We must now consider the most vital and difficult portion of our subject. How are we to determine the cola? The colon is the very soul of the rhythm. The period is generally too long for the ear to receive it as one artistic impression. The foot is too short; moreover, the mere foot too often tends to play one false: irrational syllables and τονή are against us. But the colon is neither too long nor too short. The colon-division serves the same purpose as non-commissioned officers in a regiment, or the determination of watersheds in geography—it gives a sense both of grouping and of control. What precisely _is_ a colon? It is as much of a strophe as can be uttered without making a new start. It is the embodiment of rhythm, as the foot is the embodiment of metre. In other words, it is a series of feet bound into a rhythmical unity by the presence of one main ictus. Three questions, then, arise. (i) What is an ictus? (ii) Which is the main ictus of a series? (iii) Can we with certainty determine the beginning and end of a colon when we have identified the main ictus? (i) Ictus is stress-accent. The ictus of any single word is usually obvious. In the word “maritime” it falls upon the first syllable, in “dragoon” upon the second, in “cultivation” upon the third. In πάντων, λυσαμένοις, and κατάπαστος, it falls upon the first, second, and third respectively. Greek metre is based upon quantity, but Greek rhythm (like all other rhythm) is based upon ictus. A strophe can, and must, be scanned foot by foot on quantity alone; but when we go beyond the foot-division to exhibit the structure of the whole, we must refer to ictus and nothing but ictus—for structure is an affair of cola, and the colon is created by the main ictus. (ii) Among the many word-ictuses of a considerable passage, a few will be found which are heavier than the rest. These are simply the ictuses of the most important words. Each of these prominent ictuses gathers the neighbouring minor ictuses into a group round itself. We should begin then by fixing some obvious example, one (that is) where the main ictus is unmistakable, and on this basis attempt, by the help of the correspondences which we expect, to determine other main ictuses. The strophe will thus gradually fall into cola. This leads us at once to our third question. (iii) Can we with certainty determine the extent of each colon? Unfortunately no simple invariable rule can be given for the settlement of this vital point. But certain useful principles may be mentioned. (_a_) A well-trained ear is the chief guide. Intelligent and careful reading aloud of an English prose-passage will show this. Take first (the best-known version of) a famous sentence of John Bright:— The Angel of Death is abroad in the land: you may almost hear the beating of his wings.