Anatomy

14697580, 2017, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/joa.12595 by INASP/HINARI - INDONESIA, Wiley Online Library on [30/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Journal of
J. Anat. (2017) 230, pp734--742 doi: 10.1111/joa.12595

Distribution of the characteristics of barbs and

barbules on barn owl wing feathers
Matthias Weger and Hermann Wagner
Institute of Zoology, RWTH Aachen University, Aachen, Germany

Abstract
Owls are known for the development of a silent flight. One conspicuous specialization of owl wings that has
been implied in noise reduction and that has been demonstrated to change the aerodynamic behavior of the
wing is a soft dorsal wing surface. The soft surface is a result of changes in the shape of feather barbs and
barbules in owls compared with other bird species. We hypothesized that as the aerodynamic characteristics of
a wing change along its chordwise and spanwise direction, so may the shape of the barbs and barbules.
Therefore, we examined in detail the shapes of the barbs and barbules in chordwise and spanwise directions.
The results showed changes in the shapes of barbs and barbules at the anterior and distal parts of the wing,
but not at more posterior parts. The increased density of hook radiates at the distalmost wing position could
serve to stiffen that vane part that is subject to the highest forces. The change of pennulum length in the
anterior part of the wing and the uniformity further back could mean that a soft surface may be especially
important in regions where flow separation may occur.
Key words: biomimetics; feather morphology; flow separation; silent flight.

American barn owl (Tyto furcata pratincola) or the eagle

Introduction
The silent flight of owls (Strigiformes) is a model system for
biomimetic research (Vad et al. 2006; Bachmann et al. 2011;
Kla€n et al. 2012; Winzen et al. 2012, 2014a,b,c, 2015; Clark
et al. 2014; Winzen & Roidl, 2016). Morphological special-
izations such as the leading-edge serrations, the trailing-
edge fringes, a soft dorsal surface and a wing shape opti-
mized for slow flight are well known, but their exact func-
tion is still a matter of debate and often not well
understood (Fatio, 1866; Mascha, 1904; Graham, 1934; Sick,
1937; Hertel, 1963).
The surface of bird wings is formed by pennaceous feath-
ers. Feathers consist of a shaft (scapus), which is subdivided
into the calamus and the rachis, and barbs. Barbs originate
at the rachis. The barb shafts give off barbules (also called
radiates). The radiates can be interconnected to form a
closed vane, as found for example in pennaceous feathers.
The radiates consist of a base and a filamentous end called
pennulum. In most bird species, the pennula are short and
do not overlap neighboring barb shafts. The wing surface in
these cases feels rough. By contrast, owls such as the
Correspondence
Matthias Weger, Institute of Zoology, RWTH Aachen University,
Worringerweg 3, D-52074 Aachen, Germany. E: weger@bio2.rwth-
aachen.de
Accepted for publication 3 January 2017
Article published online 3 March 2017
owl (Bubo bubo) have elongated pennula on the radiates of
their wing feathers so that neighboring barb shafts are cov-
ered (Bachmann et al. 2007; Chen et al. 2012). The surface
of owl wings feels soft. Because this softness is reminiscent
of velvet, the surface of the wings of the owls is often called
velvet-like. We adopt this terminology in the following.
Bachmann et al. (2007) described several morphological
characteristics of the pennula: (i) variation with feather
type; (ii) shorter on the outer vane than on the inner vane
on most positions; (iii) variation along the vane. Bachmann
et al. (2007) described these structural properties on three
different wing feathers (10th primary feather, 8th secondary
feather, 1st greater primary covert). What these authors did
not do, and what is missing so far, is to relate the character-
istics of the velvet-like surface to wing position. This is
interesting because it was shown that the addition of a vel-
vet-like surface to an artificial wing reduced flow separation
and enabled boundary-layer control (Kla €n et al. 2009;
Winzen et al. 2012, 2014a). Since the aerodynamic charac-
teristics of a wing, such as the boundary layer, Reynolds
number, shear stress and pressure distribution, change in
both the spanwise and chordwise directions (Usherwood
et al. 2005; Kla €n et al. 2009, 2012; Winzen et al. 2012,
2014a, 2015; Winzen & Roidl, 2016), the shapes of the barbs
and barbules that form the velvet-like surface may do so,
too. Thus, here we tested the hypothesis that the character-
istics of the velvet-like surface change along the chordwise
and spanwise directions of the T. furcata pratincola wing.

