3/8 TJie Americayi Geologist. December, isoo 
without interest to apply these methods to the resuhs of the 
analyses given, and to show what relations each method puts 
|)articularly in relief in the special cases considered. 
The method proposed by Michael Levy*has the advantage 
of giving a diagram for each rock, and may be employed there- 
fore not only for a petrographic series, whose various types 
may readily be compared, but also for any rock considered 
alone. The method of Brogger is a modification of the pre- 
ceding, which loses some of the advantages of Michael Levy's 
method, particularly in the case of some of the rocks here 
considered. 
The methods of Becke and of Iddings adapt themselves to 
a less general usage than the preceding methods, since they 
are destined to present the characteristics of a seriesof rocks, 
and, in certain cases, its continuity, but not of any rock con- 
sidered alone. 
The method of Michael Levy may be considered in detail 
first, as it is more particularly suggestive for certain purposes 
of special importance in the present instance. 
According to this method the various elements K, Na, Ca, 
Mg, and Fe, in the form of their oxides, as found in the anal- 
ysis, are inscribed on two rectangular axes in the following 
way : the potassium oxide {k') is represented on the negative 
ordinate^ the sodium oxide (//) on the positive abscissa, the 
calcium oxide entering into the composition of the feldspar 
(c) on the positive ordinate; and thus we have the summits 
of the "alkaline-earth" triangle {ktic), or in most cases, the 
feldspathic triangle. The oxide of magnesium (w) is repre- 
sented on the positive ordinate, the excess of calcium oxide 
(^')on the negative abscissa and the oxides of iron (/and/') 
on the negative ordinate, and the summits of the ferromag- 
nesian triangle {//ic '/) are thus fixed. This triangle is tinted 
in black. 
It will be noted that two important elements. Si and Al, 
remain unrepresented ; the silica is provided for by inscribing 
the numerical per cent on the horizontal axis ; the alumina as 
a whole is never directly represented, but if an excess of alum- 
ina {a = alumina not combined in the form of feldspars) oc- 
*Michel Levy: Qassification des Magmas dcs Roches Eruptives: 
Bull. Soc. Geol. Fr. 1897, XXV. pp. 326-376. 
