For instance, no mountain in the Alps produces a more vigorous impression of peakedness than the Matterhorn. In Professor Forbes's work on the Alps, it is spoken of as an "obelisk" of rock, and represented with little exaggeration in his seventh plate under the outline Fig. 31. Naturally, in glancing, whether at the plate or the mountain, we assume the mass to be a peak, and suppose the line _a b_ to be the steep slope of its side. But that line is a perspective line. It is in reality _perfectly horizontal_, corresponding to _e f_ in the penthouse roof, Fig. 30. [Illustration: FIG. 33. Angles with the horizon _x y_. _a f_ 56 deg. _a e_ 123/4 _e b_ (from point to point) 441/2 _b c_ ( ditto, ditto ) 671/4 _c d_ (overhanging) 79 deg. _a x_ (irrespective of irregularities) 56 _a y_ 383/4 ] Sec. 12. I say "perfectly horizontal," meaning, of course, in general tendency. It is more or less irregular and broken, but so nearly horizontal that, after some prolonged examination of the data I have collected about the Matterhorn, I am at this moment in doubt _which is its top_. For as, in order to examine the beds on its flanks, I walked up the Zmutt glacier, I saw that the line _a b_ in Fig. 31 gradually lost its steepness; and about half-way up the glacier, the conjectural summit _a_ then bearing nearly S. E. (forty degrees east of south), I found the contour was as in Fig. 32. In Fig. 33, I have given the contour as seen from Zermatt; and in all three, the same letters indicate the same points. In the Figures 32 and 33 I measured the angles with the greatest care,[61] from the base lines _x y_, which are accurately horizontal; and their general truth, irrespective of mere ruggedness, may be depended upon. Now in this flank view, Fig. 32, what _was_ the summit at Zermatt, _a_, becomes quite subordinate, and the point _b_, far down the flank in Forbes's view taken from the Riffelhorn, is here the apparent summit. I was for some time in considerable doubt which of the appearances was most trustworthy; and believe now that they are _both_ deceptive; for I found, on ascending the flank of the hills on the other side of the Valais, to a height of about five thousand feet above Brieg, between the Aletsch glacier and Bietschhorn; being thus high enough to get a view of the Matterhorn on something like distant terms of equality, up the St. Nicholas valley, it presented itself under the outline Fig. 34, which seems to be conclusive for the supremacy of the point _e_, between _a_ and _b_ in Fig. 33. But the impossibility of determining, at the foot of it, without a trigonometrical observation, _which is the top_ of such an apparent peak as the Matterhorn, may serve to show the reader how little the eye is to be trusted for the verification of peaked outline.