The wing feathers can be divided into remiges (or flight feathers) and coverts. Furthermore, both types of feathers can be subdivided into primaries and secondaries. This nomenclature is commonly used  and will be used here to describe the position of the feathers on the wing (Fig. 1A).
Five feathers p10 (Fig. 1A; for this and the further abbreviations please see List of abbreviations) from five wings of three different owls were examined; feathers from the other positions were taken from two wings of two individuals. The eight feathers p10 of the pigeon were taken from eight wings of four individuals; all other feathers from two wings of two individuals. Fig. 1A shows the positions of the examined feathers. In the following, the relevant morphological characteristics of the feathers (Fig. 1B) as outlined in the Methods section will be presented. Afterwards, measurements of the important parameters of the barbs are presented (see Fig. 1C,D). The data on the feathers of the owl will be compared to the corresponding data on the pigeon feathers.
Characteristics of feathers
The feather consists of a shaft (rachis) and a vane. The vane is divided into an outer and inner part by the rachis (Fig. 1B). The basal part of the rachis, called quill, is embedded in the bird's skin and therefore does not show any barbs. The examination included the depth, size and shape of the inner and outer vanes. The results of the measurements of the feathers p10 varied only marginally (SEM of approximately 6% for the barn owl and less than 3% for the pigeon) (Table 1). Therefore, only two feathers of the other positions were examined.
Barn owls and pigeons were nearly of the same body weight, but owl feathers were typically larger than pigeon feathers, resulting in a larger wing span, wing area and wing chord. For example, the feather p10 of the barn owl had a mean length of 24.5 cm, while the length of the pigeon's feather p10 was only 19.3 cm (Table 1). For both species, feather p10 was the second longest and was only exceeded by feather p9. This difference in size could also be found in the vane's depth. The normalised depth of the inner and outer vanes varied (Fig. 2) with the depth of the barn owl's outer vanes of the primaries being larger than that of the pigeon. Especially for the feather p10 the difference between the two species was obvious. The comparison of the outer vanes of the secondaries showed a contrary result: The outer vanes of the pigeon's secondaries were larger than those of the barn owl (compare Fig. 2A with Fig. 2B).
The depth of the vane as a function of its length was measured and the mean value was calculated. For all examined feathers (except for the secondary coverts in the pigeon), the outer vane was smaller than the inner vane (Fig. 2). The asymmetry index AId, introduced in the Methods section (Eqn. 1), revealed two morphometric characteristics (Fig. 3): On the one hand, the asymmetry in the pigeon's remige was smaller and, on the other hand, showed a much higher variation along the length of the feather compared to the barn owl. The mean asymmetry of the pigeon's remiges decreased from lateral to medial (p10, AId = -0.61; p1 AId = -0.23; s8, AId = -0.1), whereas the asymmetry of the barn owl's remiges changed only little (p10, AId = -0.66; p1, AId = -0.44; s8, AId = -0.42) (for position of the feathers see Fig. 1A).
The distribution of the coverts' asymmetry in the barn owl and in the pigeon was similar to the distribution of the pigeon's remiges (Fig. 3C,D). Only the greater secondary coverts of both species differed. A clear asymmetry could be found in the barn owl's coverts, in contrast to the pigeon's coverts, where the secondary coverts were almost symmetrical. However, the asymmetry of the coverts' vanes was typically smaller than the asymmetry of the primaries (Fig. 3C,D).
The feathers of both species showed different shapes. This can be seen in the normalised depth of vanes. For example feather p10 of the barn owl had its maximum of depth of the inner vane at approximately 70% of the vane's length (Fig. 2A). For the pigeon, the maximum of depth of feather p10 was located at 40% of the vane's length (Fig. 2B). The inner vane of the pigeon's feather p10 had a characteristic emargination at 60% of its length in all eight investigated feathers (Fig. 2B). This emargination was unique in feather p10 and could not be found in any other feather.
Only marginal changes in depth were observed in the outer vanes of the feathers p10 and p9 ranging at 10% to 90% of its length (Fig. 2A,B). This observation applied for both species. The outer vanes of the other remiges formed a double s-shape curve with two bulges in the first and last quarter. This shape was more pronounced in the barn owl's feathers (Fig. 2A,B). The bulge at the tip of the feather was mainly due to curvature of the rachis and not due to changes in the feather's margin. The bulge at the base of the feather was mainly due to the plumulaceous barbs. The combination of the s-shape of the outer vane and the ellipsoid shape of the inner vane resulted in a maximum of AId at 50% of the normalised length for most remiges of the barn owl (Fig. 3A).