© 2017 Anatomical Society

 14697580, 2017, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/joa.12595 by INASP/HINARI - INDONESIA, Wiley Online Library on [30/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Barb characteristics on barn owl wings, M. Weger and H. Wagner 735

The investigations were performed independent of feather Furthermore, for reasons of comparability between different
type. The data presented in the following demonstrate an samples, the following deviations from the strict geometrical deter-
influence of wing position on the characteristics of the vel- mination of the position examined were applied.

vet-like surface at anterior and distal wing positions. • In the case of several overlapping feathers, the dorsal-most
feather at a given measurement position which was exposed
to the airflow, was investigated.
Materials and methods • If at a selected position a covered part occurred (as indi-
cated by its whitish coloration), we interpreted this as an
The investigation was carried out on eight wings of T. furcata prat-
artifact of the preparation, and chose the closest uncovered
incola. The wings were obtained from animals of the colony at the
feather part in spanwise position at the given chordwise
Institute of Biology II at RWTH Aachen University that had been sac-
position for analysis. This was done because major differ-
rificed at the end of other experiments under a local authority per-
ences were observed in the quality of the velvet-like surface
mit [Landespra €sidium fu € r Natur, Umwelt und Verbraucherschutz
between the covered and uncovered feather parts (Sick,
Nordrhein-Westfalen, Recklinghausen, Germany (LANUV)]. The
1937; Bachmann et al. 2007). We chose only to investigate
wings were prepared as they appear in gliding flight. We chose this
the velvet-like surface that occurs at the brownish, uncov-
wing form because in gliding flight the wing is fully extended and
ered feather part because this feather part is exposed to
because it offered the possibility to compare our data with earlier
the airflow, and because the covered, whitish part does not
work (Bachmann et al. 2007, 2011).
occur in all feathers.
• Samples were investigated at the central or proximal part of a
Selection of positions barb. This took into account that the radiates and pennula
tend to be reduced at the barb tips (Bachmann et al. 2007).
To obtain samples at distinct wing positions, a wing was subdi-
vided lengthwise into different sampling positions. The wings
were positioned according to the description in Bachmann et al. Preparation of samples
(2011; Fig. 1). A two-dimensional coordinate system was estab-
Preparation of samples included the following steps (Fig. 2). First, a
lished in the plane formed by wingspan and wing chord. The
sample containing a short rachis length and vane area and several
chordwise axis was set as the x-axis with the leading edge being
barbs was cut off the rachis of a given feather and fixed on a small
set to 0 and the trailing edge being set to 1. The y-axis was set
aluminum block with double-sided adhesive tape so that the area
parallel to the wingspan, with the position of the root chord
of the vane was horizontal. Care was taken that the feather vane
being set to 0 and the wing tip being set to 1. Four spanwise
was closed.
positions were investigated in this study, at 0.2, 0.4, 0.6 and 0.8
of the y-axis. At each spanwise position, 12 measurement posi- Afterwards, images were taken from a horizontal perspective
tions were investigated in the chordwise direction. Note that the with a microscope camera (Scope Tek DCM500, Hangzhou Scopetek
origin of the x-axis was always at the leading edge at the respec- Opto-Electric Co., Hangzhou, Zhejiang Province, China) fixed to a
tive spanwise position, whereas 100% of the x-axis was at the stereomicroscope (Nikon SMT 10-A, Nikon Corp., Tokyo, Japan) at a
trailing edge. Thus, the absolute length of the x-axis differed for magnification of 259.
the different spanwise positions. The first six measurement posi-
tions had a distance of 0.05 in the normalized x-axis, starting at
0.05 and ending at 0.3 of the normalized x-axis. This was done
because most effects of the flow field, including boundary layer
separation at artificial owl-based wing models, were found in this
particular region (Kla €n et al. 2009, 2012; Winzen et al. 2012,
2014a). Between 0.4 and 0.9 of the normalized x-axis, the mea-
surement positions were set in steps of 0.1 of the x-axis. This
resulted in 48 measurement positions per wing. In the following
we will use a combination of the respective values along the
x-axis and the y-axis to describe the measurement positions
on the wing. For example, the position x05y20 describes the
chordwise position (x-axis) 0.05 at 0.2 of the wing length (y-axis).
It is important here to remind the reader that different feather
types such as marginal coverts, lesser, median and greater sec- Fig. 1 Measurement positions. The velvet-like surface was investi-
ondary coverts, primary coverts, alular coverts, alular feathers, pri- gated at four different parts of the wing (0.2, 0.4, 0.6 and 0.8 of
mary remiges and secondary remiges occur on a bird wing. Most of wing length) which is indicated by the corresponding lines. Twelve
these feathers can be numbered by their position, which holds chordwise measurement points were set at each investigated wing
especially for remiges. We decided to categorize the different part. The first six measurement positions had a distance of 0.05 in the
feather types and formed three categories (Fig. 1). The first feather normalized x-axis, between 0.05 and 0.3 of the normalized x-axis.
category includes only the marginal coverts. The second category Measurement points between 0.4 and 0.9 were set in steps of 0.1 of
includes lesser, median and greater secondary coverts, primary cov- the x-axis. Note that samples were only taken from feather parts that
erts and alular coverts. For reasons of simplicity, in the following we exhibited the uncovered velvet-like surface. The feathers that occur at
shall refer to this category as covert feathers. The third category these wing parts include marginal coverts, covert feathers (primary
comprises the flight feathers, which consist of primary remiges, and secondary coverts) and flight feathers, depending on wing part
secondary remiges and alular feathers. and wing position.

