186 
and alba oculaia did not produce pleasing results from the 
Our alba and atropurpurea hybrids are all uniform ane Soe 
Vilmorin’s observations (1852. Sur les panachures des fleurs.) 
hrow an interesting light on these contradictory results. He 
his results. Further, this suggests that alba and atropurpurea 
were less constant formerly than they are now and that selection 
has increased fixity. 
Heinemann states that alba oculata fertilized with pollen of 
lilac flowers gave some pretty hybrids, one of which was 
For twenty-two rs Heinemann worked on the improve- 
me: r dti; he used the method of almost 
wholesale ae followed by selection. It is stated ex- 
nase that owers of 1853 were only half as large as 
randi ea ones nd were not brilliant and we 
t ibute his success undoubtedly to both selection and 
hybridization t me of his death (1876) ten grandiflora 
varieties were known (Hamburger Garten und . Blumenzeitung, 
1876, 0 s Phlox Drummondii grandiflora atro 
rea. ‘alia ee which originated, we suspect, in a cross; he 
