Mechanical Properties of Contour Feathers

Feathers are probably the most complex derivatives of the integument to be found in any vertebrate animal. However, though their structural characteristics have been extensively investigated (e.g. see the review by Lucas & Stettenheim, 1972), knowledge of the relationship between the functions of feathers and their morphology is still incomplete.

The major parts of the feather are the shaft and the vanes which it supports on either side. For flight feathers, the vanes provide an aerofoil surface; for contour feathers they cover and insulate the body. The shaft of contour feathers consists of two layers: a porous, inner medulla, and a solid, outer cortex. The cross-section gradually changes shape, from approximately rectangular near the base, to elliptical as the shaft tapers towards the tip (Lucas & Stettenheim, 1972).

Purslow & Vincent (1978) have studied the relationship between the structure of a flight feather shaft and its bending behaviour. Although contour feathers are not necessarily subjected to such large forces as are flight feathers, they must still contend with a great variety of loads. This has been observed recently for domestic hens (Hughes, 1978).

This paper reports measurements of the mechanical properties of contour feathers, dealing a gradient along the shaft which has not previously been described.

Young’s modulus was then determined in the dorso-ventral plane of bending for a similar set of samples using the static cantilever method (Fig. 1).

In bending, the shape of the cross-section has a great bearing on stiffness. The resistance to bending is determined both by the amount of material in the cross-section and by its distribution about the neutral axis of the section, as well as by the modulus of elasticity. The first two factors are measured by the second moment of area of the section I, which gives more significance to material further away from the neutral axis than to that close to it (see Fig. 1). In the present study established formulae for the calculation of I were used (Alexander, 1968). Equations for a rectangle and an ellipse were used for the proximal and distal regions of the feather shaft, respectively (Fig. 2).

One end of each sample was embedded in Araldite which was then rigidly clamped. The point of emergence from the mount was made abrupt and clear, to reduce error in the measurement of the bent length. The shaft was bent by a load applied 10 mm from the fixed end by means of a delicate spring. The deflexion at the point of application of the load was measured to + 0·02 mm with a travelling microscope. The dimensions of the shaft (Table 1), as well as the bent length, were measured at ten points along the length by using a second microscope (× 100 magnification, as previously). Initially, increasing loads were applied to one sample for each species. For all the subsequent samples a load was used which fell in the linear region of the load-deflexion curve.

Finally, two complementary procedures were adopted to confirm the discrepancies between the Young’s moduli in extension and bending. Part of the cortex was removed from one set of samples by abrasion with emery paper, while in a second set most of the medulla was bored out with a hypodermic needle. The Young’s modulus in bending was determined as previously, with values for the second moment of area corrected for the quantity of material removed.

Since the length of shaft chosen for the measurement of Young’s modulus in bending was short, shear deformation could be significant. For short beams Ugural & Fenster (1975) suggest a correction factor equal to 0·75 (1 + n/2) (h²/L²), where n is the Poisson ratio for the material, and h and L are the thickness and length of the sample respectively. The Poisson ratio for the feather is likely to be similar to the published value of 0·3 determined for wool fibres (W.I.R.A., 1955), while the largest value of h/L for the present samples was 0·1. Therefore the maximum correction applicable to these measurements is less than 1 %. This is insignificant in relation to the other sources of error in the measurements and has therefore been neglected.

The measurements of Young’s modulus (E) in extension are presented in Table 2. The interspecific variance is large and the values for the distal region are systematically higher than those for the proximal region. E ranged from 45 to 181 MN m⁻² and from 210 to 682 MN m⁻² for the proximal and distal regions, respectively. In both cases the highest values were for gull feathers, and the lowest values were for the chicken, turkey and pheasant feathers.

In the proximal region, the modulus for extension is higher than Young’s modulus for bending (Table 3); in the distal region the converse is true. This suggests that the the birds studied, the cortex of the proximal region has a lower modulus than the medulla, while for the distal region this is reversed. Both the scatter between species and the variation with shaft position are much greater in the case of the Young’s modulus in bending than in extension. Values obtained for the proximal and distal regions in bending were from 5 to 24 MN m⁻² and 457 to 1849 MN m⁻², respectively.

The treated shafts showed a similar qualitative difference in elastic properties between proximal and distal regions. The results of the treatments confirm the above suggestion of different relative contributions by medulla and cortex in the two regions; for removal of that part of the feather which either has a comparatively low or high E value should respectively increase or decrease the modulus measured for the remaining structure.