© 2017 Anatomical Society

 14697580, 2017, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/joa.12595 by INASP/HINARI - INDONESIA, Wiley Online Library on [30/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
736 Barb characteristics on barn owl wings, M. Weger and H. Wagner

B C
Fig. 2 Preparation of measurement samples.
(A) Flight feather of Tyto furcata pratincola.
(B) Cut off part of the outer vane. The arrow
from B to A points to the feather part that
was investigated. Note that the barb next to
the left outer barb was partially cut to obtain
a better view on complete barb. (C) Image of
barbs of the investigated vane part. The
arrow from C to B points to the part of the
sample that was used for the recording.

Quantification of velvet-like surface

The images were analyzed in IMAGEJ (National Institutes of Health,
Bethesda, MD, USA). Several parameters were extracted from the
photographs (Fig. 3). The length of a pennulum was measured
from the base to the tip of the pennulum. The length of the hook
radiates was measured from the basis of the barb to the origin of
the pennulum. The length of the pennulum was divided by the
base of the hook radiate to estimate the pennulum–base ratio.
The overlap (O) of the pennula (Sick, 1937; Bachmann et al. 2007)
describes the extent to which pennula reach over neighboring
barbs, determined by the formula:

rHR
O¼
dbarb
with the range of the hook radiates (rHR) being measured as the
distance between the barb and the tip of the pennula, normal to
the barb axis and dbarb being measured as the distance between
two adjacent barbs. The dbarb was measured in regions where two
adjacent barb ran in parallel, which occurred throughout the major
part of barb length, but not close to the distal end. Finally, the den-
sity of the pennula was determined as the number of barbules per
mm for hook radiates.
Fig. 3 Overview of measured parameters. The investigated barb
parameters are presented in the sketch. The overlap was estimated as
the ratio between the range of hook radiates and the distance
between two barbs. The density of the hook radiates was determined
as the number of hook radiates per mm. h.r., hook radiates; b.r., bow
radiates.
in structure. Thirdly, all 48 measured spanwise and chordwise posi-
tions were compared (1128 test cases) to find possible differences
between wing positions.

Statistical analysis Results

The data obtained were analyzed statistically with EXCEL (Microsoft
Occurrence of feathers at investigated wing parts
Corporation, Redmond, WA, USA) and MATLAB (The MathsWorks,
Inc., Natick, MA, USA). The data of the T. furcata pratincola speci- Eight wings were examined in this study. We chose to sub-
men were tested statistically using a multiway ANOVA in MATLAB. A
divide the occurring feathers in three categories: marginal
Tukey HSD test was used for post-hoc analysis. The multiway ANOVA
coverts, covert feathers, and flight feathers. The distribution
investigated three factors. First, the differences between all individ-
uals were checked. This resulted in 28 test cases. In a second step, of the different feather types was similar for the investi-
the difference among the three different feather types was tested gated owl specimen (Fig. 4). Note, however, that due to
(three test cases). This was done to determine possible differences our position-based method of analysis it sometimes

