Introduction

Most plumage coloration results either from structural feather properties (e.g. white, blue or UV) or from pigment molecules, mainly carotenoids and melanins, embedded in the feathers (e.g. red, yellow or black, Brush 1978). A current debate on the evolution of mate choice through plumage coloration centres on whether different shades of colour convey different kinds of information (Gray 1996; Badyaev and Hill 2000). Since birds cannot synthesise carotenoids they must be ingested (Fox 1976), and since they play an important role as precursors of vitamins (Olson and Owens 1998) and in the neutralisation of potentially harmful secondary metabolites (Von Schantz et al. 1999), it is widely recognised that carotenoid-based plumage coloration may confer information on the nutritional, parasitic and general body condition of the bearer (Hill 1999; Møller et al. 2000). Structural plumage coloration has also been shown to be condition-dependent (Gustafsson et al. 1995; Qvarnström 1998) and to be related to nutritional and parasitic condition (Potti and Merino 1996; Keyser and Hill 1999, 2000).

Data from melanin-based plumage coloration, however, is more controversial, since different studies on different species either support (Møller et al. 1996; Veiga and Puerta 1996; González et al. 1999; Poiani et al. 2000; Fiske and Richner 2002) or reject (Hill and Brawner 1998; Figuerola et al. 1999a; McGraw and Hill 2000; Hill 2000; Senar and Escobar 2002) the view that melanin coloration signals body condition. Further support for the lack of condition dependence comes from the fact that melanin pigments may be synthesised from several common dietary components (Fox 1976) and that no physiological link of melanin to general body condition has yet been described (Hillgarth and Wingfield 1997; Olson and Owens 1998).

The aim of this paper was to explore the potential value of melanin-based coloration as an indicator of nutritional condition by analysing the relationship between the size of the great tit’s (Parus major) central black breast-stripe and the rate of tail feather growth (ptilochronology). Previous studies have shown that the central black breast-stripe of the great tit is related to the quality of male parental care (Norris 1990a, 1993) and to female investment in reproduction (Norris 1990b). We used ptilochronology as a measure of nutritional condition because of its straightforward interpretation and its previously proven usefulness (Grubb 1995), but most importantly because it has already been found to be related both to carotenoid-based (Hill and Montgomerie 1994) and structural (Keyser and Hill 1999) colorations; this would thus allow, for the first time, a comparison of the three kinds of coloration within the same frame of reference.

Materials and methods

Great tits were captured in baited funnel traps from 1997 to 2000, in two contrasted areas near Barcelona, north-eastern Spain, situated approximately 10 km apart but with no exchange of birds. Age was determined according to the method of Jenni and Winkler (1994).

The size of the melanin-based black breast band was measured on digital photographs (n=144) with Image Tool software (for more technical details see Figuerola and Senar 2000). Measurement repeatability (measured as intra-class correlation) was r i=0.98, P<0.001, n=12 (see Figuerola and Senar 2000). General yellowish plumage coloration on the breast was measured with a Minolta CR200 colorimeter (Minolta Corporation 1994) using a standard D65 light flash (n=221, see Figuerola et al. 1999b). Colour was characterised according to the tri-stimulus variables of hue, chroma and lightness. The repeatability of the three colour variables was high (lightness, r i=0.85, P<0.001; chroma, r i=0.87, P<0.001; hue, r i=0.92, P<0.001; n=12; Figuerola et al. 1999b). Two data point outliers were discarded from analyses because of a very low value of hue (<75°) which was probably due to an error in measurement, although the results still hold if they are retained. We did not measure the extent of yellow coloration in the breast because the whole area is generally covered in yellow, rendering this measurement impractical.

