THE HIPPEASTRUM (AMARYLLIS). 



251 



interest for attempting the improvement of the old race of 

 Amaryllis that was not to be neglected, and we accordingly 

 commenced in 18G7 a series of experiments with that object. 

 The first experiments with pardinum were made by Seden, who 

 raised Chclsoni and Brilliant. The general results obtained 

 with pardinum were, however, disappointing, and we soon 

 abandoned the use of that species. We also tried aulicum in our 

 earlier operations, but we obtained nothing of value, the ill shape 

 of the flower of that species, and its large green eye, prepon- 

 derating in the progeny. In Leopoldi we found a more potent 

 subject ; in fact it is not saying too much in affirming that 

 throughout the genus no single species has exercised so prepon- 

 derating an influence in the production of the best forms as this. 

 Our first results with Leopoldi were encouraging, for although 

 we found, as Dean Herbert had found half a century before, 

 that when a particular species is crossed with a mule, or any 

 particular mule with a species, the individuals of the resulting 

 progeny have not only a great resemblance to each other, but 

 also many of them come so near the species as to be practically 

 the same thing, or the same but slightly varied, yet we are able 

 to select several distinct new forms showing a marked improvement 

 on their progenitors in breadth and substance of segment, size 

 and symmetry of flower, &c, while preserving the main features 

 of Leopoldi. The hybrids of this type culminated, as regards 

 form and size of flower, in a variety we called John Heal, as a 

 recognition of the patient and intelligent perseverance of the 

 foreman who now cultivates our collection. The scapes of 

 Leopoldi, it will be remembered, are two-flowered, while those 

 of some of the older species, and especially some of the best 

 hybrids derived from them, are four to six flowered. We at- 

 tained great perfection in form and size in the variety John Heal. 

 Our next step was to attempt to obtain four to six flowered 

 scapes with flowers equally good in form and size, with diversity 

 in colour ; we therefore crossed De Graaffs Empress of India, 

 a noble four to six flowered variety of brilliant colour (a specimen 

 of which is before you), with the best Leopoldi forms, with the 

 result of obtaining not only four to six flowered scapes, but also 

 decided breaks of colour into various shades and tints, and by 

 the intercrossing of the forms so obtained has resulted the race 

 now so universally admired. 



c2 