© 2017 Anatomical Society

 14697580, 2017, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/joa.12595 by INASP/HINARI - INDONESIA, Wiley Online Library on [30/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Barb characteristics on barn owl wings, M. Weger and H. Wagner 737

happened that in different specimens, different feather
types occurred at a given position.
The marginal coverts and covert feathers occurred pre-
dominantly at the frontal part of the wing at 0.2 of the
wing length and to a lesser extent also at 0.4 of the wing
length (Fig. 1). The area covered by the marginal coverts
was 8.2% of the total wing area. Barb shape in marginal
coverts varied a lot. In this feather type, prolonged hook
radiates were present, but the singular barbules were com-
paratively thin (Fig. 5A). The coverts were present over the
high cambered wing part where the patagium, skeletal
parts and adjoining tissue of the wing are present as well
(Fig. 1). Coverts were found in 26.3% of the total wing
area. The flight feathers made up the largest part of the
total wing area. They were found in the distal and posterior
parts of the wing and covered 65.4% of the total wing area.
The outer half of the wing was mainly formed by feathers
emerging from the arm and hand parts of the wing. The
distal part of the wing (0.8 of wing length) was completely
made up by flight feathers, which were also present at the
caudal part of the wing at all other positions. The marginal
coverts and secondary coverts of T. furcata pratincola
occurred over 55  2.14% of the chord at 0.2 of wing
length, 42  5.49% of the chord at 0.4 of the wing length
and 12  9.64% of the chord at 0.6 of the wing length.
Distribution of the velvet-like surface
The barbs of the feathers investigated were compared at
different spanwise and chordwise wing positions (Fig. 1).
We used six different morphological parameters to describe
the properties of the velvet-like surface of an owl wing. The
variation of these parameters in chordwise and spanwise
direction was analyzed.
First we compared the data obtained from the eight indi-
vidual wings. Two of 28 comparisons exhibited significant
differences (multiway ANOVA followed by a Tukey post hoc
test, P < 0.05). This low number of differences led us to con-
clude that data of all wings may be pooled for further anal-
yses. No differences between feather types were observed.
The pennula had a similar length at the majority of the
measured spanwise and chordwise positions in our study,
with an average length of 793  166 lm (Fig. 6). In gen-
eral, the pennulum length was constant over the wing and

Fig. 4 Feather types at different measurement positions. The feather types that occurred at the different measurement positions are depicted for
all individual investigated wings. Marginal coverts occur at the frontal part at 0.2 and to a lesser extent at 0.4 of the wings. Secondary coverts
occur mainly at the frontal and median part of 0.2–0.4 of the wings. The flight feathers are found at the distalmost and caudal part of the wings.
Note that some data points are missing which is shown at the empty square. T.f.p: Tyto furcata pratincola; y/c: spanwise measurement position.

A B C

Fig. 5 Feather types. The three different feather types are depicted in this figure. (A) Example of a marginal covert. (B) Example of a covert
feather. (C) Example of a flight feather. A lower barbule density and a resulting higher porosity can be seen for the marginal coverts in A).

© 2017 Anatomical Society


738 Barb characteristics on barn owl wings, M. Weger and H. Wagner
exhibited no significant difference between the majority of
different spanwise and chordwise positions. One conspicu-
ous deviation from this scheme was observed at the four
frontal chordwise positions at 0.2 of the length of the wing.
The length of the pennula at x20y20 was the greatest of all
investigated positions (1053  205 lm) (Fig. 6). The pen-
nula at this position were significantly longer than pennula
of several measurement positions (multiway ANOVA followed
by a Tukey post hoc test, see Supporting Information
Fig. S1, vertically and horizontally aligned squares based at
x20y20). We observed this difference after analysis of five
wings, which we initially regarded as a high enough sample
size. Since the difference occurred only at the medial and
frontal part of the wing, we were worried that the differ-
ence might represent a false positive and decided to ana-
lyze three more wings. We observed the same difference.
This led us to conclude that the difference was real. In sum-
mary, pennulum length did not change along the major
part of chordwise and spanwise directions, with the excep-
tion of a change at the leading edge of the most proximal
investigated wing part, up to a chordwise position of 0.2.
14697580, 2017, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/joa.12595 by INASP/HINARI - INDONESIA, Wiley Online Library on [30/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
The bases of hook radiates between individuals exhib-
ited differences in four of 28 tested cases. No differences
were observed between the three feather types. The
length of the base of the hook radiates from where
the pennula emerged showed no positional differences
despite a length range between 131 and 392 lm
(224  39 lm, Fig. 7).
The ratio between the pennula and the base of hook
radiates can be used to describe the elongation of the pen-
nula. The mean ratio between pennula and hook radiates
was 3.58  92. No difference between individuals and
between feather types were observed for this ratio. The
ratio between pennula and hook radiates showed few
positional differences (Fig. 8). Only five of 1128 positional
combinations exhibited significant differences. These differ-
ences were between x05y20 and x20y20, as well as between
x15y40, x25y60 and x10y80, and between x10y80 and
x40y80 (Fig. S2). Since significance changed when only the
first five wings were considered compared with when all
eight wings were considered, some of these differences
may represent false positives.
Fig. 6 Pennulum length. The length of the
pennula was measured at all four wing
positions on 12 different chordwise positions.
Fig. 7 Hook radiate length. The length of the
hook radiates was measured at all four wing
positions on 12 different chordwise positions.
© 2017 Anatomical Society

