Pharmacokinetic and pharmacodynamic data are useful to interpret the efficacy and safety of a product. However, information on avian pharmacotherapy is scarce. Most of the pharmacokinetic studies on antimicrobials in chickens are based on drugs applied orally using Ôoral gavageÕ. However, it is very difficult to find a clear description of the type of catheter used for the administration of the antimicrobials. Aivlosin Ò (ECO Animal Health, London, UK), a water-soluble powder for the treatment of mycoplasmosis and other diseases in poultry and swine production, contains the macrolide tylvalosin (acetylisovaleryltylosin) (Cerdá et al., 2002(Cerdá et al., , 2006a. During pilot studies on the pharmacokinetics of tylvalosin in chickens, we found a high variation in the absorption profile not only between individuals but also within individuals when used on separate occasions. To investigate the cause of this variation, a study was designed with the objective of analysing the impact of the quality of the catheter used on the plasma kinetics of tylvalosin.A total of 12, 2-week-old chickens were used. Group A (six chickens) received tylvalosin orally, through a flexible 25-cmlength catheter (polyvinyl chloride), at a dosage of 20 mg ⁄ kg bodyweight (b.w.); group B (six chickens) received tylvalosin orally, through a rigid 25-cm-length catheter (stainless steel cannula), at a dosage of 20 mg ⁄ kg b.w. Blood samples (0.5 ml) were collected from the wing vein in heparinized tubes at the following times: 0, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 4, 6, 8, 10, 12 and 24 h postadministration.Plasma was separated and frozen at )20°C until analysis. All samples were assayed on the week following collection. Tylvalosin plasma concentration was determined using a microbiological assay (Bennet et al., 1966) with Micrococcus luteus ATCC 9314 as test micro-organism (Cerdá et al., 2006b,c). The limit of quantification of the method was 0.05 lg ⁄ ml. The method was linear between 0.025 and 10 lg ⁄ ml (r = 0.9974). Inter-and intra-assay coefficients of variation were <10%.Individual birdÕs tylvalosin concentration vs. time curves were analysed by nonlinear least square regression analysis using PCNONLIN (SCI Sofware, Lexington, KY, USA), applying a onecompartment open model with first order absorption and elimination. Initial estimates were determined using the residual method (Gibaldi & Perrier, 1982) and refitted by nonlinear regression. Most pharmacokinetic parameters were calculated using classic equations associated with compartmental analysis (Gibaldi & Perrier, 1982). Area under the plasma concentrationtime curve [AUC (0-¥) ] and mean residence time (MRT) were calculated by the trapezoidal rule (linear for the ascending and log-linear for the descending part of the curve).Pharmacokinetic parameters are expressed as mean ± standard deviation, except for half-lives which are reported as harmonic mean ± pseudo-standard deviation. Main parameters for each animal were statistically compared for the two administration methods by applying Wilcox...