Column 2: Number of Flowers fertilised.

Column 3: Number of Capsules produced.

Column 4: Average Number of Seeds per Capsule.

Long-styled by pollen of longest stamens of short-styled. Legitimate union : 19 : 15 : 57.4.

Long-styled by pollen of longest stamens of mid-styled. Legitimate union : 4 : 3 : 59.0.

Long-styled by pollen of own-form mid-length stamens. Illegitimate union : 9 : 2 : 42.5.

Long-styled by pollen of own-form shortest stamens. Illegitimate union : 11 : 0 : 0.

Long-styled by pollen of shortest stamens of mid-styled. Illegitimate union : 4 : 0 : 0.

Long-styled by pollen of mid-length stamens of short-styled. Illegitimate union : 12 : 5 : 30.0.

Mid-styled by pollen of mid-length stamens of long-styled. Legitimate union : 3 : 3 : 63.6.

Mid-styled by pollen of mid-length stamens of short-styled. Legitimate union : 4 : 4 : 56.3.

Mid-styled by mixed pollen from both own-form longest and shortest stamens.

Illegitimate union : 9 : 2 : 19.

Mid-styled by pollen of longest stamens of short-styled. Illegitimate union : 12 : 1 : 8.

Short-styled by pollen of shortest stamens of mid-styled. Legitimate union: 3 : 2 : 67.

Short-styled by pollen of shortest stamens of long-styled. Legitimate union: 3 : 3 : 54.3.

Short-styled by pollen of own-form longest stamens. Illegitimate union: 5 : 1 : 8.

Short-styled by pollen of own-form mid-length stamens. Illegitimate union : 3 : 0 : 0.

Short-styled by both pollens mixed together, of own-form longest and mid-length stamens. Illegitimate union: 13 : 0 : 0.

Short-styled by pollen of longest stamens of mid-styled. Illegitimate union : 7 : 0 : 0.

Short-styled by pollen of mid-length stamens of long-styled. Illegitimate union: 10 : 1 : 54.

We here see that thirty-six flowers on the three forms legitimately fertilised yielded 30 capsules, these containing on an average 58.36 seeds. Ninety-five flowers illegitimately fertilised yielded 12 capsules, containing on an average 28.58 seeds. Therefore the fertility of the six legitimate to that of the twelve illegitimate unions, as judged by the proportion of flowers which yielded capsules, is as 100 to 15, and judged by the average number of seeds per capsule as 100 to 49. This plant, in comparison with the two South American species previously described, produces many more seeds, and the illegitimately fertilised flowers are not quite so sterile.

Oxalis rosea.

Hildebrand possessed in a living state only the long-styled form of this trimorphic Chilian species. (4/14. 'Monatsber. der Akad. der Wiss. Berlin' 1866 page 372.) The pollen-grains from the two sets of anthers differ in diameter as 9 to 7.5, or as 100 to 83. He has further shown that there is an a.n.a.logous difference between the grains from the two sets of anthers of the same flower in five other species of Oxalis, besides those already described. The present species differs remarkably from the long-styled form of the three species previously experimented on, in a much larger proportion of the flowers setting capsules when fertilised with their own-form pollen. Hildebrand fertilised 60 flowers with pollen from the mid-length stamens (of either the same or another flower), and they yielded no less than 55 capsules, or 92 per cent. These capsules contained on an average 5.62 seeds; but we have no means of judging how near an approach this average makes to that from flowers legitimately fertilised. He also fertilised 45 flowers with pollen from the shortest stamens, and these yielded only 17 capsules, or 31 per cent, containing on an average only 2.65 seeds. We thus see that about thrice as many flowers, when fertilised with pollen from the mid-length stamens, produced capsules, and these contained twice as many seeds, as did the flowers fertilised with pollen from the shortest stamens. It thus appears (and we find some evidence of the same fact with O.

speciosa), that the same rule holds good with Oxalis as with Lythrum salicaria; namely, that in any two unions, the greater the inequality in length between the pistils and stamens, or, which is the same thing, the greater the distance of the stigma from the anthers, the pollen of which is used for fertilisation, the less fertile is the union,--whether judged by the proportion of flowers which set capsules, or by the average number of seeds per capsule. The rule cannot be explained in this case any more than in that of Lythrum, by supposing that wherever there is greater liability to self-fertilisation, this is checked by the union being rendered more sterile; for exactly the reverse occurs, the liability to self-fertilisation being greatest in the unions between the pistils and stamens which approach each other the nearest, and these are the more fertile. I may add that I also possessed some long-styled plants of this species: one was covered by a net, and it set spontaneously a few capsules, though extremely few compared with those produced by a plant growing by itself, but exposed to the visits of bees.