Barb distance was different in two out 28 cases for
T. furcata pratincola individuals, again leading to the con-
clusion that the data may be pooled. In our investigation
the distance between adjacent barbs was significantly
higher in flight feathers (292  42 lm) than in covert
feathers (258  43 lm, multiway ANOVA, P < 0.001). Barb
distance varied little (284  43 lm) for most positions on
the feather. Yet, several positional differences occurred
(Fig. 9). Forty-two of 1128 positions exhibited significant
positional differences. These differences were especially
conspicuous between position x05y40 and most (nine of
12 comparisons) of the chordwise positions at 0.8 of the
wing length (Supporting Information Fig. S3, horizontally
aligned squares in sector y40y80). This suggested that barb
distance was increased at 0.8 wing length. Similar differ-
ences occurred between caudal chordwise positions at 0.8
of the wing length and frontal to central chordwise posi-
tions at the other spanwise positions (vertically aligned
squares in sectors y20y80, y40y80 and y60y80). Both obser-
vations are consistent with earlier findings (Bachmann
et al. 2007).
Fig. 8 Ratio between pennula and base of
hook radiates. The ratio of pennula to hook
radiates was estimated for all four wing
positions on 12 different chordwise positions.
Fig. 9 Barb distance. The distance of two
adjacent barbs was measured at all four wing
positions on 12 different chordwise positions.
© 2017 Anatomical Society
14697580, 2017, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/joa.12595 by INASP/HINARI - INDONESIA, Wiley Online Library on [30/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Barb characteristics on barn owl wings, M. Weger and H. Wagner 739
The overlap describes the extent to which the pennula
reach over neighboring barbs. If the pennula were to
exactly reach the adjacent barb, the overlap would be 1.
We observed a mean overlap of 2.1  0.58 (Fig. 10). When
comparing different individuals, no differences were
observed. In contrast, differences were observed between
the different feather types. The overlap on flight feathers
(2.26  0.52) was significantly higher than the overlap on
marginal coverts (1.64  0.49, multiway ANOVA, P < 0.001)
and covert feathers (1.98  0.55, n-way ANOVA, P < 0.001).
Two of 1128 positional comparisons exhibited significant
differences [x05y20 differed from x60y20 (multiway ANOVA,
P = 0.018) and x70y40 (multiway ANOVA, P = 0.02)] Support-
ing Information Fig. S4).
The number of hook radiates per mm was the sixth inves-
tigated parameter (Fig. 11). The number of hook radiates
per mm can be used to describe the density of the velvet-
like surface and further indicates the porosity of the feather
vane. The mean value was 25.62  4.81 hook radiates/mm).
No differences between the individuals were observed.
However, the different feather types exhibited a

