The Vedanta Sutras of Badarayana, Commentary by Sankara (SBE38), tr. by George Thibaut  at sacred-texts.com
26. The extending beyond is as in the case of odour.
Just as odour, although a quality, extends beyond the odorous substance--as appears from the fact of our perceiving odour even without actually grasping flowers which are the seat of odour--so the quality of intelligence also may extend beyond the soul although the latter be atomic. It therefore is an undue stretch of inference to maintain that a quality, such as colour and the like, cannot separate itself from the substratum in which it inheres, because it is a quality; for we see that odour although a mere quality does separate itself from its substratum.--The objection that odour also separates itself from its substance
only with the substance (i.e. parts of the substance) we do not admit, because that would involve the dwindling away of the fundamental substance from which the separation of parts takes place. But that it does not so dwindle away, we conclude from its remaining in its former condition; otherwise it would lose the heaviness and other qualities belonging to it in its former state.--Well, but perhaps the separation of the particles in which odour resides is not noticed on account of their minuteness. Nevertheless the fact may be that minute odorous atoms spreading in all directions enter the cavity of the nose and there produce the sensation of smell.--This we cannot admit, because the atoms are suprasensible, and because in some cases, as, for instance, from the blossoms of the nâgakesara-tree, a very strong odour is perceived 1. According to the generally prevailing idea, moreover, it is not the odorous substance which is smelled, but ordinary people rather think that they smell the odour only.--The objection that, because we do not perceive colour and so on to extend beyond their substratum, we have no right to assume that odour does so, we cannot admit, because there is no room for that conclusion 2, on account of the (actually existing) perception (of the smell apart from the odorous substance). Logicians must shape their inferences in such a way as to make them agree with ordinary observation, not in any other way. For, to quote another instance, the circumstance that one of the qualities, viz. taste, is perceived by the tongue, certainly does not entitle us to draw the general inference that colour and the other qualities also are perceived by means of the tongue.
41:1 Single atoms could not produce any sensations; trasarenus, i.e. combinations of three atoms even could not produce lively sensations.
41:2 Viz. that smell cannot exist apart from the odorous substance, because it is a quality like colour.