This peculiarity is well exemplified in the tribe _Gaudichaudieae_ of the order _Malpighiaceae_. A. de Jussieu, in his monograph, speaks of these flowers as being very small, green, destitute of petals, or nearly so, with a single, generally imperfect anther; the carpels also are more or less imperfect, but not sufficiently so to prevent some seeds from being formed. A similar production of imperfect flowers has been noticed in many other orders, _e.g._ _Violaceae_, _Campanulaceae_, &c. In some cases these supplementary blossoms are more fertile and prolific in good seeds than are the normally constructed flowers. M. Durieu de Maisonneuve alludes to a case where flowers of this description are produced below the surface of the ground. The plant in question is _Scrophularia arguta_, and it appears that towards the end of the summer the lowest branches springing from the stem bend downwards, and penetrate the soil; the branches immediately above the lowest ones also bend downwards, but do not always enter the earth. These branches bear fertile flowers: those which are completely below the soil are completely destitute of petals; those which are on the surface have a four-lobed corolla whose divisions are nearly equal, like those of _Veronica_.[369] To Sprengel, and specially to Darwin, physiologists are indebted for the demonstration of the relation of di- and trimorphic flowers to fertilisation. In certain genera of orchids, such as _Catasetum_, &c., flowers of such different form are produced that botanists, without hesitation, considered them as belonging to different genera, until the fact of their occasional production on the same plant showed that they were not of even specific importance. It was reserved for Mr. Darwin to show experimentally that these very different flowers are really sexual forms of one and the same species, ordinarily occurring on different plants, i.e. dioecious, but occasionally formed on the same spike. The same excellent observer has demonstrated that the di- and trimorphic forms of _Primula_, of _Linum_, _Lythrum_, and other plants--forms differing mainly in the relative length of the stamens and styles, are also connected with striking differences in the number of perfect seeds produced. The most perfect degree of fertility is obtained when the stigma of one form is fertilised by the pollen taken from stamens of a corresponding height. On the other hand, when the union is, as Mr. Darwin states, illegitimate, that is, when the pollen is taken from stamens not corresponding in length to the style, more or less complete sterility ensues in the progeny, sometimes even utter infertility, such as happens when two distinct species are crossed, so that, in point of fact, the offspring of these illegitimate unions correspond almost precisely to hybrids.[370]