740 Barb characteristics on barn owl wings, M. Weger and H. Wagner
significantly different number of hook radiates. The flight
feathers (26.98  4.11 per mm) had a significantly higher
number of hook radiates per mm compared with covert
feathers (23.51  6.03 per mm, multiway ANOVA, P = 0.0013)
and marginal coverts (19.2  3.95 per mm, multiway ANOVA,
P < 0.001). Covert feathers had a significantly higher num-
ber of hook radiates per mm compared with marginal cov-
erts (multiway ANOVA, P < 0.001). In marginal coverts the
hook radiates were widely spread. This results in a higher
porosity. The differences in between the three feather types
also resulted in a higher number of positional differences.
Of 1128 positional combinations, 158 were significantly dif-
ferent (Supporting Information Fig. S5). As demonstrated
by the thick horizontal stripe of squares in Fig. S5, the dif-
ferences were present between the first three chordwise
positions (905, 910, 915) at 0.2 of the wing length (y20) in
comparison with the majority of the other positions. Addi-
tionally, as indicated by the vertical and horizontal stripes
of squares based at x05y80, the hook radiate density was
especially high at the first two chordwise positions at 0.8 of
wing length.
14697580, 2017, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/joa.12595 by INASP/HINARI - INDONESIA, Wiley Online Library on [30/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Discussion
The velvet-like surface on wings of T. furcata pratincola was
investigated with stereomicroscopy at preassigned wing
positions to evaluate whether the characteristics of the sur-
face depend on wing position. We observed a change in
pennulum length at medial and anterior wing positions
and an increase in the density of hook radiates at the distal-
most wing position. In the following, we shall discuss the
results in the context of the existing literature and attempt
to draw conclusions with regard to biological aspects and
technical applications.
Distribution of the velvet-like surface on an owl wing
Several attempts have been made qualitatively to describe
(Sick, 1937) or quantitatively to compare the occurrence of
the elongated pennula that are fundamental for the forma-
tion of a velvet-like feather surface (Bachmann et al. 2007;
Chen et al. 2012). The feather-type related analysis of Bach-
mann et al. (2007) detected a lower number of barbules
Fig. 10 Overlap. The overlap of the pennula
in regard to the adjoining barb shafts was
measured at all four wing positions on 12
different chordwise positions.
Fig. 11 Number of hook radiates. The
number of the hook radiates per mm was
measured at all four wing positions on 12
different chordwise positions.
© 2017 Anatomical Society

