BULLETIN OF THE ALLYN MUSEUM
THE ALLYN MUSEUM OF ENTOMOLOGY
Number.. 42 10 May 1977
OBSERVATIONS ON MALE U-V
REFLECTANCE AND SCALE
ULTRASTRUCTURE IN PHOEBIS (PIERIDAE)
Arthur C. Allyn, Jr.
Director, Allyn Museum of Entomology
John C. Downey
Biology Department, University of Northern Iowa, Cedar Falls, Iowa and
Research Associate, Allyn Museum of Entomology
The genus Phoebis and its relatives are medium to large sized sulphurs
limited to the new world. Their colors, as with many pierids, range from white
to yellow, orange and red, and mixtures of these. While most of the visible color is
probably due to pigments in the wing scales, some species show a brilliant iri-
descence in the near ultra-violet range of the spectrum, which is apparently
caused by optical interference in a lamellar system found on the ridges of the outer
wing scales. The latter ultrastructure has been described in other Pieridae: in
Eurema by Ghiradella et al 1972i. Colias by Hirata et at (1957, 19:i9 1960), by
Silherglied and Tavlor 1197:1 and by Ghiradella (1974); and in Pieris by Obara and
Hidaka (1968). Sellier (1971) published an SEM photograph of Genopteryx scales
without further comment. In a series of papers Descimon (1965, 1966a, 1966b,
1969, 1971, 1976) contributed significantly to the understanding of pigment for-
mation in the Pieridae, and to the cytoplasmic ultrastructure of scales. About
25 pierid species have males in which U-V reflectance has been indicated, mostly by
photographic techniques. Mazokhin-Porshnyakov (1954, 1957, 1969) studied
the U-V reflecting properties of representatives of several families of butterflies,
but did not relate color to scale type. He recorded reflectance percentages and
patterns in five wing areas of several species including two Phoebis (philea and
argante) and concluded that the U-V reflectance was "undoubtedly perceived" by
Since Phoebis has marked color variants and overlapping forms in both sexes
in most of its species in both the U-V and the visible parts of the spectrum, it was
our feeling that it would be an excellent study group in which to attempt to relate
reflectance and scale ultrastructure. Further, since the male wings show marked
color reflectance, but seem to be less variable in color characters than the females
of Phoebis species, we hypothesized that the contribution of scale ultrastructure
to color might be more easily sought in the males.
Brown (1929, 1931: 10) was of the opinion that Phoebis and its close relative
Aphrissa. are very recently evolved genera, which show much diversification
under varying environmental conditions. Knowledge of the nature and origin
of color production in Phoebis will not only contribute to the taxonomic and
evolutionary concepts within the genus, but will broaden our understanding of
the biological significance of ultrastructure and its contribution both to general
morphology and to behavior.
Over 50 years ago Mattram and Cockayne (1920) noted that scales contained
"fluorescent pigments" in the ultraviolet part of the spectrum which are not visible
to the unaided human eye. They suggested (loc. cit.: 38-39) that the fluorescent
dimorphism between the sexes aids in sexual recognition. Cockayne (1924) demon-
strated U-V reflectance in a wide variety of lepidoptera.
Except for studies such as Koehler (1941) where U-V light was used as an
irradiating source to study induced variation, it is somewhat surprising that after
Cockayne's introductory work, there was a thirty-year gap in studies involving
U-V reflectance and possible advantages to the species. In the late 1960's and
early 1970's, there has been an increasing number of studies on ultraviolet re-
flectant scales, and their optical properties.
Since 1965 when Nekrutenko first suggested and used ultraviolet reflectance
photography as a taxonomic tool, it has been used in several pierid groups (Ferris,
1972, 1973; Silberglied and Taylor, 1973; Scott, 1973, Nekrutenko, 1970a, 1970b,
1973). Ultraviolet video-viewing has also been suggested as an exciting teaching
aid (Eisner et al, 1969: 1174). Photographic techniques have also been used in
detecting gynandromorphs (Nekrutenko, 1966.)
MATERIALS AND METHODS
Most of the described taxa in the genus were closely examined visually and
photographs were taken of select species using appropriate filters (see below) to
cut out all but the near ultra-violet wavelengths.
Based on both visual color and U-V reflectance properties, we eventually
selected three species for detailed SEM studies: Phoebis philea (Linn.), P. thalestris
(Illig.) and P. avellaneda (Herr-Schaef.). These had an increasing component of
orange and red scales from the almost monotone lemon-yellow phase of philea,
through the heavily infused orange-yellow of thalestris to the red cast of avellaneda,
where the yellow scaling is reduced to a relatively small fraction of the visual com-
ponent. Further, the U-V reflectance patterns on the upperside of these three
species have consistent and significant differences. We were convinced that dif-
ferences of the magnitude observed between them would also be observable with
scanning electron microscopy if they were correlated in any way with ultrastructure.
Hirata and Kubota (1957) noted that some differences between their observa-
tions on ultrastructure in Colias and those of other authors were explained by
the methods used, rather than just the specimens involved. It thus seemed advisable
to briefly note the procedures and equipment used so that results could be con-
sistently duplicated for interpretation.
