A second and all-important element of all music is the use of fixed intervals which may be transposed from one point of the tone series to another and which are always recognized as equivalent. In singing, these intervals are naturally inaccurate, for intonation is uncertain and wavering and depends upon the intensity of emotional excitement. Intervals are liable to increase, as the emotions of the singers are raised to a higher pitch. It is, therefore, difficult if not impossible to say what the singers intend to sing. The musical interval may be compared with the melody of language. Most languages do not use pitch in such a way that it is an important, significant part of articulation. The use of pitch in language is more widely distributed than is generally known. It is not by any means the exclusive feature of Chinese, but it occurs in Africa as well as in America, not to speak of its familiar use in the Scandinavian languages and in ancient Greek. Theoretically it is conceivable that early human speech might have used fixed intervals and musical phrasing of vowels and voiced consonants just as well as different timbre of vowels (that is our _a_, _e_, _i_, _o_, _u_, and other vowel values), to express different ideas, but it cannot be proved that such was done. It is much more likely, according to available linguistic evidence that musical tone in language is a secondary development due to the disappearance of formative elements. We must also consider that in languages with tone, glides are of great importance and that these are not typical parts of the melodic sequence, although they occur as endings of phrases. Furthermore the intervals of speech are not fixed and vary considerably according to the position of the word in the phrase. It does not seem likely, therefore, that the melody can be derived directly from speech, as Herbert Spencer tried to do. I rather adhere to the opinion of Stumpf who demands a different origin for the fixed interval. The sustained cry is much more likely to use fixed intervals and stable tones. Whatever their origin may have been, we must recognize the existence of fixed intervals and their transponability as the fundamental requirements of all music. It is true that in some languages the value of the fixed interval is keenly felt. This is demonstrated by the so-called drum language of West Africa in which the speech melody and rhythm is repeated on drums of definite tones and where these tone sequences are understood.