That Dalton did not rank high as an experimenter is evident from the many mistakes in matters of fact which are to be found in the second part of his "New System." A marked example of his inaccuracy in purely experimental work is to be found in the supposed proof given by him that charcoal, after being heated to redness, does not absorb gases. He strongly heated a quantity of charcoal, pulverized it, and placed it in a Florence flask, which was connected by means of a stopcock with a bladder filled with carbonic acid: after a week he found that the flask and its contents had not sensibly increased in weight, and he concluded that no carbonic acid had been absorbed by the charcoal. But no trustworthy result could be obtained from an experiment in which the charcoal, having been deprived of air by heating, was again allowed to absorb air by being pulverized in an open vessel, and was then placed in a flask filled with air, communication between the carbonic acid and the external air being prevented merely by a piece of bladder, a material which is easily permeated by gases. Dalton used a method which can only lead to notable results in natural science when employed by a really great thinker; he acquired a few facts, and then thought out the meaning of these. Almost at the beginning of each investigation he tried to get hold of some definite generalization, and _then_ he proceeded to amass special facts. The object which he kept before himself in his experimental work was to establish or to disprove this or that hypothesis. Every experiment was conducted with a clearly conceived aim. He was even willing to allow a large margin for errors of experiment if he could thereby bring the results within the scope of his hypothesis. That the _law of multiple proportions_ is simply a generalization of facts, and may be stated apart from the atomic theory, is now generally admitted. But in Dalton's mind this law seems to have arisen rather as a deduction from the theory of atoms than to have been gained as a generalization from experiments. He certainly always stated this law in the language of the atomic theory. In one of his walking excursions he explained his theory to a friend, and after expounding his views regarding atomic combinations, he said that the examples which he had given showed the necessary existence of the principle of multiple proportions: "Thou knowest it must be so, for no man can split an atom." We have seen that carburetted hydrogen was one of the compounds on the results of the analysis of which he built his atomic theory; yet we find him saying of the constitution of this compound that "no correct notion seems to have been formed till the atomic theory was introduced and applied in the investigation."