248 
The Garden Magazine, June, 1921 
Bliss (System One for Class I and Class V, and in System Three 
throughout) is a simple and logical one. 
Since very many of the blue Irises and practically all of the 
so-called red Irises contain some violet — such as Caprice or 
Edouard-M icbel which are violet with a varying amount of 
white or red in them — it seems reasonable to put all of these 
in a class “blue to violet’’ with subdivisions as indicated in 
System Five. 
A class “lilac to rose” is used in System Two. In System 
Five this is changed to read “red to pink in general effect” 
which makes provision not only for such varieties at present 
extant but for the clearer reds and pinks of the future as well. 
Each system contains a class to include “blends.” Such 
varieties as Eldorado and Quaker Lady cannot well be included 
elsewhere. However, a variety such as Iris King, which is 
yellow in general effect, should be included in that class rather 
than among the blends. This may be the way to handle a 
number of others now classified as blends. 
Terms used to characterize the groups in the old classification 
have been omitted entirely in System Three and System Five; 
such terms as neglecla and variegata seeming particularly mean- 
ingless and unsuitable. 
For subdivisions also, color of flower is a helpful basis of iden- 
tification, though to use this to the exclusion of other valuable 
characters would be unfortunate indeed. In every really effi- 
cient classification known to the writer, root, leaf, stem, flower, 
and seed characters are used in the tracing and determination of 
plants. 
Classification by Time of Flowering, Height, and Form of Stem 
T IME of flowering is one of the bases of classification in Sys- 
tem Two and, allowing for some variation from season to 
season, may be broadly divided into: (i) Early-flowering, (2) 
Medium-flowering, (3) Late-flowering. 
Another division useful for purposes of classification is that of 
height of plant at flowering time, roughly: (1) Dwarf, (2) 
Medium-height, (3) Tall. Both height and time of flowering 
vary, of course, according to conditions of soil, climate, etc.; 
but there are certain fundamental differences of growth inherent 
in varieties that external conditions have no power to alter. It 
is scarcely conceivable, for example, that the slender leaves and 
medium-high stem of Trautlieb could by any combination of 
sunshine, soil, and water, take on the broad-leafed, tall-stemmed 
character of Violacea grandiflora. 
Fleight and time of flowering of the different types of Iris are 
factors to be borne in mind when planting. 
Another more constant and especially serviceable character on 
which to base identification is the form of the flower stem. As 
a result of hybridization the newer varieties are frequently char- 
acterized by branching stems; for example, Trojan, lsoline, Nine 
Wells, I. aphylla, and others which contrast strikingly with the 
usual, straight-stemmed varieties. This is a feature of value in 
classifying before the flowers open and for a brief period 
after they are gone. 
Classification by Color of Leaf 
A DEFINITE plant character, very easily seen in most 
varieties possessing it and of great value in identifying 
garden varieties of Iris, is coloration of the bases of the leaves. 
(This character appears more valuable in the identification of 
varieties which show it plainly than the red coloration of the leaf 
margins mentioned by W. R. Dykes in the Genus Iris as distinc- 
tive of the variety Kharput; or for distinguishing var. Rubro 
marginata from others of the Mellita group, mentioned in The 
Garden No\. LXXX 1 1 , No. 2450, Nov., 1918, by thesame writer.) 
Since 1908, when I first noticed that a number of varieties were 
purple at the leaf-base, I have kept this plant character steadily 
under observation and have found that the amount of color in 
varieties varies considerably, but, whenever present, it remains 
constant throughout the year. Also I am led to conclude 
1U' 
that leaf coloration is much less influenced by soil and 
climate than it is commonly thought to be. The presence 
of anthocyanin, a pigment, in the cell sap is the real cause of 
such coloration. 
