It is important to note that by this treatment of the two instruments we have not changed the way in which they usually indicate temperature. For thermometrical measurement is in actual fact never anything else than a recording of the movement of the indicator from one level to another. We choose merely to take a certain temperature level - that of melting ice or something else - as a fixed point of reference and mark it once for all on the instrument. Because we find this mark clearly distinguished on our thermometers, and the scales numbered accordingly, we fail to notice what lies ideally behind this use of the same zero for every new operation we undertake. What the zero signifies becomes clear directly we start to work with thermometers not marked with scales. For in order to be used in this form as real thermometers, they must be exposed on each occasion first of all to some zero level of temperature, say, that of melting ice. If we then take them into the region of temperature we want to measure, we shall discern the difference of levels through the corresponding movement of the column. The final position of the column tells us nothing in itself. It is always the change from one level to another that the thermometer registers - precisely as does the sense of warmth in our hands in the experiment just described. Hence we see that in the ordinary operation with the thermometers, and when we use our hands in the prescribed manner, we are dealing with the zero level in two quite different ways. While in the/two instruments the zero level is the same, in accordance with the whole idea of thermometric measurement, we make a special arrangement so as to expose our hands to two different levels. So we need not be surprised if these two ways yield different results. If, after placing two thermometers without scales in hot and cold water, we were to assign to each its own zero in accordance with the respective height of its column, and then graduate them from this reference point, they would necessarily record different levels when exposed to the tepid water, in just the same way as the hands do. Our two hands, moreover, will receive the same sense-impression from the tepid water, if we keep them in it long enough. Seen in this light, the original experiment, designed to show the subjective character of the impressions gained through the sense of warmth, reveals itself as a piece of self-deception by the onlooker-consciousness. The truth of the matter is that, in so far as there is any subjective element in the experience and measurement of heat, it does not lie on the side of our sense of warmth, but in our judgment of the significance of thermometrical readings. In fact, our test of the alleged proof of the absolute superiority of pointer-readings over the impressions gained by our senses gives us proof of the correctness of Goethe's statement, quoted earlier, that the senses do not deceive, but the judgment deceives.