Vibration Frequencies of Passerine Bird Song•

Vol.1938 55]J BRAND,Vibration Frequencies ofPassefine Song 263 VIBRATION FREQUENCIES OF PASSERINE BIRD SONG• BY ALBERT R. BRAND T•E recordingof bird songson film producesa medium from which soundscan be studiedobjectively; hence we needno longerrely entirely on aural impressions.There, on the film, is printedin black and white, on a soundtrack of varyingdensity or varyingarea, dependingon whichof two methodsof recordinghas been used, the picturethat the physicalvibrations have madeupon the sensitizedfilm. The numberof thesevibrations per seconddetermines the pitch of a sound. •Forexample: a soundof 4096 vibrationsper secondis pitchedat the highestnote of the pianokeyboard (C?) and is said to have a frequencyof 4096. Bird songsmay alsobe re- cordeddirectly on wax, from whencephonograph records can be manu- factured. This method,while satisfactoryfor purposesof reproduction,is not so accurateas film for objectivestudy. In 1934 in Germany,and in 1935-37 in England, Ludwig Koch recordedbird songsdirectly on wax. He reportson the pitchof someten passerincbird songs(Nicholson, E. M., andKoch, Ludwig, 'Songs of Wild Birds,'London, p. xxviii, 1936).The data are, however,fragmentary; he givesonly the highestand lowestfrequencies found. Thosehe reportson are lower,so far as the highestnotes are con- cerned,than the resultsof my studieswould indicatefor Americanbirds. The highestfrequency he mentionsis 5500 for the Yellow-hammer(ibid., p. 3). But one gathersfrom what he writes that it is unlikely that he is getting all the highspresent in bird song;probably somewhere in his ap- paratushe losesor partially suppressesthem. Thereare a numberof places wherethis lossis likely to occur. Unlessmicrophone, amplifiers and re- corderare very carefullycoordinated and speciallyconstructed, this is in- evitable; in addition, it is doubtful whetherit is practicalto constructa disk recordersuch as Koch uses,that will reproducemuch over 5000 vibrations per second. In other words,songs containing frequencies over 5000 vibrationswould probablybe reproducedonly in part in disk recording. On film, however,we have recordedin Nature, bird and insectsounds of a frequencyof over 10,000vibrations, and we havemade laboratory tests up to 17,000. It is thereforeunsafe to includethese European recordings in our calculations,at leastuntil numericallyfar more studieshave beenmade than are yet reported,and until we have definiteproof that theserecordings includethe higherfrequencies. In almostevery casemy studieswere madefrom originalnegative film, producedby either the variable-densityor the variable-area recording • Read at the Fifty-fifth Stated Meeting of the American Ornithologists' Union, Charleston, November 17, 1937. 264 BR•N.,Vibration Frequencies ofPassefine Song [April[Auk method. Perhapshalf a dozen studieswere made from positive prints. Variable-area and variable-densityfilms give essentiallysimilar results, with the main differencethat varying-area track producesmore visual detail than oneof varyingdensity. As at presentwe are concernedwith the pitch-giving frequencyof the songsof birds, either type of recordingis satisfactoryfor study. Were we discussingthe important factor of bird song,quality, or what musicianscall timbre, we shouldfind variable-area film more valuable for study. But thus far, in most of the studies,I have not attempted to analyze the sound further than to discoverthe main or pitch-giving frequency; and these remarks are restricted to pitch. The study of quality is highly involved,requiring in most easesa study of com- plicated wave shapes in order to determine the overtones or harmonies. This necessitatesintricate mathematical calculations. It is agreed that quality--that whichenables us to recognizea Meadowlark'snote immedi- ately, and not to confuseit with, let us say, a warbler'snote given on the samepitch--is of fundamentalimportance. At somefuture time I hopeto have sufficientmaterial to offer the resultsof investigationson that phase. For the present,let us restrict our considerationsto the main pitch-giving frequency. In the past few years, in the Laboratory of Ornithologyat Cornell Uni- versity, we have studied intensivelyunder a low power of magnification (25 diameters),several hundred pieces of film comprisingthe songsand calls of about one hundredspecies of birds. In orderto simplifymatters, I am in this paper, consideringsongs only. That is, I have taken what is generallyaccepted as song,and giventhe results;and in order to simplify further, I shall consideronly passefinesong. Thus I have reducedthe results so that we have studiesof fifty-nine passefinebird songs. The familiesand the numberof speciesincluded are as