the Gladiolus and Alstroemeria, where they are very numerous and follow the rule of commencing to emerge in the middle, and then proceed upwards and downwards. Though parietal placentas seem generally to have their ovules developed from below upwards, yet, as seen above, it is not uncommon with an axile placentation. If any interpretation be sought, I should feel inclined to associate it somewhat with a more primitive state of things, since a parietal placentation presents a more rudimentary character than an axile. But why they are developed thus, sometimes upwards, sometimes downwards, or both ways at once, is at present as inexplicable as the fact that leaves develop both basipetally and basifugally, either in their entirety, or as to their lobes and notches, which may be formed on either plan. Perhaps there may prove to be a common cause for both.

CHAPTER XXII.

HETEROGAMY * AND AUTOGAMY.

PROTANDRY, PROTOGYNY, HOMOGAMY, AND CLEISTOGAMY.-These conditions prevail in nature in varying degrees of frequency. The first is common to all conspicuous flowers habitually visited by insects, and is accompanied by heterogamy. The fact that anthers mature their pollen before the stigmas of the same flower are ready to receive it, is due to the extra stimulus given to the androecium, which mostly effects simultaneously the enhancement of the corolla or perianth which attracts the insects (see p. 191). Like everything else in nature, it is very far from being absolute, and any flower may be protandrous at one time or place, while it may at another mature the essential organs together, and then it becomes homogamous, or it may be even protogynous.

These latter conditions prevail in less conspicuous flowers and all those which are fluctuating between a condition * Heterogamy, i.e. union by intercrossing different flowers. Autogamy, i.e. union by self-fertilising one and the same flower. Protandry, i.e. stamens maturing the pollen before the stigmas of one and the same flower are ready to receive it.

Protogyny, i.e. pistil maturing the stigmas before the pollen of one and the same flower is shed.

Homogamy, i.e. pollen and stigmas of one and the same flower, maturing simultaneously.

Cleistogamy, i.e. autogamous within an unopened perianth.

requiring insect agency and self-fertilisation or autogamy; as well as in the majority of flowers which are too inconspicuous to invite insects at all, or which never expand. The series of such flowers terminates in perfect and perpetual cleistogamy.

The first condition, or Protandry, does not now require special discussion or illustration; as it is the prevailing one in most conspicuous flowers: though it must be distinctly borne in mind that the exceptions are rare in which a flower cannot fertilise itself at some period or other before it fades ; even though a large order, as Orchidea, may furnish many examples.

Protogyny may arise from several causes. Müller has mentioned about twenty species of plants irrespective of the Grasses which are more or less decidedly protogynous; and what one notices is that many are Alpine species of genera which have other species dispersed elsewhere that are homogamous or protandrous. Thus Anemone alpina is protogynous, but A. Narcissifolia is protandrous. Ranunculus montanus, R. parnassifolius, R. pyrenæus are all protogynous. These may be compared with the smaller-flowered forms of R. aquatilis which are homogamous; but R. flammula, R. acris, R. repens and R. bulbosus are protandrous with the outermost stamens only. Thus, this genus supplies a progressive series. Other protogynous and mountain species are Dryas octopetala, species of Saxifrage, as S. androsacea and S. muscoides, and S. Seguieri: but Müller found S. oppositifolia and S. tridactylites to be sometimes feebly protandrous, at others protogynous. On the other hand, S. rotundifolia, S. aizoides, etc. are protandrous. Loiseleuria procumbens, Trientalis Europæa, Bartsia alpina, Hutchinsia alpina, and Thalictrum alpinum are all protogynous.

Secondly, a group of plants, the flowers of which have

the habit of blossoming early, as in the spring or the beginning of the summer, are protogynous; such are species of Hellebore, Prunus, and Crataegus, as well as the Horse-chestnut and Mandragora vernalis.

Some species are characterized by the habit of living in shady places, as Geum urbanum and G. rivale, Chrysosplenium oppositifolium, Gagea lutea, Paris quadrifolia.

Lastly, others have minute flowers, as Geranium pusillum, Veronica serpyllifolia, Toffieldia, and many other species, some of which I have mentioned when treating of the emergence and development of the floral whorls, where I have explained the cause. *

Wind-fertilised or anemophilous flowers are for the most part protogynous; for these flowers have been accompanied by strong degeneracy of the corolla and pollen, while all traces of nectariferous structures are almost invariably and entirely suppressed. Hence Thalictrum minus, Poterium, Sanguisorba, Plantago sp., Callitriche, Myriophyllum, Artemisia, Chenopodium, Amentiferæ, Juncaceae, and Gramineæ are all more or less characterized by being protogynous while they are anemophilous as well.

If we are not in a position to trace the actual causes of protogyny in every instance, we can at least see several influences which can bring it about. Temperature will be seen hereafter to be a most potent one; for a relatively lower temperature very frequently checks the energy of the corolla and stamens, without having any necessarily corresponding effect on the pistil, and several compensating processes then come into play; so, conversely, the pistil now gains the ascendancy and can mature first. This, therefore, will *See Chaps. XX. and XXI.

Intercrossing by insects may be recovered in anemophilous flowers; when honey may be again secreted, as in Salix capræa and Sanguisorba officinalis; see Fertilisation, etc., p. 236, fig. 77.

account for some mountain species, as well as those blossoming early or in shady places, being protogynous.

It must not be regarded as universally true. If flowers so situated or circumstanced be abundantly visited by insects, they will respond to their influence; and the consequence is, that many Alpine plants are even strongly protandrous, as well as spring-flowering plants and some which grow in shady places, as Sanicula Europaea, Odontites serotina, etc. It is when we compare the protogynous species with others of the same genus, that the influences of a lower temperature, shade, etc., more especially suggest themselves as true causes of protogyny in some species, while others may be homogamous or protandrous.

Many plants normally provided with conspicuous flowers, but accidentally growing in shady places, may often be found having them half opened or as quite closed buds, and yet fully fertile. The same occurs late in the season, when the flowering period is drawing to a close. Such flowers represent the preliminary stages leading to a permanently homogamous or protogynous condition, as the case may be, which are mostly autogamous as well.

Whatever may be the direct cause, and there may be others besides those I have mentioned, protogyny is easily brought about temporarily in individuals, or it may become hereditary and a permanent feature.

It need now hardly be added that, before protogyny is reached and emphasized, all degrees of passage can be met with from strong to weak protandry; then homogamy is acquired: and, after passing through oscillating conditions, permanent protogyny can be finally the result.

Many individual plants vary in this respect, being sometimes or in some places in one condition, and at other times and in other places in another condition. As nothing is