butterflies are worse than useless. They remain long on the umbels, sucking, but resting their feet superficially on the flowers, and seldom effecting pollination. Of twenty-three species bearing pollinia, only three are smaller than Chrysophanus Hypophleas; while of twelve species on which no pollinia were found, with the exception of Eudamus Tityrus, which was not caught, the Chrysophanus is the largest. As stated before, butterflies have corpuscula on their claws less frequently than Hymenoptera. Since several moths were found entrapped on the flowers, pollination must often be brought about by night-flying Lep-

Flowers and Insects. XIX

I. Comparison of the genera of bees observed in Low Germany and in Illinois, with the number of species of each and their flower visits.-The results credited to Muller are taken from the Ferfilization of Flowers. They are based on observations made by Herr Borgstette at Teklenburg, in the north of Westphalia, and by Muller at Lippstadt and in Sauerland, in the central and southern parts of the same region, as well as observations made by him in Thilringia. My results are based on observations made within ten miles of Carlinville. Each species of bee is credited with a visit for each of the species of plants on whose flowers it has been taken.

Flowers and Insects. I

Delphi1ziurt tricorne Michx.-The flower agrees in most respects with D. elatum, as described by Muller.2 It is blue, but the exposed parts of the two upper petals which arch over the entrance to the spur are white, forming a sure guide to the nectar. In D. Ajacis, according to Sprengel (702),3 the upper petals also form the pathfinders by a variation in color. In D. elatum yellow hairs on the lateral petals form the guides, while in D. Consolida pathfinders are wanting (Miiller). The lower petal has disappeared, since its attractive function has been usurped by the sepals. It is unnecessary as a protection to the stamens and pistils, and its presence in the median line would only prevent proper contact with the anthers and stigmas. The parts whose function has been most imperfectly explained are the two lateral petals. These close over the numerous stamens, completely hiding them, but an entrance to the spur is left between them and the upper petals. When a bee visits the flower, the lateral petals are forced aside, and the under side of the bees head comes in contact with the

Flowers and Insects. XVI

NOTHOSCORDUM STRIATUM Kunth. N. ornithogaloides (Walt.) Kunth.-The plant is common in woods, blooming -from April ioth to May i6th. The scapes grow I or 2dm high and bear small umbels of white flowers. The flowers are about iomm long and expand IO or 12mm. The sepals are approximated below, the base of the tube being greenish and narrowed by the ovary and the filaments of the six stamens. The flowers are homogamous, the stigma occupying the center of the circle of anthers and somewhat surpassing them. Spontaneous self-pollination can hardly occur. The flower is remarkable for being abundantly visited by numerous species of bees of the genus Nomnada. On seven days, between April 20th and May 9th, I captured the following visitors:

Flowers and Insects. XIII

DODECATHEON MEADIA L.-American cowslip, shootingstar. In his arrangement of floral mechanisms Delpino (2) recognizes a class of pendulous nodding or horizontal flowers upon which the visitors cling (apparrecchi prensili), which he divides into the borrage and the verbascum types (tipo borragineo, tipo verbascino). The former includes such flowers as Dodecatheon, Cyclamnen, and Solanumn, which the bees clasp in such a way as to receive the pollen upon the undersides of their bodies, and the latter contains flowers like Verbascumn and Tradescantia, which are provided with hairs which afford a foot-hold. Kerner (I) regards the reflexed petals as facilitating access to nectar and pollen, and this is true as regards the legitimate visitors. Intruders are much more effectually excluded than would be the case if the petals were less strongly reflexed, as in the flowers of Erythronium.2 The approximation of the anthers in a compact cone also gives the flower a signal advantage over such a flower as Ribes gracile,3 whose stamens being of the ordinary form permit the visits of a number of insects which the flower cannot utilize. The reflexed petals also render the nodding flowers much more conspicuous and attractive than they would be if the expanded petals faced the ground. Loew (3) has figured and thoroughly described Dodecatheon integrifoliun Michx. (=D. Meadia L.), D. 7effreyi Moore (=-D. Meadia var. iancifoliumn Gray), and an unnamed species from material growing in the Berlin Garden. To this I shall add an account of D. Meadia as observed under natural conditions in Illinois. The plant is common in prairies and open woods, where it grows in rather large patches. The scapes rise from 3 to 6dm and bear numerous, handsome flowers, which are white or rose color. The corolla has a short tube, which for