The coverts of the barn owl showed noticeable changes along the vane's length. These changes occurred in the first 40% and were most obvious on the outer vane for gsc5 and gsc9 (Fig. 2C). In the first 40%, these feathers lack a closed vane because of the plumulaceous barbs. Pigeon coverts also had plumulaceous barbs, but they did not result in a larger depth at the feather's base. The pigeon coverts were slim and elongated in the normalised depiction (normalised depth of vane below 0.12), while the barn owl's coverts were wider (normalised depth of vane up to 0.2) (Fig. 2C,D).
The area of the outer and inner vanes was measured and an asymmetry index (AIa, Eqn. 2, see Methods section) was calculated as described in the Methods section. For example, the outer vane of the feather p10 of the pigeon had an area of 4.33 cm2, while the inner vane had an area of 17.58 cm2, resulting in an asymmetry index of -0.6 (Table 1). With a mean area of 10.44 cm2 for the outer vane of feather p10 and 52.15 cm2 for the inner vane, resulting in an asymmetry index of -0.67, the barn owl's vane was approximately three times larger than the pigeon's. The asymmetry of the feather depended upon its position on the wing. The highest asymmetry could be found in the feathers p10, the lowest in the feathers s4 and s8. The absolute values of AIa differed between pigeon and barn owl (Table 1, 2). For the barn owl, the asymmetry was higher, but the tendency of more asymmetric primaries was the same for both species.
Characteristics of the barbs
The vane consists of barbs. Therefore, a closer investigation of the barbs revealed the fine structure of a feather. The parameters introduced in the Methods section (Fig. 1C,D) were determined and a comparison between the barn owl and the pigeon was made in order to point out the special structures which evolved in the owl. At first, a general comparison will be given. Following, data on the special structures of the owl's feathers, which are the serrations, the fringes and the velvet-like surface, will be presented.
In both species, the outer vane was homogeneous, which means that the edge was smooth and the surface regular. However, by taking a closer look at the length of the barbs and the angle of attachment of the barbs to the rachis, two different principles of construction were revealed. The length (Fig. 4A) and the angle of attachment (Fig. 5A) of the barn owl's barbs were nearly constant. By contrast, the pigeon's barbs varied in both parameters (Fig. 4B, Fig. 5B). The normalised length of the barbs increased towards the middle of the rachis (feather centre) and decreased towards the tip (Fig. 4B). The highest increase was found at the pigeon's feather s8. Here, the length of the barbs increased by a factor of three compared to those at the base of the feather. For most feathers of the pigeon, the angle of attachment decreased from the base to the tip of the feather. The change in the angle of attachment together with the variation of the length resulted in an almost constant depth of the outer vane as can be seen in Fig. 2B. The pigeon's feather p10 and all feathers of the barn owl had an almost constant angle of attachment. The angle of attachment of the barbs of the inner vane showed a similar, but less pronounced distribution than that of the outer vane. The most acute as well as the most obtuse angle was measured at the inner, respectively at the outer vane of feather p10 for both species.
The barbs of the coverts did not reveal interspecific differences in length as big as was observed in the primaries (Fig. 4). The normalised length of the barbs of the barn owl's coverts increased slightly to the centre of the vane (Fig. 4C). The angle of attachment of the inner and the outer vanes decreased in an almost linear way in the owl and the pigeon. However, the differences between feathers from different positions were smaller in the barn owl (compare Fig. 5C with Fig. 5D).
The area outside the dotted lines in Fig. 4 represents regions of unconnected barbs. They play a decisive role in the description of barbs, because they form the plumulaceous barbs and the fringes of the feather edges. In one special case, they also form the serrations on the outer vane of the feathers p10 and gpc10 of the barn owl. This special structure will be discussed later. By and large, the density of the barbs was independent from the species (Fig. 6). Moreover, the variation of the density, depending on the position of the feather, was smaller than the variations measured in length and angle (compare Fig. 6 with Fig. 4 and 5). The largest variation occurred in the area of the plumulaceous barbs. This was typically in the range from 0 to 0.1 of the normalised length of the feather. The outer vanes of the pigeon's remiges showed the greatest variation along their length as well as compared to feathers from other positions (Fig. 6B).