We used tail ptilochronology to estimate the conditions of food availability and nutritional condition experienced by birds during feather growth (Grubb 1995). The second pair of rectrices was plucked for further measurement in the laboratory, where the same observer measured the width of the first ten measurable distal growth bars to the nearest 0.1 mm (see Carrascal et al. 1998). This value was transformed by dividing by ten to obtained daily growth rate (mm/day) and averaging the values for the left and right rectrices. Measurement repeatability was r i=0.99, P<0.001, n=15. We measured the first ten measurable growth bars rather than following the standard technique (Grubb 1989) because the length of the great tit feather very often would not allow for measuring ten bands. Since great tits only moult in the autumn (Jenni and Winkler 1994), rectrices and contour feathers were moulted during the same period. Since yellow and black breast feathers belong to the same feather tract, they are moulted within the same time period (Gosler 1993).

Results

Breast band size was not correlated to the tri-stimulus variables of yellow breast coloration (lightness, r=0.07, P=0.43; chroma, r=0.07, P<0.45; hue, r=0.002, P<0.98; n=144; standardising by age, sex and locality).

Width of tail growth bars was positively correlated with hue of the yellow coloration of the breast (Table 1, Fig. 1). However, width of growth bars was not correlated with the melanin-based breast band size (Table 1, age, sex, locality, year and months from moult controlled in a MANCOVA with the three colour variables and breast band size as dependent variables and width of the growth bars and months since moult entered as covariates).

Table 1. MANCOVA analysing the relationships of tail feather growth rate as measured from ptilochronology (i.e. nutritional condition) to lightness, intensity and hue of the carotenoid-derived yellow coloration of the breast and size of the melanin-derived black breast band of great tits, controlling for the effects of age, sex, locality, year and months since moult. Multivariate tests: overall regression Wilks' λ=0.84, P=0.01; fixed effects: age Wilks' λ=0.78, P<0.001, sex Wilks' λ=0.35, P<0.001, locality Wilks' λ=0.85, P=0.01, year Wilks' λ=0.36, P<0.001. Slopes are standardized to mean = 0 and SD=1
Full size table
Fig. 1.
figure 1

Relationship between tail feather growth rate, as measured from width of growth bars (ptilochronology) and A hue of the yellow carotenoid-based coloration of the breast as obtained from colorimetry, and B size of the melanin-based breast band of the great tit, as obtained from digital photography

Full size image

The two regressions of growth bars to either hue of the yellow breast and size of the black breast band significantly differed in slopes (parallelism test, F 1,352=4.42, P=0.04, controlling for age, sex and locality, including the whole data set). This shows that melanin-based coloration is less important than carotenoid-based coloration in relation to growth.

Discussion

Condition dependence is critical to the understanding of ornament evolution (Pomiankowski 1987). Ptilochronology (Grubb 1995) is a robust approach to testing the condition-signalling role of plumage, since it directly links foraging success to the expression of plumage coloration (Hill and Montgomerie 1994). Here we have shown that the width of the great tit melanin-based breast stripe was not related to feather growth rates. However, yellow carotenoid coloration (hue) of the breast was significantly related to feather growth rates. The discrepancy between our results and those of Eeva et al. (1998) is probably related to the greater repeatability and precision of our colour measurements, where the slopes of the two regressions (carotenoid and melanin) differed significantly. This supports the hypothesis that melanin-based coloration is not a good candidate for signalling nutritional condition, and that melanin- and carotenoid-based colorations might serve fundamentally different functions (Gray 1996; Badyaev and Hill 2000).

The main source of carotenes for tits are caterpillars (Slagsvold and Lifjeld 1985). Once ingested, carotenes are deposited unmodified in the feathers of tits (Partali et al. 1987). A brightly carotene-coloured great tit may therefore be directly signalling its ability to find caterpillars, a main food source for both adults and fledglings (Gosler 1993). Our data relating width of tail growth bars to yellow hue supports this higher foraging efficiency of more brightly coloured tits. Hence, it would be of direct female benefit to pair with a very yellow great tit male (see Senar et al. 2002). The study of tits could therefore neatly solve the missing link in other bird studies in relation to the question of why finding carotenes may be difficult but nevertheless highly advantageous in order to acquire a mate.