Useofarti®ciallyapplied n-alkanesasmarkers fortheestimationofdigestibility,foodselection andintakeinpigeonsColumba livia) J.-M.Hatt a,* ,R.W.Mayes b ,M.Clauss c ,M.Lechner-Doll d a Division of Zoo Animals and Exotic Pets, University of Zurich, Winterthurerstrasse 268, CH-8057 Zurich, Switzerland b Macaulay Land Use ResearchInstitute, Craigiebuckler, Aberdeen AB15 8QH, UK c Institute of Animal Physiology, Physiological Chemistry and Animal Nutrition, Ludwig Maximilians University, D-80539 Munich, Germany d Institute of Zoo Biology and Wildlife Research, D-10315 Berlin, Germany Received28November2000;receivedinrevisedform19June2001;accepted7August2001 Abstract Theexperimentwasconductedtodeterminetheaccuracywithwhichfoodselection,intake,and apparent digestibility of protein could be estimated with n-alkanes in comparison to observed intakes and digestibility estimated through total excreta collection. Four pigeons Columba livia) wereusedasstudysubjectsandfourpigeonsascontrols.Allanimalswereindividuallycaged.In Trial 1 the birds were fed a pelleted diet ad libitum which had been labelled with the n-alkanes octacosaneC 28 2000ppmdrymatterDM)),dotriacontaneC 32 1500ppmDM),andhexatriacontane C 36 2000ppmDM).Thebirdshad6daysofadaptationtothedietfollowedby4daysoftotalexcreta collection. On the basis of observed intakes and total excreta collection n-alkane recoveries were estimatedandcomparedtopublisheddataforchickens.Theintakesobtainedusinghentriacontane C 31 )andC 32 correctedforrelativerecoverieswereverysimilartotheobservedintakes.Whenalkanes werenotcorrectedforrelativerecoveries,thealkanemethodsigni®cantlyunderestimatedtheintake. TherecoveryrateofC 36 n-alkanewas80%andtheproteindigestibilitycoef®cientsobtainedafter correction with the recovery rate did not differ signi®cantly from digestibility obtained with total excretacollection. InTrial2selectiveintakewasmeasuredinfourbirdsofferedthreedifferentpelletedfeedsad libitumfor3daysofadaptationfollowedby5daysofsamplecollection.Pellet1waslabelledwith n-alkanesC 28 2000ppmDM),C 32 1500ppmDM),andC 36 2000ppmDM),pellet2waslabelled with hexacosane C 26 3000ppm DM), and pellet 3 with triacontane C 30 3000ppm DM). Very accurateestimatesofpelletselectionwereobtainedusingthepatternsof n-alkanesinexcreta,being inaveragewithin 2.5%oftheactualpelletselectioninindividualbirds. AnimalFeedScienceandTechnology 942001)65±76 * Correspondingauthor.Tel.: 41-1-635-83-42;fax: 41-1-635-89-01. E-mail address:jean-michel.hatt@access.unizh.chJ.-M.Hatt). 0377-8401/01/$±seefrontmatter # 2001PublishedbyElsevierScienceB.V. PII:S0377-840101)00294-2