780 TRANSACTIONS OF SECTION K» 



under the name A. arvensie, L., in the 'Index Kewensis,' and also in Engler's 

 ' Pflanzenreieh,' in which latter work they are given the varietal names of 

 phcenieea and ccerulea. The blue variety is stated to be without glandular hairs 

 on the margin of the petals, but in all the specimens examined such hairs have 

 been present, and the edge of the petals was not always markedly denticulate. 



In both forms the throat of the corolla is of purple colour, due to the presence 

 of a purple sap in the cells of this region, and in the centre of these purple cells 

 there is always a collection of needle-shaped crystals of deep-blue colour. Sap of 

 the same purple colour is found in the stamina! hairs, and also in the glandular 

 hairs which fringe the petals in both forms, but in the blue variety the basal 

 cells of these glands contain numerous blue crystals similar to those found in the 

 throat of the corolla. The sap of the cells appears to be acid, and, when treated 

 with potash, the blue crystals observed in these cells are dissolved. 



In the scarlet form the bright colour of the petals is due to a red sap, different 

 in colour and in its chemical nature from the purple sap of the cells just referred 

 to, at the base and the edge of the corolla. In the blue form the colour seems to 

 be due to a blue sap, though possibly the appearance is really due to very finely 

 divided particles of blue colour. When treated with dilute acetic acid the sap 

 tends to become pink, and small blue granules make their appearance in the cells. 

 Two other varieties of this species occur in nature, one somewhat salmon- 

 coloured (A. carnea of Schrank) and one of pale-pink colour. Both these forms 

 resemble the scarlet form, except in the lighter colour of their cell-sap, and have 

 the same purple throat with blue crystals in its cells. Through the kindness of 

 my friend Mr. Doeg, of Manchester, I obtained specimens of ail four varieties 

 from Anglesea, where they occur fairly commonly. 



On pollinating the scarlet form with pollen of the blue variety I obtained 

 plants which bore pure scarlet flowers, and the reciprocal cross, blue pollinated 

 with pollen of a scarlet form, produced plants with scarlet flowers indistinguish- 

 able from the scarlet parent. Only on one or two flowers out of several hundred 

 wa.s a small blue streak noticeable on the petals and indicated the hybrid 

 nature of the plant. The red colour was therefore dominant, and the blue colour 

 recessive. 



In the f., generation complete segregation took place, and the plants bore 

 either brilliant scarlet or bright blue flowers; no intermediate colours were 

 noticeable. This applied to the offspring from both of the crosses. No very 

 good Mendelian ratios were obtained in the numbers of the blue and red plants, 

 partly on account of the small number of the plants and partly owing to somo 

 interruption in the work. Up to the present the numbers in f 2 generation are— 

 in descendants of A. ccerulea $ x phcenicea £ sixty-two red and eight blue; in 

 the case of the descendants of A. phcenicea^ X ccerulea & twenty-five red and 

 two blue. 



The pale-pink form was crossed with blue, and the latter colour was in this 

 case alsu found to be recessive. The f 2 generation of this cross has not yet been 

 obtained. 



The complete dominance of the pure red colour in the crosses of the varieties 

 phcenicea and ccerulea is contrary to the suggestion which has been made (Engler's 

 ' Pflanzenreich') that A. carnea (Schrank) is the hybrid of these two forms. It is 

 more likely, I think, to be a pale variety of the scarlet pimpernel, especially as 

 there is some difference in the depth of colour of this form. Experiments are 

 now being carried on to settle this point. 



The experiments described above confirm those of Focke ('Pflanzen- 

 mischlinge,' 1881), who obtained a hybrid of Anagallis arrensis-phcenicea 9 x 

 ccerulea in which the first flower had one-tenth of the corolla — i.e., one-half of 

 one petal — blue, while the rest was red, and all the subsequent flowers were red. 

 He did not, however, carry his experiments on to the /, generation, and so did 

 not observe the complete segregation of the two forms. 



Gaertner, on the other hand (' Bastardenerzeugung,' 1849), states that the two 

 forms do not produce fertile seed when crossed. In this connection it may be men- 

 tioned that a blue form sent me by Dr. Berger, of La Mortola, cannot be crossed 

 with the ordinary red form. Other characters, too, seem to indicate that this 

 Southern blue form may be a different species. The hybrid described by P'irfler 



