166 
OK VARIETY IN FLOWER-GAEDENS. 
produced. Every fermenting mass, particularly if much urine be present, yields a volume 
of ammoniacal vapour that is conveyed into the atmosphere, and is returned to the earth by 
rain and snow showers. Liebig was the first chemist who demonstrated the actual 
presence of ammonia in rain water, a fact of great interest to the gardener, since it instructs 
him that the putrid exhalations from matters decaying on the surface of the ground are 
purified in the great meteorological laboratory of the atmosphere, instead of remaining 
below, as a source of epidemic and pestilence. 
Ammoniacal gas is much lighter than atmospheric air. From the most critical and 
correct analysis, it should appear that three measures of hydrogen and one measure of 
nitrogen gas, condensed into two measures, constitute ammoniacal gas, the specific gravity 
of which, when compared with that of air, is as 0.590 to 1000. It perhaps is unfortunate 
that hydrogen and nitrogen, when artificially mixed, show no tendency synthetically to 
combine ; but we learn that, if a succession of electric sparks be passed through a small 
portion of ammoniacal gas confined in a proper tube, over quicksilver, the gas will increase 
to twice its original bulk, and lose that easy solubility in water which characterises gaseous 
ammonia. Again, by the agency of oxygen, and refined calculations which it would be 
irrelevant now to detail, chemists have discovered the proportions in which hydrogen and 
nitrogen gas naturally combine to produce ammonia. 
In the next article, the subject of atmospheric air and its accessories will come under 
consideration. 
ON VARIETY IN FLOV^EE-GARDENS. 
By Mr. Moore. 
Variety is, after all, the great idol which the majority of mankind worships ; and for 
variety's sake it is the common fashion to make almost any sacrifice. Now, without 
attempting to decide whether or no this eagerness to bow before a shrine identified with 
inconstancy, be generally praiseworthy, as it is generally prevalent, it may surely be 
assumed, that at least, as far as regards flower-garden arrangements, there is little or no 
folly in becoming one of variety's devotees. 
If this be true, the fact militates more or less against the practice of planting large 
masses of unbroken colours to make up a Flower-Garden. The end thus attained is no 
doubt a gaudy display ; but, paradoxical as it may appear, the question may be raised, 
whether, in such cases, gaudiness is not obtained at a sacrifice of effect. Doubtless there 
is, and ever will be, more than one opinion upon such a point ; but at any rate, it may 
not be assumed as unquestionable that a gaudy display and elegance of effect are synony- 
mous. The fact appears to be that here we have a contest — not an uncommon one — 
whether quantity or quality should prevail. The great glare of colour in the one case, as 
the more attractive, is commonly set down as the more perfect result; a decision which, 
upon the face of it, appears questionable. 
But it does not therefore follow that colours should not be massed. To be effective, 
colour must be decided or obvious, and to be decided or obvious, it must not be too much 
broken or isolated. It seems, therefore, on the whole, that the effort to be made is 
rather to contract than to enlarge the groups of flower-garden plants, in order that a given 
space may show a contrast or variety of colour, instead of an unbroken monotonous hue. 
Size is however always relative, and what is large in one place would be small in another, 
so that no absolute rule as regards the size of masses can be drawn in respect to the 
distribution of colours. 
It is some years since the circle w^as first recommended as the most desirable figure 
for flower-beds, and it is now, as it was then, true, that the more perfectly angles in 
flower-beds, and especially acute angles, are avoided, the better ; that is to say, if the bed 
is to represent a mass or any formal combination of colours ; for there is a tendency 
towards rotundity in the growth of all plants, which renders it next to impossible that 
