ACCURATE OBSEEVATION 283 



the way out : I always carried my note-book and pencil 

 tied to my jacket pocket, and generally walked with them in 

 my hand. It is impossible to begin observing too soon, or 

 to observe too much : if the excursion is long, little is ever 

 done on the way home ; the bodily powers being mechani- 

 cally exerted, the mind seeks repose, and being fevered 

 through over-exertion, it can endure no train of thought, or 

 be brought to bear on a subject. (H. J., i. 247.) 



As to overhasty generalisation : 



The plants gathered near the top of Wallanchoon pass 

 were species of Compositae, grass, and Arenaria ; the most 

 curious was Saussurea gossypina, which forms great clubs 

 of the softest white wool, six inches to a foot high, its flowers 

 and leaves seeming uniformly clothed with the warmest 

 fur that nature can devise. Generally speaking, the alpine 

 plants of the Himalaya are quite unprovided with any special 

 protection of this kind ; it is the prevalence and conspicuous 

 nature of the exceptions that mislead, and induce the care- 

 less observer to generalise hastily from solitary instances ; 

 for the prevailing alpine genera of the Himalaya, Arenarias, 

 primroses, saxifrages, fumitories, Ranunculi, gentians, grasses, 

 sedges, &c., have almost uniformly naked foliage. (H. J., i. 

 225.) 



As in other matters, so he sought for accuracy in drawing 

 mountain scenery, with a deliberate endeavour 



to overcome that tendency to exaggerate heights and in- 

 crease the angle of slopes, which is, I believe, the besetting 

 sin, not of amateurs only, but of our most accomplished 

 artists. 



Confessing that, as he did not use instruments to project the 

 outlines, he could not pretend to have wholly avoided this 

 snare (while the lithographer, alas, was not always content 

 to a' bide by his plain copy), he is often careful to mention the 

 angle subtended by lofty peaks in the distance, and the true 

 slope on their sides. For, as he remarks (H. J., i. 347), 



the vagueness with which all terms are usually applied to the 

 apparent altitude and steepness of mountains and precipices, 



