T-ransaclioHd. 21 



realised that we are probably indebted to insects for the existence 

 of this particularly beautiful group. It seems to nie difficult to 

 iind cuiy other reason for CorulUjiora than their being able to 

 reserve their honey for longlipped insects, which also happen to be 

 the most intelligent and industrious pollen carriers. In other 

 ways a tubular liower is a disadvantage, as material is needlessly 

 wasted in the tube itself in the strong supporting calyx and so 

 on. One might even, I think, trace the variation which led to the 

 formation of tubular flowers a little further back ; the petals in 

 a minute forming Hower consist of four or live small pimples of 

 jelly-like substance which are aiTanged in a circle. If these little 

 bulging pimples were arranged closely side by side, they would 

 be likely to run together and rise as a single rim or cylinder 

 instead of as separate projections. If this is true, it exi)lains why 

 Ijjchids and sileue which are tubular Howers, from an insect's 

 point of view, are not CoroUi/lorct', for we tind in these forms that 

 the position of the stamens and nectaries would prevent this 

 fusion. At any rate, when a tube of this kind was once pro- 

 duced even in a rudimentary condition (such as we tind in the 

 holly and bryony), the advantage in retaining the honey and 

 preserving it for the best insects would be so great that it would 

 be immediately seized upon and improved. 



A possibility of indefinite variation was thus atibrded, and the 

 variations that actually have occurred are so numerous that it is 

 somewhat ditHcult to classify them. I think one iuay, however, 

 ti-ace three distinct types, under which probably 90 per cent, of 

 the CoroUiJiorcf niay be placed : — 



1. Flowers with a widely open corolla very n)uch like a large 

 Thalamijlor whose petals have united — Campanula, Convolvulus, 

 and the Foxglove. 



2. Flowers with a very narrow tube which ends in a spreading 

 horizontal limb — Primrose, Periwinkle. 



.3. Flowers with distinct upper and lower lip and a short or 

 moderately long txih-i— Lobelia, Salvia, and almost all the Labiatess. 



Now every one of these very widely spread types can, I think, 

 be shown to be directly adapted to the shape of the insects which 

 visit them. Thus a Foxglove is almost exactly the shape of a 

 bumble bee's body, and I think we are quite justified in saying 

 that the bee has fashioned the shape of the Foxglove flower 

 exactly as the thumb of an old glove affords the exact pattern of 



