A NATURALIST'S WANDERINGS
none, it continues to rotate through about 70A more, till it comes into contact with the face of the column, that is with the stigmatic cavity, which is very large, broad and full of viscid matter (Fig. 6). The whole surface of the lower four pollinia come into contact with the viscid matter and sink well into it, while the viscid matter finds its way gradually about all of the pollinia. The inner members of the upper row of pollinia sometimes escape this inundation, but it seems of little avail to the plant for its cross-fertilisation, for they remain throughout covered by the anther-cap. The tips of the caudicles, however, remain in most cases unaffected throughout, but I have found it difficult to remove any of their pollen grains. The inundated pollinia have no obstacles to bar the way of their tubes to the ovary. On clearing out with a blunt instrument
F G. 7.a PHAJUS BLUMET, SAME AS FIG. 6, FIG. 8.a PHAJUS BLUMET, SHOWING EXTRA WITH ANTHER-CAP MERELY DOTTED ANTHER, H ; A, B, C, AS IN PREVIOUS
IN ; A, B, C, AS IN PREVIOUS FIGURES. FIGURES.
the swollen pollinia from the stigma, it can be seen that from nearly the top of the column, along the posterior median line, a prominent ridge (Fig. 3, p. 86) runs down almost to the ovarium. In the light afforded by the dissection of an Arundina speciosa (to be mentioned below) this would appear to represent the absent rostellum. Large seed-capsules were produced by every flower of this form. This Phajus is also remarkable for producing, at times two, supernumerary anthers on the top of the column one on each side of the normal anthers (Fig. 8).
Here then we have an orchid whose flowers present every attraction to insects to pay at least a first visit (when they would find no nectar), all of them gay, with a nectary, and a beautifully painted and finger-posted labellum, yet rarely possible to be anything but self-fertilised.