388 
THE GARDENERS' CHRONICLE OF AMERICA 
pink flowers is a glorious variety; in my estimation tne 
finest of all the Dianthus. We hardly look for Primulas 
iii September, but P. Cortusoides is a continuous bloom- 
ing member of this charming family. P. Sikkimensis, 
P. Bulleyana and P. capitata, all late bloomers, passed 
many weeks ago. All Primulas do not demand shade. 
P. cortusoides does better in sunshine and the past Sum- 
mer caused plants of P. capitata in moderate shade to rot 
away altogether in some cases. 
Some Sedums, such as Laggeri and Carpathicum, are 
September bloomers, so is Erysimum ochroleucum with 
its bright yellow flowers. There are still a good number 
of flowers on Aquilegia chrysantha. Amongst taller 
growing subjects of value in rockeries, Verbascums 
Olympicum, Olympicum album and blattaria are each 
carrying some spikes. Lobelia cardinalis is hardly 
classed as a rock plant, but plants this season in quite dis- 
positions are three to five feet high and still blooming 
freely. We do not properly appreciate the brilliant car- 
dinal flower as a rock or border plant. It does surprisingly 
well in the average hardy border. Collect seeds as soon 
as ripe and sow in a frame. Purchased seeds usually fail 
to grow at all. 
Geum coccineum Mrs. Bradshaw is a hardy plant 
which blooms practically for six months and in quite dry 
pockets it seems to be quite at home. This is one of the 
finest hardy plants introduced of late years. The first 
flowers are just appeariiiL; of the Colchicums. or autumn 
Crocus, and they remain with us for some weeks. There 
are other flowers additional to those named, and no doubt 
others who grow this most interesting of all types of 
garden flowers can add many more varieties. 
Opportunities of the Garden 
IN nearly all gardens of fair size, and even in tiny ones, 
there may occur sometime some chance variation of 
a well-known flower. It may be some slight change 
in color or form, some difference in foliage, which might 
add to the delight of all gardeners and give to the plant 
itself some peculiar distinction and charm. Or, if the 
chance variation does not occur in your own garden, 
why not offer your assistance (or interference, perhaps) 
to Nature in the working-out of her well-ordered laws, 
and held to develop new varieties yourself? This will 
require, for the most part, only a little knowledge of 
kindred plant families, and care in the pollinating of the 
flowers, and later, in segregation from the parent types. 
There occurred in our own small garden, five or six 
years ago, writes Elma Loines in Landscape Architecture, 
a variation of the Rocky Mountain columbine. This 
flower, as my readers well know, is normally long- 
spurred, and has five petals, which contain the spurs and 
outside five sepals. The sepals are pure blue ; the white 
corona appears on the inner and upper edge of the 
smirred petals. The seed-pod has five divisions. The 
chance variation, which occurred in just two flowers on 
i< plant of possibly a hundred blooms, had no white at all 
in the petals, no spurs and, one might almost say, no 
petals, for what now appeared to be petals were really 
sepals. These were of pure sky-blue, and were no longer 
five, but ten in number. With this transformation, the 
whole aspect of the flower changed. It ceased to be a 
columbine and became something distinct in itself. As 
the flower unfolded from the bud and came into bloom, 
it was like a lovely blue chalice filled with the deep gold 
of many stamens. Gradually it opened more and more, 
but never lost entirely the appearance of the cup. 
From these two flowers were gathered seven rather 
spindling seeds ; and of these seven but two germinated. 
The seedlings were set apart in a little garden bv them- 
selves, but they never bloomed until the third year. Thev 
then bloomed gloriously. On one plant were counted one 
hundred blossoms, on the other fifty ; and, oddly enough, 
each plant produced nothing but the new flower, i. e., 
there was not a single reversion to the old type. This 
was remarkable, for, in the working-out of Mendel's 
law, one would fully expect either that each plant would 
bear two types of blossoms, or that one plant at least 
would bear blossoms entirely of the original type. I 
came to the natural conclusion, therefore, that if the five 
other seeds of the seven original had germinated, some 
of them at least would have continued true to type. So, 
in this case, there was no need to throw out the old, or 
spurred variety. When the two plants went to seed, a 
marked tendency toward variation was at once notice- 
able in the number of the divisions of the seed-pods, 
which varied in odd numbers anywhere from three to 
eleven. 
The next generation came entirely true to the new 
type, with the exception of two new variations, one which 
showed a tendency on the part of alternate petals (as I 
will now call them) to become concave, as if they were 
trying to revert to the spurred petals, which are slightlv 
incurved ; the other, a tendency to increase the number 
of petals, for two blossoms had twelve instead of ten. 
The seed-pods of this, the third generation, varied some- 
what in the number of divisions, but were in general 
five, i. e., like those of the spurred type. 
The new flower, after six years, seems to be pretty 
well established, although it is of course possible that 
the spurred type may occasionally recur in succeeding 
generations. The flower has been named St. Columba. 
as it was derived from a columbine, and chance has it 
that it is at perfection in the locality where it was first 
grown about June 9, at the time of the feast of the Celtic 
saint. 
It comes into bloom with the earliest of the spring 
flowers, and in northern latitudes remains in bloom for 
six to eight weeks. It is charming as a garden plant 
and, when cut, is graceful in the house. 
If the chance variation does not occur in the garden, 
another method of developing new varieties may be tried, 
in order to obtain changes in color, height, or form ; 
that is, pollinating by hand. Considerable care has here 
to be exercised. The parents of the proposed hybrid 
must be selected from plants belonging to the same gen- 
eral families. They must come into bloom about the 
same time. The plants must possess differentiating 
characters, such as dwarfness and tallness, or differences 
in color, form. etc. The flower of the seed-bearing plant 
must be kept away from the influence of foreign pollen 
during the time that it is in bloom. Just before the 
flower which is to bear the seed comes into bloom, that 
is. before the pollen ripens on the anthers of the stamens, 
all the stamens must be picked off, if the separate flowers 
are self-fertilizing. Then, when the pollen is ripe and 
fresh on the pollen-producing plant selected, the pollen 
may be gathered on a small camel's-hair brush and 
brushed well over the stigma of the seed-bearing flower. 
Care must then be taken that no insect bring pollen of 
the flowers of the seed-bearing plant to the hvbrid. This 
may be accomplished by having the plant off in a small 
bed by itself, provided that all the flowers be cross- 
fertilized ; or else by covering the special flower with 