With most of the species of Oxalis the short-styled form seems to be the most sterile of the three forms, when these are illegitimately fertilised; and I will add two other cases to those already given. I fertilised 29 short-styled flowers of O. compressa with pollen from their own two sets of stamens (the pollen- grains of which differ in diameter as 100 and 83), and not one produced a capsule. I formerly cultivated during several years the short-styled form of a species purchased under the name of O. Bowii (but I have some doubts whether it was rightly named), and fertilised many flowers with their own two kinds of pollen, which differ in diameter in the usual manner, but never got a single seed. On the other hand, Hildebrand says that the short-styled form of O.

Deppei, growing by itself, yields plenty of seed; but it is not positively known that this species is heterostyled; and the pollen-grains from the two sets of anthers do not differ in diameter.

Some facts communicated to me by Fritz Muller afford excellent evidence of the utter sterility of one of the forms of certain trimorphic species of Oxalis, when growing isolated. He has seen in St. Catharina, in Brazil, a large field of young sugar-cane, many acres in extent, covered with the red blossoms of one form alone, and these did not produce a single seed. His own land is covered with the short-styled form of a white-flowered trimorphic species, and this is equally sterile; but when the three forms were planted near together in his garden they seeded freely. With two other trimorphic species he finds that isolated plants are always sterile.

Fritz Muller formerly believed that a species of Oxalis, which is so abundant in St. Catharina that it borders the roads for miles, was dimorphic instead of trimorphic. Although the pistils and stamens vary greatly in length, as was evident in some specimens sent to me, yet the plants can be divided into two sets, according to the lengths of these organs. A large proportion of the anthers are of a white colour and quite dest.i.tute of pollen; others which are pale yellow contain many bad with some good grains; and others again which are bright yellow have apparently sound pollen; but he has never succeeded in finding any fruit on this species. The stamens in some of the flowers are partially converted into petals. Fritz Muller after reading my description, hereafter to be given, of the illegitimate offspring of various heterostyled species, suspects that these plants of Oxalis may be the variable and sterile offspring of a single form of some trimorphic species, perhaps accidentally introduced into the district, which has since been propagated as.e.xually. It is probable that this kind of propagation would be much aided by there being no expenditure in the production of seed.

Oxalis (Biophytum) sensitiva.

This plant is ranked by many botanists as a distinct genus. Mr. Thwaites sent me a number of flowers preserved in spirits from Ceylon, and they are clearly trimorphic. The style of the long-styled form is clothed with many scattered hairs, both simple and glandular; such hairs are much fewer on the style of the mid-styled, and quite absent from that of the short-styled form; so that this plant resembles in this respect O. Valdiviana and Regnelli. Calling the length of the two lobes of the stigma of the long-styled form 100, that of the mid- styled is 141, and that of the short-styled 164. In all other cases, in which the stigma in this genus differs in size in the three forms, the difference is of a reversed nature, the stigma of the long-styled being the largest, and that of the short-styled the smallest. The diameter of the pollen-grains from the longest stamens being represented by 100, those from the mid-length stamens are 91, and those from the shortest stamens 84 in diameter. This plant is remarkable, as we shall see in the last chapter of this volume, by producing long-styled, mid-styled, and short-styled cleistogamic flowers.

h.o.m.oSTYLED SPECIES OF OXALIS.

Although the majority of the species in the large genus Oxalis seem to be trimorphic, some are h.o.m.ostyled, that is, exist under a single form; for instance the common O. acetosella, and according to Hildebrand two other widely distributed European species, O. stricta and corniculata. Fritz Muller also informs me that a similarly const.i.tuted species is found in St. Catharina, and that it is quite fertile with its own pollen when insects are excluded. The stigmas of O. stricta and of another h.o.m.ostyled species, namely O.

tropaeoloides, commonly stand on a level with the upper anthers, and both these species are likewise quite fertile when insects are excluded.

