have a spiral, annular, or reticulated thickening of the wall. The endothecium varies in thickness, generally becoming thinner towards the part where the anther opens, and there disappears entirely. The walls of the cells are frequently absorbed, so that when the anther attains maturity the fibres are alone left, and these by their elasticity assist in discharging the pollen. The anther is developed before the filament, and is always sessile in the first instance, and sometimes continues so. It appears at first as a simple cellular papilla of meristem, upon which an indication of two lobes soon appears. Upon these projections the rudiments of the pollen-sacs are then seen, usually four in number, two on each lobe. In each a differentiation takes place in the layers beneath the epidermis, by which an outer layer of small-celled tissue surrounds an inner portion of large cells. Those central cells are the mother-cells of the pollen, whilst the small-celled layer of tissue external to them becomes the endothecium, the exothecium being formed from the epidermal layer.
In the young state there are usually four pollen-sacs, two for each anther-lobe, and when these remain permanently complete it is a quadrilocular or tetrathecal anther (fig. 70). Sometimes, however, only two cavities remain in the anther, by union of the sacs in each lobe, in which case the anther is said to be bilocular or dithecal. Sometimes the anther has a single cavity, and becomes unilocular, or monothecal, or dimidiate, either by the disappearance of the partition between the two lobes, or by the abortion of one of its lobes, as in Styphelia laeta and Althaea officinalis (hollyhock). Occasionally there are numerous cavities in the anther, as in Viscum and Rafflesia. The form of the anther-lobes varies. They are generally of a more or less oval or elliptical form, or they may be globular, as in Mercurialis annua; at other times linear or clavate: curved, flexuose, or sinuose, as in bryony and gourd. According to the amount of union of the lobes and the unequal development of different parts of their surface an infinite variety of forms is produced. That part of the anther to which the filament is attached is the back, the opposite being the face. The division between the lobes is marked on the face of the anther by a groove or furrow, and there is usually on the face a suture, indicating the line of dehiscence. The suture is often towards one side in consequence of the valves being unequal. The stamens may cohere by their anthers, and become syngenesious, as in composite flowers, and in lobelia, jasione, &c.
The anther-lobes are united to the connective, which is either continuous with the filament or articulated with it. When the filament is continuous with the connective, and is prolonged so that the anther-lobes appear to be united The connective. to it throughout their whole length, and lie in apposition to it and on both sides of it, the anther is said to be adnate or adherent; when the filament ends at the base of the anther, then the latter is innate or erect. In these cases the anther is to a greater or less degree fixed. When, however, the attachment is very narrow, and an articulation exists, the anthers are movable (versatile) and are easily turned by the wind, as in Tritonia, grasses (fig. 105), &c., where the filament is attached only to the middle of the connective. The connective may unite the anther-lobes completely or only partially. It is sometimes very short and is reduced to a mere point, so that the lobes are separate or free. At other times it is prolonged upwards beyond the lobes, assuming various forms, as in Acalypha and oleander; or it is extended backwards and downwards, as in violet (fig. 71), forming a nectar-secreting spur. In Salvia officinalis the connective is attached to the filament in a horizontal manner, so as to separate the two anther-lobes (fig. 72), one only of which contains pollen, the other being imperfectly developed and sterile. The connective is joined to the filament by a movable joint forming a lever which plays an important part in the pollination-mechanism. In Stachys the connective is expanded laterally, so as to unite the bases of the anther-lobes and bring them into a horizontal line.
 |
 |
 | ||||
| Fig. 71. | Fig. 73. | Fig. 74. | ||||
| ||||||
|
Fig. 71.—Two stamens of Pansy (Viola tricolor), with their two anther-lobes and the connectives (p) extending beyond them. One of the stamens has been deprived of its spur, the other shows its spur c. | ||||||
|
Fig. 72.—Anther of Salvia officinalis. lf, fertile lobe full of pollen; ls, barren lobe without pollen; e, connective; f, filament. | ||||||
|
Fig. 73.—Stamen of Lady’s Mantle (Alchemilla), with the anther opening transversely. | ||||||
|
Fig. 74.—Stamen of a species of Nightshade (Solanum), showing the divergence of the anther-lobes at the base, and the dehiscence by pores at the apex. | ||||||
|
Fig. 75.—The stamen of the Barberry (Berberis vulgaris), showing one of the valves of the anther (v) curved upwards, bearing the pollen on its inner surface. | ||||||
The opening or dehiscence of the anthers to discharge their contents takes place either by clefts, by valves, or by pores. When the anther-lobes are erect, the cleft is lengthwise along the line of the suture—longitudinal dehiscence (fig. 25). At other Antherdehiscence. times the slit is horizontal, from the connective to the side, as in Alchemilla arvensis (fig. 73) and in Lemna; the dehiscence is then transverse. When the anther-lobes are rendered horizontal by the enlargement of the connective, then what is really longitudinal dehiscence may appear to be transverse. The cleft does not always proceed the whole length of the anther-lobe at once, but often for a time it extends only partially. In other instances the opening is confined to the base or apex, each loculament opening by a single pore, as in Pyrola, Tetratheca juncea, Rhododendron, Vaccinium and Solanum (fig. 74), where there are two, and Poranthera, where there are four; whilst in the mistletoe the anther has numerous pores for the discharge of the pollen. Another mode of dehiscence is the valvular, as in the barberry (fig. 75), where each lobe opens by a valve on the outer side of the suture, separately rolling up from base to apex; in some of the laurel tribe there are two such valves for each lobe, or four in all. In some Guttiferae, as Hebradendron cambogioides (the Ceylon gamboge plant), the anther opens by a lid separating from the apex (circumscissile dehiscence).
The anthers dehisce at different periods during the process of flowering; sometimes in the bud, but more commonly when the pistil is fully developed and the flower is expanded. They either dehisce simultaneously or in succession. In the latter case individual stamens may move in succession towards the pistil and discharge their contents, as in Parnassia palustris, or the outer or the inner stamens may first dehisce, following thus a centripetal or centrifugal order. These variations are intimately connected with the arrangements for transference of pollen. The anthers are called introrse when they dehisce by the surface next to the centre of the flower; they are extrorse when they dehisce by the outer surface; when they dehisce by the sides, as in Iris and some grasses, they are laterally dehiscent. Sometimes, from their versatile nature, anthers originally introrse become extrorse, as in the Passion-flower and Oxalis.
The usual colour of anthers is yellow, but they present a great variety in this respect. They are red in the peach, dark purple in the poppy and tulip, orange in Eschscholtzia, &c. The colour and appearance of the anthers often change after they have discharged their functions.
Stamens occasionally become sterile by the degeneration or non-development of the anthers, when they are known as staminodia, or rudimentary stamens. In Scrophularia the fifth stamen appears in the form of a scale; and in many Pentstemons it is reduced to a filament with hairs or a shrivelled membrane at the apex. In other cases, as in double flowers, the stamens are converted into petals; this is also probably the case with such