and longer pennula on T. furcata pratincola wings in com-
parison with pigeon (Columba livia domestica) wings, but
no differences between different primary and secondary
barn-owl feathers. Our position-dependent analysis demon-
strated an influence of wing position on the characteristics
of the velvet-like surface for anterior positions. These differ-
ences were missed in earlier studies. This shows that it is
important to take feather position into account when it
comes to the characteristics of the velvet-like surface.
Apart from the frontal positions, our results suggest a
high uniformity of the specialized barb modifications on
owl wings. We interpret this uniformity such that an evolu-
tionary pressure for the creation of a position-dependent
morphology existed only for frontal positions.
One interesting finding of this study was the change of
the pennulum length at the most proximal wing positions
investigated, reaching a maximum at 0.15 or 0.2 chord
length. This was paralleled by an increase in the number of
hook radiates for the most medial lengthwise position.
Interestingly, the start of the boundary layer on an owl-
based wing model was found slightly behind this chordwise
position (Winzen et al. 2012, 2013, 2014a,b). This leads to
the speculation that the presence of a velvet-like surface
may be especially important in regions where flow separa-
tion may occur (see also below).
Another conspicuous deviation is the increased density of
hook radiates at the frontal part at the distalmost wing
position. This observation reflects the high number of hook
radiates that were observed earlier (Bachmann et al. 2007)
on 10th primary feathers. One function of this increased
density could probably be to stiffen that vane part most
likely subject to the highest forces. Future experiments will
have to clarify whether this conclusion is correct.
Our findings of the pennulum and hook radiate length
on flight and covert feathers correspond well to the obser-
vations reported in Bachmann et al. (2007), which can now
be extended to the major part of the wing.
We chose to restrict our investigation to uncovered
feather parts because of large differences in the characteris-
tics of the pennula in uncovered and covered areas. For
example, the shape of the pennula in the brownish, uncov-
ered feather parts differed from the shape in the covered,
whitish feather parts (Bachmann et al. 2007).
Functional aspects of the velvet-like surface
The velvet-like surface on owl wings has been related to
reduction in flight noise. Lilley (1998), for example, pro-
posed that the downy surface helps in dampening the aero-
dynamic effects that create flight noise. Vad et al. (2006)
demonstrated noise reduction in the human hearing range
in an RAF-6E profile with an artificial downy surface that
was inspired by the surface of owl feathers. Chen et al.
(2012) conducted noise measurements on feathers of
B. bubo, and the common buzzard (Buteo buteo), a
© 2017 Anatomical Society
14697580, 2017, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/joa.12595 by INASP/HINARI - INDONESIA, Wiley Online Library on [30/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Barb characteristics on barn owl wings, M. Weger and H. Wagner 741
member of the family of diurnal raptors. Feathers of
B. bubo exhibited better sound-absorbing properties in
flight than the feathers of the buzzard (by approximately
10 dB for frequencies below 1 kHz). Clark et al. (2014)
demonstrated that an artificial fabric based on a velvet-like
surface exhibited noise reduction capabilities up to a maxi-
mum of about 30 dB, depending on the fabric model, sur-
face structure, flow velocity and frequency. Jaworski (2016)
suggested a model in which filaments of the velvet surface
act as inclined fibers that achieve noise reduction by the
interaction of fiber motion and a moving line vortex.
Although these data demonstrate the noise-reducing
capabilities of the velvet-like surface, recently a further func-
tion has been described. Wind tunnel studies showed
reduced flow separation for artificial wings covered with
€n et al. 2009; Winzen et al.
artificial velvet-like surfaces (Kla
2012, 2014a,b). Thus, the velvet-like surface has also an aero-
dynamic effect. Whereas these authors used a uniform vel-
vet-like surface on the whole wing, we show here that at
anterior and proximal wing positions the pennula are
shorter. They reach their full length only in wing regions
where flow separation may occur. Thus, it might be interest-
ing to test wings with a velvet-like structure that resembles
the positional distribution on natural wings. This suggestion
is also based on the findings of Winzen et al. (Winzen et al.
2014b, 2015; Winzen & Roidl, 2016) that demonstrate an
even better aerodynamic behavior of natural owl wings
than of artificial wings equipped with a velvet-like surface.
Conclusions and outlook
Several questions relating to the evolution and function of
the velvet-like surface remain unanswered. First, if the major
advantage of the velvet-like surface was a reduction of
flight noise in combination with passive hearing, why did it
evolve in nightjars (Sick, 1937), which mainly hunt insects
that are not expected to perceive flight noise? To our best
knowledge, nothing is known about the situation in other
night-active birds such as frogmouths, potoos and owlet
nightjars. Secondly, if the major advantage was a reduction
of flow separation, why did the velvet-like surface not
evolve in other bird species? A more thorough comparative
study may help to answer part of these questions.
Within the owl clade, it is known among ornithologists
that the small, diurnal pygmy owl (Glaucidium passerinum)
does not possess a velvet-like surface (H-H Bergmann, pers.
comm.). We have investigated feathers of the diurnal little
owl (Athene noctua) and the nocturnal B. bubo and
observed longer pennula in the three owl species than in C.
livia domestica, but shorter pennula in A. noctua than in
T. furcata pratincola and B. bubo (M. Weger & H. Wagner,
unpubl. obs.). Thus, it would be helpful to collect data on
more owl species.
It would also be interesting to make a further comparison
of the velvet-like surface on covered and uncovered feather