follows:Tyrannidae, six; Corvidae,one; Paridae, one; Troglodytidae, three; Mimidae, one; Turdidae, six; Bombyeillidae,one; Sturnidae,one; Vireonidae,three; Compsothlypi- dae, sixteen;Ieteridae, six; Thraupidae, one; Fringillldae, thirteen. These specieswere selected at randomand give a fair generalsample of passerine song. Almost no bird notesremain on the samepitch from instant to instant. That is, birds almostinvariably slur either upwardor downward,and often in both directions. I have attempted to estimatethe bird songson what appearsto be the approximatemean of all frequenciesin a note, and also the mean of the song. Though a note may contain a very short portion (sometimes0.002 secondsin duration) of either excessivelyhigh or rather low frequency,I have attemptedto approximatethe main pitch of the note, which in many eases,is the pitch of longestduration and sometimesthe pitch which is Vol.1938 55]J BRAND,Vibration Frequencies ofPasserinc Song 265 loudest. In many notesthe pitch is changingconstantly, and in thesecases I have, as nearly as possible,estimated the average. While it is possibleto count accuratelythe frequencies,these estimatesare merely approxima- tions. However, they are correctwithin, let us say, two or three per cent. Thus a note which I have called a frequencyof 5000 vibrationsmight be 4850 or 5150, but the error will rarely be greater than that, and in most easesis considerablyless. Thesedifferences are so slight as to be negligible, being lessthan one-quarterof a note. Errors may occur,in the following ways. In making a recording,the motor that drives the film through the recordermay occasionallybe slightly slow or fast. However, this is very carefullywatched, and it rarely happensthat the film speedis not accurate; and when it is not, the differenceis so slight as to be negligible. In esti- matingfrequencies by inspeetlonno attempt is made at absoluteaccuracy. Here again the introducederror is very slight. This has been proved a number of times by estimatingindependently, on separatedays, the results of the studiesof successivesongs of the same bird. The results of these cheeksare always the same. The introducederror of estimationis very small; often two separatelymade countsgive absolutelyidentical results. We thus have assurance that the method used is accurate for all practical purposes. Of the fifty-nine songsincluded in this study we find, by the method described,that the averagepitch of passerincbirds' songis 4280 vibrations, or a quarter note higher than C7, the highestnote of the piano keyboard. This is about an octaveand a half higher than the highestnote that an operatleeoloratura soprano can reach. The averageof 4280 for fifty-nine speciesgives, I believe, a fairly correct estimationof all passerincsong. However, there is great variation within the group. The averageof Wood Warbler songsis considerablyhigher: for the sixteenspecies included in thesestudies, it was found to be 5350, or a shadehigher than the E above the highestnote on the piano keyboard. This is two noteshigher than the averageof passerincsong. In the thrush group, of which the songsof six specieswere studied, the averagewas about 2890, which is a shadeunder the highestF sharpon the piano,or F sharp6, approximatelythree notes lower than the average. Ieterine songalso is comparativelylow, and we foundin the six speciesstudied, exactly the same, i.e., 2890, as in the Turdidae. Fringillinebird songon the wholeis high, althoughit includesthe Cardinal whoseclear, whistling tune averageslower than one would presume. The Cardinal songis 2800, or a shadeabove the highestF (F6) of the keyboard. The songsof thirteenfringilline species studied, showed an averageof 4800 or just underD sharp7, the D sharpabove the highestnote (C?)of the keyboard. The highestsong studied was the Black-pollWarbler's. Its averagenote 266 BRANV,Vibration Frequencies ofPasserins Song œArril]'Auk was8900 or overan octaveabove the highestnote on the piano. It is mid- way betweenC sharps and Ds. The Blackpoll also producedthe highest avian frequencystudied, 10,225, a note about a quarter under Es. Other high averageswere countedfor birds in the followingorder: Grasshopper Sparrow,8600; CedarWaxwing, 8400; Blue-wingedWarbler, 7675; Black- burnJanWarbler, 7300; Black and White Warbler, 6900. The differencein frequencybetween the first three, Blackpoll, GrasshopperSparrow and CedarWaxwing, is only about a half a note, and is so smalland the pitch so high, that an ear would have to be remarkably accurateto recognizethe pitch difference. COUNT OF PITClt OF PASSERn•IE BIRD SONGS For further referencefor thoseinterested in the subject,the following table givesfigures for all songsreported on in this paper. The birds are listed under their common names as given in the fourth edition of the A. O. U. 'Check-list,'grouped by familiesin systematicorder.

Load more