With respect to O. acetosella, Hildebrand says that in all the many specimens examined by him the pistil exceeded the longer stamens in length. I procured 108 flowers from the same number of plants growing in three distant parts of England; of these 86 had their stigmas projecting considerably above, whilst 22 had them nearly on a level with the upper anthers. In one lot of 17 flowers from the same wood, the stigmas in every flower projected fully as much above the upper anthers as these stood above the lower anthers. So that these plants might fairly be compared with the long-styled form of a heterostyled species; and I at first thought that O. acetosella was trimorphic. But the case is one merely of great variability. The pollen-grains from the two sets of anthers, as observed by Hildebrand and myself, do not differ in diameter. I fertilised twelve flowers on several plants with pollen from a distinct plant, choosing those with pistils of a different length; and 10 of these (i.e. 83 per cent) produced capsules, which contained on an average 7.9 seeds. Fourteen flowers were fertilised with their own pollen, and 11 of these (i.e. 79 per cent) yielded capsules, containing a larger average of seed, namely 9.2. These plants, therefore, in function show not the least sign of being heterostyled. I may add that 18 flowers protected by a net were left to fertilise themselves, and only 10 of these (i.e. 55 per cent) yielded capsules, which contained on an average only 6.3 seeds. So that the access of insects, or artificial aid in placing pollen on the stigma, increases the fertility of the flowers; and I found that this applied especially to those having shorter pistils. It should be remembered that the flowers hang downwards, so that those with short pistils would be the least likely to receive their own pollen, unless they were aided in some manner.

Finally, as Hildebrand has remarked, there is no evidence that any of the heterostyled species of Oxalis are tending towards a dioecious condition, as Zuccarini and Lindley inferred from the differences in the reproductive organs of the three forms, the meaning of which they did not understand.

PONTEDERIA [SP.?] (PONTEDERIACEAE).

Fritz Muller found this aquatic plant, which is allied to the Liliaceae, growing in the greatest profusion on the banks of a river in Southern Brazil. (4/15.

"Ueber den Trimorphismus der Pontederien" 'Jenaische Zeitschrift' etc. Band 6 1871 page 74.) But only two forms were found, the flowers of which include three long and three short stamens. The pistil of the long-styled form, in two dried flowers which were sent me, was in length as 100 to 32, and its stigma as 100 to 80, compared with the same organs in the short-styled form. The long-styled stigma projects considerably above the upper anthers of the same flower, and stands on a level with the upper ones of the short-styled form. In the latter the stigma is seated beneath both its own sets of anthers, and is on a level with the anthers of the shorter stamens in the long-styled form. The anthers of the longer stamens of the short-styled form are to those of the shorter stamens of the long-styled form as 100 to 88 in length. The pollen-grains distended with water from the longer stamens of the short-styled form are to those from the shorter stamens of the same form as 100 to 87 in diameter, as deduced from ten measurements of each kind. We thus see that the organs in these two forms differ from one another and are arranged in an a.n.a.logous manner, as in the long and short-styled forms of the trimorphic species of Lythrum and Oxalis. Moreover, the longer stamens of the long-styled form of Pontederia, and the shorter ones of the short-styled form are placed in a proper position for fertilising the stigma of a mid-styled form. But Fritz Muller, although he examined a vast number of plants, could never find one belonging to the mid-styled form. The older flowers of the long-styled and short-styled plants had set plenty of apparently good fruit; and this might have been expected, as they could legitimately fertilise one another. Although he could not find the mid-styled form of this species, he possessed plants of another species growing in his garden, and all these were mid-styled; and in this case the pollen-grains from the anthers of the longer stamens were to those from the shorter stamens of the same flower as 100 to 86 in diameter, as deduced from ten measurements of each kind. These mid-styled plants growing by themselves never produced a single fruit.

Considering these several facts, there can hardly be a doubt that both these species of Pontederia are heterostyled and trimorphic. This case is an interesting one, for no other Monocotyledonous plant is known to be heterostyled. Moreover, the flowers are irregular, and all other heterostyled plants have almost symmetrical flowers. The two forms differ somewhat in the colour of their corollas, that of the short-styled being of a darker blue, whilst that of the long-styled tends towards violet, and no other such case is known. Lastly, the three longer stamens alternate with the three shorter ones, whereas in Lythrum and Oxalis the long and short stamens belong to distinct whorls. With respect to the absence of the mid-styled form in the case of the Pontederia which grows wild in Southern Brazil, this would probably follow if only two forms had been originally introduced there; for, as we shall hereafter see from the observations of Hildebrand, Fritz Muller and myself, when one form of Oxalis is fertilised exclusively by either of the other two forms, the offspring generally belong to the two parent-forms.

