NOTKS AND EXHIBITS. 161 



oliLS^ wliicli liad been grown for five years in some instances, for 

 the purpose of testing the characters of growtli and noting any 

 changes in the colour of the flowers and seed-coats, etc., without 

 resorting to artificial pollination. — (1) The various forms of 

 Fhaseobis vulgaris were found to be fairly constant, reproducing 

 their flowers and seeds true to the parent-forms. In two instances, 

 however, dwarf forms were found intermixed with climbers; these 

 were marked, and grown again separately; and in each case, the 

 dwarf or climbing habit of their respective parents was retained. 

 With regard to certain forms of P. innltiflorus, considerable 

 variation occurred in the colour of the flowers and seed-coats. 

 For example, from seeds of the form known in the trade as 

 "Papilio" or "Butterfly-Runner," plants were obtained producing 

 three distinct forms; viz., (1) flowers and seeds like the parent- 

 form, (2) flowers pale scarlet, and seed with cream background 

 with cupreous markings, and (3) flowers similar to No. 2, but with 

 seed having a pale reddish-mauve or amethyst background, and 

 marked with purplish splashes or stripes. The latter had flowered 

 again, producing flowers exactly like the original " Papilio "- 

 forms, namely, scarlet standard and white wings. Another form 

 producing seeds somewhat resembling those of the typical 

 " Scarlet Runner," but paler, with less dark-coloured markings, 

 and flatter irregular seeds, which may be temporarily designated 

 a " Pale form of Scarlet Runner," occasionally produces plants 

 having white flowers, and white seeds, almost identical with 

 those of the form known in the trade as " Czar " or so-called 

 " White Scarlet-Runner." As the different forms were grown 

 quite close to each other, it is quite possible that cross-pollination 

 had taken place, as bees were observed, on two occasions, visiting 

 the flowers, irrespective of colour. — (2) Examples of P. lunatics 

 and the forms of var. macrocarpus were likewise exhibited; these, 

 so far, have given no variants. — (3) Pods and seeds of three 

 species of Stizolobium (Velvet Bean); these, too, so far, have 

 given no v^ariants. — (4) Examples of Canavalia ohtusifolia DC, 

 C. ensifo7-mis DC, and C. gladiata DC, which appear to be con. 

 stant. The last two are regarded by some authors as belonging 

 to one species, but C. ensiformis (commonly known as "Jack 