Standard errors in all the experiments were of the order of 10%. The major part of this was associated with the measurement of shaft dimensions. In extension the area of the shaft was measured at the point of fracture, but the strain before fracture varied to some extent between shafts. To account for this, samples whose strain at breakage was more than 2 standard deviations from the mean strain of 10% were discarded.

Mechanically, a feather shaft can be treated as a tapering beam. Its elastic properties may therefore be determined by its deformation under load, as treated mathematically in engineering texts (see Ugural & Fenster, 1975). The equations usually assume that the material concerned is perfectly elastic, homogeneous and isotropic. When this assumption is true, then the stiffness of the material will be the same in bending as in extension. In a perfectly elastic system the relationship between stress and strain is linear and reproducible, shows no hysteresis and is independent of the direction in which the test piece is cut from the sample. Most animal fibres, including feathers, cannot be expected to fulfil these criteria since they are heterogeneous, anisotropic and viscoelastic. For the contour feathers studied here the initial part of the stress-strain relationship was linear and reproducible and all measurements were taken in this region.

Many determinations of Young’s modulus have been made on mammalian hair acres. For example, Khayatt & Chamberlain (1948) obtained mean values for the Young’s modulus of human hair of 3·6 GN m⁻²and 1·9 GN m⁻² in stretching and bending, respectively. The values of E measured in the present work are lower than those generally reported for wool and hair, perhaps because wool and hair are both of an alpha-keratin type structure, whereas feather is of a beta-keratin type (Woods, 1955). Purslow & Vincent (1978) quote E values estimated from cantilever beam tests of 7·75 and 10 GN m⁻² for the shaft cortex of two types of pigeon flight feathers. These are about two orders of magnitude higher than the values reported here for intact feathers in bending. However, in measurements made on intact shafts, the air-filled cavities within the medulla (Lucas & Stettenheim, 1972), which could not be accounted for in cross-sectional area measurements, would tend to reduce the stiffness of the shaft as compared to that of the shaft cortex. This suggests that in both bending and extension the effective stiffness depends on the feather structure as much as on the material.

Purslow and Vincent (1978) reported measurements of the second moment of area of the shaft of pigeon flight feathers. Taking a mean value for I of 2·3 mm⁴ for a 400 g bird from their results, this is three orders of magnitude greater than the equivalent value for a contour feather from a 1600 g bird in the present study. Purslow and Vincent also found that the cortex provided most of the resistance to bending.

The variation in the elastic properties of the shaft layers with position (Fig. 3) may be related to the bending forces to which they are subjected. A flexible thin-walled beam, such as the proximal region of a feather shaft, would tend to fail by elastic instability if it were not supported internally. Struts have been noted in certain bone cavities which have been presumed to serve this purpose (Alexander, 1968). The substantial medulla in the basal region of a feather shaft could play the same role.

At the outer end of a contour feather the vane is generally narrower and any bending moment about the tip will be small. These factors allow tapering of the feather shaft, thus contributing to lightness. However, the shaft in this region must maintain the. vanes as part of a wind resistant coat cover, so maximum rigidity for a given cross Section would be an advantage. The distal region is relatively more stiff than the base of the feather, as shown by the present results. In the distal region, the cortex typically provided 60% of the area. Strength in bending for this region is therefore more likely to be limited by the tensile strength of keratin rather than by elastic instability. Thus a wide-bore tube can be lighter than a narrow tube of the same rigidity, for the wall can be thinner. The results reported here suggest that the outer end of a contour feather can be considered as a wide-bore tube with the medulla being of minor significance. The lowest Young’s modulus for the distal region was observed for the gull (at P > 0·05). In contrast to the feathers of the other birds the medulla also had a higher modulus than the cortex in this region, as confirmed by the experimental treatments. Such a structure may be advantageous in maintaining the aerodynamics of the coat during flight.

Purslow & Vincent (1978) suggest that the differences in flexural stiffness which they found among pigeon flight feathers are more likely to be due to changes in morphology rather than changes in the material of the shaft. In the present study, however, the cortex and medulla of contour feather rachis have both been shown to have mechanical properties which vary with position along the shaft. This may be due to variations in the keratin but further study is required to resolve the matter.

The work was supported by a British Egg Marketing Board Research and Education Trust studentship, supplemented by the University of Nottingham. The author also wishes to thank J. A. Clark for his interest and guidance.