Fh 
272 A.STUDY OF WOOD HYACINTHS. 
Let us now take "que organ by organ, of the amount of variation 
which the series show: 
Bulbs precisely similar i in structure, but varying in shape from ovoid through 
globose de depresso-globose, and in thickness from three-quarters of an ineh to an 
a half. 
pres $ varying greatly in size, shape, and imn In length from half a foot 
to a foot, above the soil; in breadth from half an inch to an inch and a half 
width either gradually reduced through the upper “thir 
within a sh nce of the tip, the point blunt or acute; the texture firm, so 
that the leaf maintains a sub-erect position till t is over, or weak, 80 
that it falls backwards over at an early stage ; the face deeply or se: Mieres 
channelled, and the back conspicuously or hardly at all convex, and in the sub- 
ct i ed rved. 
m as long as the leaves to sé as lon 
Raceme varying mu gn in density, and the pedicels pas - - length and 
ion. 
Colour of flowers.—Three pes distinct shades, the commonest hyacinth- 
blue, the other i a pale lilac-pink and a pure white, these running through the 
forms clearly without any correlation with other differences. 
ipe of the flower.— Varying from the s segments pern inam MAD p 
tube in the lower two-thirds to spreading falcately from paan p: the base, the 
length of the wiped pim. from pelt to three-quarters aan an den 
an t 
ent he filaments.—' The sepaline filaments always 
attached distinctly higher than the petaline three The attachment usually cor- 
relative with the f co ut not invariably so, the absolute length 
being mereased, proportion that is adnate greater the higher and more 
sentinels the segments are ly connivent. In the closed-flowered forms 
Supel shaped. In describing the fo ms, they are arranged in series according 
e dvo olas characters, pore with t the most closed, and ending with the most 
ed 
nite avoiding mere descriptions as ambiguous, let us compare the 
published figures with the forms above described. The history of the 
l 
and campanulata. Rh and Kunth admit four species, nutans, 
cernua, = and campanu 
ow to the figu a na are drawn in the ‘Botanical Magazine.’ 
Turni 
Plate 12 197. (but. letterpress accidentally numbered 128), called Scilla cam- 
Panutata, agrees with our No. 4, exce ept that the leaves are narrowed to @ 
: . campanulata, is most like our mes 
but with stronger deeply channelled leaves; but there are wo 
rs ith it as bei : 
only, which he calls patula, quotes this plate as representing it. 
1461, called Scilla nonscripta variety, es very well our No. 
douté’s plates do not sd so natural, ie are less easy to pronounes ai 
upon. His 1. 224 » Scilla n cripta 1 
Js com 
cripta, has very short pedice t 
iis me i inch sita Setieicttly connivent «three-quarters bee 
y up. t. 225, Scilla patula, has still larger flowers, a pie the 
connivent half t ; a 
or two-thirds of the lin de on leaves 
* 
permanentl n e 
e two threes are distinetly tà in length, but it is not so when the flowers 
form. is t. 435, Scilla campanula : 
t long, an inch broad, and the divisions of the perianth e x 