742 Barb characteristics on barn owl wings, M. Weger and H. Wagner
parts. The velvet-like surface of covered feather parts exhi-
bits qualitative differences to this surface and has been
related to friction noise reduction (Bachmann, 2010) and
was probably the inspiration for the ‘canopy’ structure used
in the experiments of Clark et al. (2014). The differences
between the two different types of surfaces have been
mentioned in previous studies, but a quantitative compar-
ison and a possible relation to function is missing and
should be examined further in future studies.
Acknowledgements
We would like to thank J. Fischoeder for the help on data acquisi-
tion and S. Brill for the help with the statistical analysis.
Author contributions
M.W. and H.W. conceived and designed the experiments.
H.W. contributed reagent, materials and analysis tools.
M.W. performed the experiments. M.W. and H.W. analyzed
the data and wrote the paper.
Conflict of interest
The authors declare no conflict of interest.
References
Bachmann T (2010) Anatomical, Morphometrical, and Biome-
chanical Studies of Barn Owls’ and Pigeons’ Wings. Aachen:
RWTH.
Bachmann T, Kla €n S, Baumgartner W, et al. (2007) Morphomet-
ric characterisation of wing feathers of the barn owl Tyto alba
pratincola and the pigeon Columba livia. Front Zool 4, 23.
Bachmann T, Mu € hlenbruch G, Wagner H (2011) The barn owl
wing: an inspiration for silent flight in the aviation industry?
In: SPIE Smart Structures and Materials+ Nondestructive Evalu-
ation and Health Monitoring. International Society for Optics
and Photonics; 2011. pp. 79750N
Chen K, Liu Q, Liao G, et al. (2012) The Sound Suppression Char-
acteristics of Wing Feather of Owl (Bubo bubo). J Bionic Eng
9, 192–199.
Clark IA, Devenport W, Jaworski JW, et al. (2014) The Noise
Generating and Suppressing Characteristics of Bio-Inspired
Rough Surfaces. AIAA/CEAS 20th Aeroacoustics Conference,
Atlanta, GA. 2014.
Fatio V (1866) Des diverses modifications dans les formes et la
coloration des plumes. Geneva (Switzerland): Ramboz and
Schuchardt.
Graham RR (1934) The silent flight of owls. J R Aeronaut Soc 38,
837–843.
Hertel H (1963) Struktur, Form, Bewegung. Mainz: Otto Kraus-
kopf-Verlag.
14697580, 2017, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/joa.12595 by INASP/HINARI - INDONESIA, Wiley Online Library on [30/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Jaworski JW (2016) Vortex sound generation from flexible
fibers. AIAA/CEAS 22nd Aeroacoustics Conference, Lyon,
France. 2014 (p. 2752).
Kla€n S, Bachmann T, Klaas M, et al. (2009) Experimental analysis of
the flow field over a novel owl based airfoil. Exp Fluids 46, 975–989.
Kla€n S, Burgmann S, Bachmann T, et al. (2012) Surface structure
and dimensional effects on the aerodynamics of an owl-based
wing model. Eur J Mech B – Fluids 33, 58–73.
Lilley GM (1998) A study of the silent flight of the owl. AIAA
Paper 4, 1–6.
€
Mascha E (1904) Uber die Schwungfedern. Z Wiss Zool 77, 606–651.
Sick H (1937) Morphologisch-funktionelle Untersuchungen u € ber
die Feinstruktur der Vogelfeder. J Ornithol 85, 206–372.
Usherwood JR, Hedrick TL, McGowan CP, et al. (2005) Dynamic
pressure maps for wings and tails of pigeons in slow, flapping
flight, and their energetic implications. J Exp Biol 208, 355–369.
Vad J, Koscso G, Gutermuth M, et al. (2006) Study of the aero-
acoustic and aerodynamic effects of soft coating upon airfoil.
JSME Int J 49, 648–656.
Winzen A, Roidl B (2016) Combined particle-image velocimetry
and force analysis of the three-dimensional fluid-structure
interaction of a natural owl wing. Bioinspir Biomim 11, 026005.
€n S, Klaas M, et al. (2012) High-speed PIV mea-
Winzen A, Kla
surements of the influence of artificial surface structures on
the near-wall flow field of 3D wing models based on an owl
geometry. 16th Int Symp on Applications of Laser Techniques
to Fluid Mechanics, Lisbon, Portugal.
Winzen A, Klaas M, Schro € der W (2013) High-speed PIV measure-
ments of the near-wall flow field over hairy surfaces. Exp Flu-
ids 54, 1–14.
Winzen A, Klaas M, Schro € der W (2014a) High-speed particle
image velocimetry and force measurements of bio-inspired
surfaces. J Aircr 52, 471–485.
Winzen A, Klaas M, Schro € der W (2014b) PIV measurements com-
paring natural and model owl wings. Exp Fluids 46, 975–989.
Winzen A, Roidl B, Kla €n S, et al. (2014c) Particle-image
velocimetry and force measurements of leading-edge serra-
tions on owl-based wing models. J Bionic Eng 11, 423–438.
Winzen A, Roidl B, Schro € der W (2015) Particle-image velocime-
try investigation of the fluid-structure interaction mechanisms
of a natural owl wing. Bioinspir Biomim 10, 056009.
Supporting Information
Additional Supporting Information may be found in the online
version of this article:
Fig. S1. Statistical differences of the pennulum length between
different measurement positions.
Fig. S2. Statistical differences of the pennula-base ratio between
different measurement positions.
Fig. S3. Statistical differences of the barb distance between dif-
ferent measurement positions.
Fig. S4. Statistical differences of the overlap between different
measurement positions.
Fig. S5. Statistical differences of the number of hook radiates
between different measurement positions.
© 2017 Anatomical Society