The number of barbs of the remiges decreased towards the tip of the vane. In contrast to the barn owl, the number of barbs of the pigeon's remiges increased slightly at the tip of the vane. Interestingly, the outermost feathers (p10; gpc10) had the lowest density of barbs on the outer vane, but a very high density on the inner vane. This was observed for both species. Again, the interspecific differences in barb density were less distinct in the coverts than in the remiges.
The leading edge of the barn owl's feathers p10 and gpc10 formed comb-like serrations (Fig. 7A). These structures could not be found in any other feather. Each serration was formed by the tip of a single barb and might be divided into a proximal base and a distal, tooth-shaped tip (Fig. 7A). The shape of each serration was curved in a way that the tip was pointing towards the proximal end of the feather (Fig. 7A). Additionally, each serration was bent to the dorsal side. The tooth-shaped tip had a mean length of 1.8 mm (Table 1). The mean density of serrations was 18/cm, which was, naturally, equivalent to the barb density as shown in Fig. 6A (red line, outer vane). Therefore, the base of each serration was 555 μm wide. The width of the serrations tapered in an almost linear mode towards the tip, resulting in a mean width of 254 μm (+/- 4.3 SEM) at 50% of its length.
Barbs do not only build the vane, but also the edges of the feathers and thus the edges of the wing. The barn owl evolved fringes at the edges of their feathers (Fig. 7B). A fringe is formed by the tip of a barb. In the region of fringes, hook and bow radiates were present. However, they were not connected, because the hook radiates lacked hooklets. Additionally, the barb shafts became thinner towards their ends. Therefore, the barb ends could float freely (Fig. 7B). Fringes were found on the outer as well as on the inner vane of almost each investigated barn owl remige and covert. The only exceptions were the outer vanes of feathers p10 and gpc10 of the barn owl, because the barbs formed serrations. Fringes on the inner vanes were more obvious than those of the outer ones. The fringes on the outer vanes were shorter and often oriented parallel to the leading edge of a feather. A typical fringe on the inner vane of the barn owl's feather p10 had a mean length of 3.45 mm (+/-0.15 SEM) (Table 1). The fringes on the inner vane were smallest in feather p1 with a mean length of 1.68 mm (+/-0.2 SEM) (Table 1). The mean calculated density of the barbs on the inner vanes of the barn owls' remiges was 28.5/cm (+/- 0.76 SEM) resulting in a spacing of 358 μm between two fringes. The edges of the pigeon's feathers were typically smooth or slightly undulated (Fig. 7D,E). The hooklets of the radiates at the end of the barbs remained connected and therefore did not form any fringes. The only area in which unconnected barbs were found was the region of plumulaceous barbs at the base of the inner and outer vanes (Fig. 7I). Only in this area fringe-like structures were formed. In the coverts of the barn owl, the fringes were even more distinct than in the remiges (between 3.78 mm and 6.1 mm). Once again, the coverts of the pigeon did not have fringes apart from the plumulaceous barbs.
Characteristics of the radiates
Hook radiates (distal barbules) are distal extensions from the barbs, while bow radiates (proximal barbules) are proximal extensions. Each radiate can be divided into a base and a pennulum . Radiates did not show many intraspecific differences, not even between remiges and coverts. However, interspecific differences occurred (Fig. 1C). The average mean density of hook radiates over all investigated feathers was determined to 31 (+/-5) per mm for the barn owl (Table 3). The pigeon's averaged mean density was higher (44 (+/-6) per mm). The number of bow radiates (br) was lower than the number of hook radiates (hr) for all investigated feathers of both species (barn owl br/hr = 0.71; pigeon br/hr = 0.73, Table 3). We observed barn owl feathers to be more porous than pigeon feathers which could also be seen in different translucency. This higher porosity in the barn owl is a consequence of the lesser density of radiates than in the pigeon. To demonstrate this qualitatively, we dyed vanes of barn owl and pigeon feathers with black hair tinting lotion to avoid influences of different keratin colours. Afterwards, a transparent foil with writing (barn owl, pigeon) was placed between feathers and light source. The feathers were illuminated from below and each feather was photographed using the same resolution and exposure time. In contrast to the pigeon, the barn owl lettering could easily be recognised (Fig. 7D,I). A quantitative description of the different porosity in both species can be found in Table 3 (number of barbules).