Flowers and Insects. VIII

basales 0.017 x o.025 mm., trigonis mnagnis acutis. Lobulus diametro caulis duplo longior, decurrens, carina arcuata sinu lunato in folium excurrente, apice exciso-truncatus, angulo brevidentato, ceterum valde convexus, margine supero involuto. Ainphigastria caulina ovata, caule plus duplo latiora, ad medium fere bifida, sinu angusto laciniis acutis. Flores feminei in caule ramisque pseudolaterales, raro in angulo furcarum. Folia floralia caulinis minora, arcuatim divergentia, e basi angusta falcato-oblonga, lobulo lanceolato profunde soluto, acuto. A mphigastriumn florale foliis suis aequimagnum, oblongum, ad 3 incisum, rima angusta, laciniis muticis. Perianthia et androecia ignota. HAB.: Florida (Underwood). Lejezinea Caroliniana monoica est. Lejeunea Austini cellulis multo minoribus gaudet. A remarkable feature in this plant is the large incrassations at the angles of the cells, which form very distinct triangles with acuminate points. Kaiser Wilhel/n str. 9., Leipzig, Germany.

Flowers and Insects. XIV

GENTIANA PUBERULA Michx.-From the abundant observations on European species of Gentianua it appears that most of the species which have been investigated are proterandrous, though several are homogamous and a few proterogynous. Most of them are adapted to bumblebees, many to Lepidoptera, while quite a number are intermediate, being visited by both kinds of insects. One species, G. lutea, has exposed nectar, and is visited by a miscellaneous list. Nothing has been done with our species, except G. crinitc and Andrews.i. In the case of G. Andrewsii, BeaL (6) observed that it was visited by bumblebees, but overlooked the proterandry, supposing that cross-pollination was favored by the stigma standing, far above the anthers. A statement of Meehan, that the flower never opens, evidently taking it for granted that it is never visited by insects, is quoted by Henslow (I2) in spite of Beals observations. Vausenburg (Io) objects to Beals conclusions, and supposes that the stigma is pollinated as it passes the anthers. Kunze (i8) regards the flower as cleistogamous, the nectar being of no significance. Bailey (I7) records that nectar is secreted by the walls of the corolla. Gray (I9, 2I, 25) states that the flower opens a short time in sunshine, which I have never observed; notes the proterandry and that spontaneous self-pollination may finally occur by the lobes of the stigma curling back until they touch the anthers. Finally I have shown the adaptation to bumblebees and have recorded the abundant visits of Bomnbus americanorum F. 8W (4i). According to Beal (6) G. crinita is visited by bumblebees and resembles G. Andrewsii, of which, however, as we have noted, he had failed to recognize the proterandry. G. Andrewsii and puberitla, the only species I have found in my neighborhood, are the very latest of the bumblebee flowers, the former beginning to bloom by September 14th, and the latter on the 27th, both running nearly through October. (?entzana pitberula has the stem terminated by a cluster of

Flowers and Insects. XII

CLEMATIS VIRGINIANA L.-The flower, with its horizontally expanded sepals, measures about 2cm across. The flower clusters form large, white masses upon the shrubs upon which the plants climb. The flowers are dioecious. The white color and the easily accessible nectar attract numerous, mostly small, short-tongued insects. The visitors, however, are principally flies, three-fifths of the species being of this kind. An interesting comparison may be made between this plant and Isopyrum biternatum, an account of which is given in the GAZETTE, 17: I73-5. I892. The flowers of both species are white, though the Clematis is more conspicuous, and the extent of nectar concealment is almost identical. Nevertheless, the species show a marked difference in the kinds of insects visiting them, as the following table will show:

Flowers and Insects. X

and we find that the cortex cells are round and have many spaces between the cells, while fig. i9 represents the roots of Actaca alba, which grows on dry banks, and here the cortex cells are angular and the spaces between cells are almost none. The differences in vascular development are also well shown in these figures. In this part of the paper I have not discussed the general structure of the species studied as regards the occurrence and distribution of the histological elements, since I find that Marie, Hegelmaier, and others have described the general structure of many species of Ranunculacee, and in many cases my descriptions would have been but a repetition of theirs.