Pot 5 : 26 6/8 : 10 4/8.

Total : 522.25 : 428.50.

The average height of the nineteen crossed plants is here 27.48, and that of the nineteen self-fertilised 22.55 inches; or as 100 to 82. All these plants were cut down in the early autumn and weighed: the crossed weighed 11.5 ounces, and the self-fertilised 7.75 ounces, or as 100 to 67. These two lots having been left freely exposed to the visits of insects, did not present any difference to the eye in the number of seed-capsules which they produced.

The remainder of the same two lots of seeds were sown in two adjoining rows in the open ground; so that the plants were exposed to only moderate compet.i.tion. The eight tallest on each side were measured, as shown in Table 4/38.

TABLE 4/38. Reseda odorata, growing in the open ground.

Heights of plants measured in inches.

Column 1: Crossed Plants.

Column 2: Self-fertilised Plants.

24 4/8 : 26 5/8.

27 2/8 : 25 7/8.

24 : 25.

26 6/8 : 28 3/8.

25 : 29 7/8.

26 2/8 : 25 7/8.

27 2/8 : 26 7/8.

25 1/8 : 28 2/8.

Total : 206.13 : 216.75

The average height of the eight crossed plants is 25.76, and that of the eight self-fertilised 27.09; or as 100 to 105.

We here have the anomalous result of the self-fertilised plants being a little taller than the crossed; of which fact I can offer no explanation. It is of course possible, but not probable, that the labels may have been interchanged by accident.

Another experiment was now tried: all the self-fertilised capsules, though very few in number, were gathered from one of the semi-self-sterile plants under a net; and as several flowers on this same plant had been fertilised with pollen from a distinct individual, crossed seeds were thus obtained. I expected that the seedlings from this semi-self-sterile plant would have profited in a higher degree from a cross, than did the seedlings from the fully self-fertile plants. But my antic.i.p.ation was quite wrong, for they profited in a less degree. An a.n.a.logous result followed in the case of Eschscholtzia, in which the offspring of the plants of Brazilian parentage (which were partially self-sterile) did not profit more from a cross, than did the plants of the far more self-fertile English stock. The above two lots of crossed and self-fertilised seeds from the same plant of Reseda odorata, after germinating on sand, were planted on opposite sides of five pots, and measured as in the last case, with the result in Table 4/39.

TABLE 4/39. Reseda odorata (seedlings from a semi-self-sterile plant).

Heights of plants to the summits of the leafy stems, flower-stems not included, measured in inches.

Column 1: Number (Name) of Pot.

Column 2: Crossed Plants.

Column 3: Self-fertilised Plants.

Pot 1 : 33 4/8 : 31.

Pot 1 : 30 6/8 : 28.

Pot 1 : 29 6/8 : 13 2/8.

Pot 1 : 20 : 32.

Pot 2 : 22 : 21 6/8.

Pot 2 : 33 4/8 : 26 6/8.

Pot 2 : 31 2/8 : 25 2/8.

Pot 2 : 32 4/8 : 30 4/8.

Pot 3 : 30 1/8 : 17 2/8.

Pot 3 : 32 1/8 : 29 6/8.

Pot 3 : 31 4/8 : 24 6/8.

Pot 3 : 32 2/8 : 34 2/8.

Pot 4 : 19 1/8 : 20 6/8.

Pot 4 : 30 1/8 : 32 6/8.

Pot 4 : 24 3/8 : 31 4/8.

Pot 4 : 30 6/8 : 36 6/8.

Pot 5 : 34 6/8 : 24 5/8.

Pot 5 : 37 1/8 : 34.

Pot 5 : 31 2/8 : 22 2/8.

Pot 5 : 33 : 37 1/8.

Total : 599.75 : 554.25.

The average height of the twenty crossed plants is here 29.98, and that of the twenty self-fertilised 27.71 inches; or as 100 to 92. These plants were then cut down and weighed; and the crossed in this case exceeded the self-fertilised in weight by a mere trifle, namely, in the ratio of 100 to 99. The two lots, left freely exposed to insects, seemed to be equally fertile.

The remainder of the seed was sown in two adjoining rows in the open ground; and the eight tallest plants in each row were measured, with the result in Table 4/40.

TABLE 4/40. Reseda odorata, (seedlings from a semi-self-sterile plant, planted in the open ground).

Heights of plants measured in inches.

Column 1: Crossed Plants.

Column 2: Self-fertilised Plants.

28 2/8 : 22 3/8.

22 4/8 : 24 3/8.

25 7/8 : 23 4/8.

25 3/8 : 21 4/8.

29 4/8 : 22 5/8.

27 1/8 : 27 3/8.

22 4/8 : 27 3/8.

26 2/8 : 19 2/8.

Total : 207.38 : 188.38.

The average height of the eight crossed plants is here 25.92, and that of the eight self-fertilised plants 23.54 inches; or as 100 to 90.

9. VIOLACEAE.--Viola tricolor.

Whilst the flowers of the common cultivated heartsease are young, the anthers shed their pollen into a little semi-cylindrical pa.s.sage, formed by the basal portion of the lower petal, and surrounded by papillae. The pollen thus collected lies close beneath the stigma, but can seldom gain access into its cavity, except by the aid of insects, which pa.s.s their proboscides down this pa.s.sage into the nectary. (4/5. The flowers of this plant have been fully described by Sprengel, Hildebrand, Delpino, and H. Muller. The latter author sums up all the previous observations in his 'Befruchtung der Blumen' and in 'Nature' November 20, 1873 page 44. See also Mr. A.W. Bennett in 'Nature' May 15, 1873 page 50 and some remarks by Mr. Kitchener ibid page 143. The facts which follow on the effects of covering up a plant of V. tricolor have been quoted by Sir J.

Lubbock in his 'British Wild Flowers' etc. page 62.) Consequently when I covered up a large plant of a cultivated variety, it set only eighteen capsules, and most of these contained very few good seeds--several from only one to three; whereas an equally fine uncovered plant of the same variety, growing close by, produced 105 fine capsules. The few flowers which produce capsules when insects are excluded, are perhaps fertilised by the curling inwards of the petals as their wither, for by this means pollen-grains adhering to the papillae might be inserted into the cavity of the stigma. But it is more probable that their fertilisation is effected, as Mr. Bennett suggests, by Thrips and certain minute beetles which haunt the flowers, and which cannot be excluded by any net.

Humble-bees are the usual fertilisers; but I have more than once seen flies (Rhingia rostrata) at work, with the under sides of their bodies, heads and legs dusted with pollen; and having marked the flowers which they visited, I found them after a few days fertilised. (4/6. I should add that this fly apparently did not suck the nectar, but was attracted by the papillae which surround the stigma. Hermann Muller also saw a small bee, an Andrena, which could not reach the nectar, repeatedly inserting its proboscis beneath the stigma, where the papillae are situated; so that these papillae must be in some way attractive to insects. A writer a.s.serts 'Zoologist' volume 3-4 page 1225, that a moth (Plusia) frequently visits the flowers of the pansy. Hive-bees do not ordinarily visit them, but a case has been recorded 'Gardeners'