PARTHENOGENESIS 133 hatched aphis by conveying it upon a twig be- neath a glass shade dipping into water. Of fourscore offspring produced alive by this in- sect, one was isolated in like manner, and with similar result; and this was repeated as long as the observations continued, or for nine successive broods. As the young aphides are ready for propagation in about two weeks, it follows that in the course of a summer a single parent may have a progeny of millions, and all without renewed intervention of the male ele- ment. Kyber found that when warmth and food were abundantly supplied, this agamic production would go on for two or three years ; but these broods, winged or wingless, consist almost wholly of imperfect females, seldom any males. The true females, always wingless, produce only after sexual union, and then eggs, not living offspring. And ordinarily, as the cold of autumn increases and the supply of food fails, the agamic young give place to true males and females, the latter laying eggs which the next spring hatch out again viviparous or imperfect females. Thus there is a cycle of changes ; a large but varying number of links of non-paternal, being interposed between any two of paternal generation. The imperfect females have, in place of ovaries, certain tubu- lar organs, the germs lying in which develop into living insects. Thus the case is only ap- parently, not really, anomalous ; the real indi- vidual of the aphides is the perfect male or fe- male only, and union of these must occur for the perpetuation of the race ; but under favor- ing conditions, by a sort of exuberance of vital activity, an intercurrent production by gem- mation or budding sets in, terminating finally in a return to the normal individual. Accord- ing to this view, the drone bees are another instance of production by budding ; and still others are said to be found in the dapJinice (water fleas), and in some species of butterfly. In plants, the occurrence of parthenogenesis, the development of an embryo in the ovule, and the production of perfect seed without the agency of the pollen or male element, was maintained in the last century by Spallanzani, who cited hemp and spinach as plants, among others, in which this took place. Since then the subject has been discussed by botanists, in- cluding some of the most eminent of the pres- ent day, without very decisive results ; as ex- periments by different observers upon plants of the same kind have led to decidedly opposite conclusions, the question of the occurrence of parthenogenesis cannot be regarded as settled. The great difficulty attending experiments on hermaphrodite or bisexual plants has led ob- servers to use those with separate sexes, and monoecious, or more generally dioecious plants, have been selected. A euphorbiaceous shrub from Australia, ccelebogyne (now alcTiornea) ili- cifolia, produced in Europe female flowers and perfected seed, while no male plant was known to be in the country ; the plant was supposed to be perfectly dioecious, neither male flowers nor stamens being detected, and the production of fertile seeds in this case was regarded as proof that, in this plant at least, the presence of pol- len was not necessary to their formation and development. In 1857 Baillon asserted that he had found a stamen in one of the female flow- ers of ccelebogyne, but this was denied by De- caisne, who asserted that Baillon had mistaken a glanduliferous bract for a stamen ; in 1860 Karsten announced that he had discovered two hermaphrodite flowers upon the plant, in the Berlin botanic garden, between May and Au- gust, which was regarded as sufficient to account for the fruiting. It is said that figs developed in summer contain no male flowers, yet the pistils of these produce seed containing an em- bryo ; but both kinds of flowers in the fig are exceedingly small, and being enclosed within the hollow receptacle, accurate observation is surrounded with difficulties. The experiments of Naudin and Decaisne (Paris) with hemp were conducted with female plants, some in the open air surrounded by a high fence, and others in pots placed in a room in the second story of the house ; no male flowers could be discovered on these plants, yet all bore fruit, and the female plants from these seeds, simi- larly isolated, ripened seeds also. On the oth- er hand, Regel of St. Petersburg, in experi- menting upon spinach and mercurialis, which Naudin and Decaisne had cited as giving seeds upon the female plant when isolated, cut back his specimens of these in order to re- duce the number of flower clusters, and found that in every instance the female plants thus treated produced more or less male flowers, very much reduced and stunted, but with stamens which produced pollen, though the flowers containing them were so insignificant that they might have been unnoticed had not great care been taken in the search. Another instance cited by Naudin and Decaisne is bryo- ny, a dioecious plant of the gourd family ; the pistillate plants of this, from which access of pollen was carefully shut out, produced fruit in the greatest abundance; in 100 of these fruits 12 had no seeds, 45 had one seed, 29 two seeds, 11 three seeds, two had four, and one had five seeds. These illustrations are sufficient to show the difficulties in determining whether perfect seeds are formed without the influence of pollen upon the ovule. Besides the fact that male flowers may sometimes be developed upon female plants, and thus clandestinely supply pollen, there is another which must be taken into account: in flowers of separated sexes rudiments of the organs of the other sex are often distinctly seen ; in the staminate flower, a knob or protuberance stands in the place of the pistil, and in pistillate flowers we have the places of the stamens occupied by glands, or abortive filaments, as if one or the other series of organs had been suppressed to make the flower male or female. The many well known instances in which a plant produces all three kinds of flowers, staminate, pistillate, and per-
