"In 35 records of this song, the number of notes varies from 9 to 25, the average being 17. Only one song has a true trill in it, that is, notes so rapid that they cannot be counted. Each song is of either two or three parts, each part composed of a series of notes on the same pitch and in uniform rhythm. The parts differ from each other in pitch, time, or loudness. In a number of songs, one of the parts is a repetition of 2-note phrases. Loudness generally increases to the end of the song, but sometimes the reverse is true. Some songs rise in pitch to the end and others fall; my records are about evenly divided in this matter. A typical three part song would be something like _tit it it it it it it pita pita pita pita pita chit chit chit chit chit_. "Pitch varies from G´´´ to E´´´´, or four and a half tones. Single songs vary from half a tone to three and a half tones, averaging one and a half. The length of songs varies from 1-4/5 to 3-1/5 seconds. An individual bird may sing a dozen different variations of the song in a short time. On the other hand, I have heard three birds in one tree singing alternately, the songs of all three being exactly alike so far as my ear could determine." Francis H. Allen gives me his impression of the song as follows: "The song bears some resemblance to that of the Nashville warbler, but is easily distinguished. I have written it _wi-chip wi-chip wi-chip wi-chip, wi-chip wi-chip chip chip chip chip chip chip chip_. The higher notes in the middle sometimes appear to be monosyllables, and they are sometimes omitted. The series of _chips_ at the end are very emphatic, and the last one is perhaps accented somewhat. All the notes are staccato." Various other renderings of the song have appeared in print, but they all give the same impression of a variable, loud, striking song which, once learned, can be easily recognized. The bird is a very persistent singer rivaling the red-eyed vireo in this respect. Bowdish and Philipp (1916) write: "As a basis for estimating the frequency of song repetition, counts were kept on three singing birds for a period of 5 minutes each, with a result of 32, 36, and 22 songs, respectively, within the period. In one instance, a bird was observed to sing while on the wing, repeating the song twice in the course of a short flight." Albert K. Brand (1938) found the pitch of the Tennessee warbler's song to be well above the average, the approximate mean count being 6,600 vibrations per second, the highest note about 9,150 and the lowest 4,025; this compares with an approximate mean of 8,900 vibrations per second for the black-poll warbler, and about 4,000 for the average passerine song.