In his article: “ Mendelian Characters in Bearded Irises,” 
(, Journal of the Royal H orticultural Society, Vol. XLV, pp. 289, 
292; 1920), Mr. A. J. Bliss of Morwellham, England, shows that 
the coloration of the leaf base is definitely inherited in Mendelian 
ratios, and is not linked with any particular flower color. I 
have, however, observed that in some instances small divisions, 
or weak plants of a number of varieties having colored leaf bases 
show less coloration than strong plants of standard size, and at 
times none at all until they have become well rooted and made a 
growth approaching normal size. 
Here then is a definite plant character very easily seen in most 
varieties possessing it, which is very valuable in identification 
of the varieties in the garden, and in a key to facilitate the tracing 
down of unknown varieties, and in descriptions in order to make 
actual identification more certain. 
Among the very best varieties at present are a considerable 
number which show this character plainly. A few are 
listed below under the color groups as arranged in System 
Five. 
1 
- 
1 
ft 
Pr 
friA 
l 
)• 
L? ; 
Sib 
luf 
Group I. Flowers white in general effect 
Mrs. H. Darwin (Foster), Rhein Nixe (Goos & Koenemann) 
Tristram (Bliss), Thorbeck 
Group 11. Flowers yellow in general effect 
Loreley (Goos & Koenemann), Marsh Marigold, and Omega (Bliss), 
Sherwin Wright (Kohankie), Maori King 
Group 111. Flowers blue to violet in general effect 
Azure (Bliss), Lugarda (Sturtevant), aphylla 
Group IV. Flowers red to pink in general effect 
Monsieur Aymard, and T roost (Denis) 
Group V. Flowers with blended colors 
Cretonne, Dusky Maid, and Romany (Bliss), Mme. Cheri (Sturte- 
vant), Dejazet (Vilmorin), Quaker Lady (Farr), A. E. Kunderd 
(Fryer), Jacquiniana (Lemon), Niebelungen and Peacock’s Eye 
(Goos & Koenemann) 
die: 
In addition to the varieties listed above, Prestige (Sturtevant) 
in Group 1 1 ; Avalon (Sturtevant), and Sybil in Group IV; Dora 
Longden and Mrs. Cowley (Bliss), Dalila (Denis), and Dalmarius 
in Group V. have some coloration of basal leaves. 
The coloration of the variety Cretonne (Bliss) is very strong 
and extends well up on the leaves to the flowering stems and 
spathe valves. This is true also for Iris aphylla described by 
W. R. Dykes in The Garden, September 28th, 1918, page 363. 
R. S. Sturtevant also mentions (The Garden Magazine, Vol. 
XXXI I, No. 2. pp. 103), the enhanced decorative effect of 
Lugarda and A. E. Kunderd due to the coloration of the 
leaf bases. This suggests the possibility of producing a 
variety with the leaves largely or entirely purple, which would 
increase the effectiveness of Iris foliage throughout the 
season. 
Coloration of leaf base as a practical guide is readily appre- 
ciated when it is known that Albicans (non-colored base), 
Mrs. H. Darwin (colored base), White Knight (non-colored 
base), Thorbeck (colored base), La Neige (non-colored base), 
Rhein Nixe (colored base), and Kashmir White (non-colored 
base) — all white varieties in general effect — may be planted in 
the order named with no reason for making mistakes when plants 
are removed at other than flowering time, even should one or 
more of the markers be removed by accident. Likewise, Aurea 
(non-colored base), Sherwin Wright (colored base), and Mrs. 
Neubronner (non-colored base), may be planted in the order 
named in rows or bed, and the separation can be made without 
difficulty out of flowering season though the markers are lost. 
This may suggest other uses in nurseries and in private 
gardens. 
Every lover and grower of Iris likes to be able to call the in- 
habitants of his garden by name. Careful study of plant charac- 
ters throughout the season will enable him to do this; of the 
various points on which identification may be based none is more 
constant than the color of leaf base. The system which 1 have 
found very useful will, I hope, prove equally serviceable toothers. 
n 
ta 
ce 
ch 
t\\ 
ce 
er 
in 
P 
dt 
A 
P 1 
la 
e: 
(1 
P 
1 
r 
t 
i 
