always the simple structure which it looks to be from the outside. That is, it may be either simple

A compound pistil is one which

more united

bears evidence of containing two or parts or units. The common test of a compound pistil is the presence of more than

but this is not

one locule,

always designative, for in some cases false partitions partitions grow out from the walls into the cavity of the ovary. The presence of more than one style to a single ovary also indicates a compound pistil; and, more especially, the Occurrence of more than one placenta. The separable compound pistil are known as carpels. The theory of a compound pistil is that

FIG. 145.

units or parts in a

it is made up of the union of two or more simple pistils.

159a. Thus the hepatica has one carpel, the tulip has three, the mustard has two, the catnip has two 2-lobed carpels, the apple has five, and even the unilocular cockle (Fig. 143) is thought to be 5-carpelled because of the five styles (two being cut away in the figure) and of certain peculiarities in related plants; that is, there is evidence that some plants which were once 5-loculed are now 1-loculed because of the loss of partitions; and sometimes this elision can be traced in the different ovaries of a single plant.

1596. A flower, therefore, may contain one simple pistil, several simple pistils, or one compound pistil; and there are instances in which it contains more than one compound pistil.

SUGGESTIONS.-When taking up

for the pistil.

The ovary is the

any unfamiliar flower, look first best distinguishing mark, for the pistil is often much disguised. Determine what relation exists between the numbers of stigmas, styles, or locules in any pistil. Also observe the number of ovules, and the placentæ.

XXVI. THE STAMENS

160. The most striking feature of the stamens in the flowers which we have seen is the great difference in length and shape. Most of the stamens are slender, and have prominent stalks or filaments; but the anthers of the currant (Fig. 146) are nearly sessile, and in some flowers they are completely sessile. It is, therefore,

K

FIG. 146.

Flower of garden

currant.

apparent that a filament is not essential to a stamen any more than a petiole is essential to a leaf.

wise employed (156a). As no confusion has arisen from the application of the word cell to both pistils and stamens, none may be anticipated from a like use of locule. It has been suggested

to use locellus (diminutive of loculus) for the anther compartment, but it seems to be unnecessary to introduce another word, and, moreover, locellus has no accepted anglicized form (although it might be shortened to locel).

162. The anther of the tulip and willow is attached by the base to the very top of the filament, but that of the water-lily (Fig. 147) seems to be joined to the filament in its entire length. The mustard and the lily (Fig. 148) show still a third method, the anther being poised by attachment to its back, and standing cross-wise the filament. These three methods, with numerous intergradations, will impress the pupil, if he were to examine numbers of flowers, as being the types of the ways in which the anther is

FIG. 150.

Sensitive stamens of barberry, showing a single flower, and the dehiscence of the anthers at a and d.

FIG. 149.

Pores in

azalea

stamen.

look inward (towards the pistil), or are said to be introrse; those of the lily look outwards, or are extrorse. The pupil should determine if innate and adnate anthers differ in this regard, also.

164. The anthers of the mustard and the tulip seem to open along the side of each locule. The azalea, however (Fig. 149), opens by a hole or pore in the tip of the locule. Heaths and huckleberries open in the same way. We should examine