Fritz Muller has recently discovered, as he informs me, a third species of Pontederia, with all three forms growing together in pools in the interior of S.

Brazil; so that no shadow of doubt can any longer remain about this genus including trimorphic species. He sent me dried flowers of all three forms. In the long-styled form the stigma stands a little above the tips of the petals, and on a level with the anthers of the longest stamens in the other two forms.

The pistil is in length to that of the mid-styled as 100 to 56, and to that of the short-styled as 100 to 16. Its summit is rectangularly bent upwards, and the stigma is rather broader than that of the mid-styled, and broader in about the ratio of 7 to 4 than that of the short-styled. In the mid-styled form, the stigma is placed rather above the middle of the corolla, and nearly on a level with the mid-length stamens in the other two forms; its summit is a little bent upwards. In the short-styled form the pistil is, as we have seen, very short, and differs from that in the other two forms in being straight. It stands rather beneath the level of the anthers of the shortest stamens in the long-styled and mid-styled forms. The three anthers of each set of stamens, more especially those of the shortest stamens, are placed one beneath the other, and the ends of the filaments are bowed a little upwards, so that the pollen from all the anthers would be effectively brushed off by the proboscis of a visiting insect.

The relative diameters of the pollen-grains, after having been long soaked in water, are given in Table 4.d, as measured by my son Francis.

TABLE 4.d. Pontederia. Diameters of pollen-grains, after having been long soaked in water, in divisions of the micrometer.

Column 1: Source of Pollen-grains.

Column 2: diameter.

Long-styled form, mid-length stamens (Average of 20 measurements): 13.2.

Long-styled form, shortest stamens (10 measurements): 9.0.

Mid-styled form, longest stamens (15 measurements) : 16.4.

Mid-styled form, shortest stamens (20 measurements): 9.1.

Short-styled form, longest stamens (20 measurements): 14.6.

Short-styled form, mid-length stamens (20 measurements): 12.3.

We have here the usual rule of the grains from the longer stamens, the tubes of which have to penetrate the longer pistil, being larger than those from the stamens of less length. The extreme difference in diameter between the grains from the longest stamens of the mid-styled form, and from the shortest stamens of the long-styled, is as 16.4 to 9.0, or as 100 to 55; and this is the greatest difference observed by me in any heterostyled plant. It is a singular fact that the grains from the corresponding longest stamens in the two forms differ considerably in diameter; as do those in a lesser degree from the corresponding mid-length stamens in the two forms; whilst those from the corresponding shortest stamens in the long- and mid-styled forms are almost exactly equal.

Their inequality in the two first cases depends on the grains in both sets of anthers in the short-styled form being smaller than those from the corresponding anthers in the other two forms; and here we have a case parallel with that of the mid-styled form of Lythrum salicaria. In this latter plant the pollen-grains of the mid-styled forms are of smaller size and have less fertilising power than the corresponding ones in the other two forms; whilst the ovarium, however fertilised, yields a greater number of seeds; so that the mid-styled form is altogether more feminine in nature than the other two forms. In the case of Pontederia, the ovarium includes only a single ovule, and what the meaning of the difference in size between the pollen-grains from the corresponding sets of anthers may be, I will not pretend to conjecture.

The clear evidence that the species just described is heterostyled and trimorphic is the more valuable as there is some doubt with respect to P.

cordata, an inhabitant of the United States. Mr. Leggett suspects that it is either dimorphic or trimorphic, for the pollen-grains of the longer stamens are "more than twice the diameter or than eight times the ma.s.s of the grains of the shorter stamens. Though minute, these smaller grains seem as perfect as the larger ones." (4/16. 'Bulletin of the Torrey Botanical Club' 1875 volume 6 page 62.) On the other hand, he says that in all the mature flowers, "the style was as long at least as the longer stamens;" "whilst in the young flowers it was intermediate in length between the two sets of stamens;" and if this be so, the species can hardly be heterostyled.

CHAPTER V.