The pigeon's hook radiates had typically more hooklets (5) than those of the barn owl (3) (Table 3). The number of hooklets decreased towards the tip of the feather (not shown) on both, the inner and outer vane. For both species, the hook radiates were always attached in a more acute angle than the bow radiates (Table 3, Fig. 1C). The largest angles were found in the barn owl and varied between 24 and 60 degrees (Table 3). The pigeon's hook radiates were attached in an angle of 22 to 49 degrees (Table 3). The interspecific difference in the angle between bow radiates and barb shaft was smaller (barn owl: 15–39 degrees, pigeon: 16–34 degrees) (Table 3). No clear differences could be found between remiges and coverts.
Length and shape of the radiates changed in the region of the serrations (Fig. 8A). They shortened towards the tip of the barb (Table 3) and the number of hooklets decreased to zero. The base of the bow radiates merged directly into the pennulum without a clear differentiation between both. Therefore, in Fig. 9A the total length of the bow radiates at 75% barb length is listed. The separation of barbs is mainly due to the lack of hooklets, shorter radiates and a change of the barb shaft in its form and shape. One serration tapered towards the tip and was bent in two different directions. As seen in Fig. 7A, the barb shaft was bent towards the feather base (calamus) and also to the dorsal side (not shown). Apart from the outer vanes of the feathers p10 and gpc1 (see above), every inner and outer vane of the barn owl was equipped with fringes. In the area of the fringes, the hooklets on the hook radiates were missing as well. By contrast to the serration, the bow and the hook radiates were not shortened. Thus, the fringes consisted of the unconnected elongated radiates and the barb shafts, leading to a fluffy structure (Fig. 7B).
The velvet-like appearance of the barn owl feathers was predominantly a consequence of elongated pennula (Fig. 1C, Fig. 7C). The pennula covered the barb shafts. This clearly differentiated the velvet-like structure of the barn owl's feathers from the homologue area in the pigeon (Fig. 7H), where a straight alignment of the barb shafts was obvious. A velvet-like structure could not be found on any of the pigeon's feathers. The length of the pennula was measured for feathers p10, s8 and gpc1 of both species on the outer vane as well as on the inner vane (Table 4). The mean length of the pennula of the inner and outer vane was larger in the barn owl than in the pigeon (Fig. 8, Fig. 9; Table 4). The pennula of the outer vane were always shorter than those of the inner vane (Fig. 8, Fig. 9; Table 3). For instance, an average pennulum of the barn owl's outer vane was 601 μm, while the average pennulum of the pigeon was 79 μm (Table 4). There was a greater divergence in the length of the inner vane, with the pigeons' pennulum length being 136 μm and the barn owls' pennulum being 1271 μm (Table 3). The pigeon's pennula were short and did not extend to the shaft of the next barb (Fig. 1C). By contrast, the barn owl's pennula overlapped up to four neighbouring barbs shafts (Fig. 1C). This difference in length was due to the length of the pennula and not to the length of the base of the radiates (Fig. 8, Fig. 9). We found that the length of pennula augmented in length in areas which were covered by another feather (Fig. 9). Remiges and coverts were arranged in an imbricate way meaning outer vanes overlapped wide ranges of inner vanes of adjacent feathers.
In both species the pennula of the outer vane of feather p10 decreased in length towards the tip (Fig. 8A). The radiates' length of the inner vane of the pigeon's 10th primary remained nearly constant (Fig. 9A). By contrast, the barn owl's radiates of the inner vane of the 10th primary increased in length towards the tip, especially the hook radiates (Fig. 8A, Fig. 9A). A similar effect could be noticed for all other investigated feathers of the barn owl. For instance, the feather gpc1 was positioned at the wrist (Fig. 1A). Hence, the main covered areas of this feather were found at its base (covered mainly by feathers of the median coverts (Fig. 1A)) and its inner vane (covered by feather gsc1 (Fig. 1A)). The pennula on the inner vane (Fig. 9C) were longer than those on the outer vane (Fig. 8C). From the density of barbs (e.g. 33.29/cm on the inner vane of p10) and the density of hook radiates (29.98/mm on the inner vane of p10) in the barn owl, we calculated the average density of the pennula to 99.8/mm2 (Table 4). The average density for the homologue structure of the pigeon was 152.5/mm2 (Table 4). By comparing the homologous structures of both species (Table 4) it was found that the density of pennula was higher for the pigeon than for the barn owl.