Reflectance and absorbance were measured on a Beckman (DB-GT) double-
beam grating spectrophotometer with a 100 specular reflectance accessory. A
Beckman 10-inch recorder charted samples in a continuous readout from 700 nm
to 190 nm. The SEM instrument used was a JSM-U3. Samples were coated with
60/40 gold-palladium in a DE10 vacuum evaporator. Dried museum specimens
Separation photographs were made with Wratten tight-band filters numbered
18a, 50, 75, 73 and 72b. Exposure times were balanced to a standard gray scale
using Panatonic X film, which is essentially insensitive to deep red and infra-red
(650 nm and higher). Originally such photographs permitted determination of
the dominant frequencies in specific wing areas. One millimeter square samples
were then excised from areas (see Fig. 4), color photographed and enlarged to 200
diameters for precise orientation of scales in SEM examination. As techniques
were perfected, much of this original orientation procedure was shortened, parti-
cularly as expertise was gained at recognizing specific areas, colors, and scales.
The ultra-violet photographs were taken with an 18a Wratten filter using a
25w/s ring flash. Focal length was increased 3 mm to compensate for focal length
difference in the shorter wavelengths. Scale vocabulary follows that given in
Downey and Allyn (1975).
RESULTS AND DISCUSSIONS
Photographic and Spectrophotometric Comparisons
Photographs "a" through "e" were taken with Wratten tight-band filters
of Phoebis philea (Fig. 1), P. thalestris (Fig. 2) and P. avellaneda (Fig. 3). Fig-
ures la, 2a and 3a were obtained with an 18a filter which peaks at 360 nm (310-
400 nm). A comparison of the three figures shows the similarities and differences
in the patterns of reflectance in the near ultra-violet range for the three species.
Note that thalestris has the largest of the high intensity reflective spots in the
forewing. The scent patch in the costal area of the hindwing of avellaneda Fig
Dorsal surface of male Phoehis photographed using tight-band filters. Figs.
1. P. philea: Figs. 2, P. thalestris; Figs. 3, P. avellaneda. Photos (a) were taken using
an 18a Wratten filter with a peak at 360nm. (Range: 310-400nm). Photos (b)
were taken using a 50 Wratten filter with a peak at 450nm. (Range: 430-480nm).
3a) shows U-V reflectance with a peak at 270 nm, compared to the 350 nm peak
(see below) of the forewing patch. The scent patches of the hindwings are not
visible in Figures la and 2a because they are covered by the forewing. Scent
patches of all male Phoebis on the ventral forewing, as well as the dorsal hind-
wing, show U-V reflectance at approximately the same intensity, but were not
included in the present report.
The actual U-V reflectance for these wings may be noted in Fig. If, 2f, and
3f. Graphs were made from samples from the forewing discal cell, and oriented in
the same direction in a grating spectrophotometer so that the readings might be
more comparable. All three species were similar in having the peak intensity of
U-V at 350 nm. Minor peaks below this wavelength are shown in all species with
thalestris showing an unusually high intensity peak at 310 nm.
The graphs in Figures If, 2f, and 3f also clearly reflect the light frequency
output obtained from the discal spots of the three species in the visible parts of the
spectrum. The heavy yellow component in philea (Fig. If) is apparent, as is the
shift toward the red end of the spectrum in avellaneda (Fig. 3f).
Dorsal surface of male Phoebis photographed using tight-band filters. Fig.
1, P. philea; Figs. 2, P. thalestris; Figs. 3, P. avellaneda. Photos (c) were taken using
a 75 Wratten filter with a peak at 490nm. (Range: 450-540nm). Photos (d) were
taken using a 73 Wratten filter with a peak at 576nm (Range: 550-600nm).
A comparison of the photographs taken with the various tight-band filters
establishes an unmistakable pierid pattern, particularly in marginal and sub-
marginal areas. One can note in the "a" photos for example, that each of the three
species has a minimum of four similar reflectance areas in the near ultra-violet
part of the spectrum: 1) the high intensity discal area, referred to above; 2) less
intense, but reflective scent patches (some not visible); 3) a marginal to submar-
ginal "glossy" area, which shows a very low intensity reflectance comparable
perhaps to that shown on many wings veins, and 4) dark areas which are U-V
absorbing. Clearly one could select scale samples from the same regions from each
of these species, and expect that they might be similar in those ultrastructural
characters which may account for this reflectance. By studying each of the "tight-
band" photos, careful selection of sample areas would maximize the chances of
relating color to scale structure.
Figure 4 shows the wing areas from which scale samples were taken for ultra-
structural comparison on the scan electron microscope (SEM). It may be noted
that the majority of reflectance patterns observed in the photographs are included
wat. Iegth, in
Dorsal surface of male Phoebis photographed (e) using a 72b Wratten filter
with.a peak at 605nm. (Range: 590-650nm). Fig. 1, P. philea; Fig. 2, P. thalestris:
Fig. 3, P. avellaneda. Reflectance curves (f) of the forewing discal cells. Figures
(f) are spectrophotometric records for the three species.
Figure 4, Wing areas from which scale samples were taken for comparative
in the samples.
The U-V reflectance photographs of males of a number of Phoebis species
and near relatives including the allied genera (considered subgenera by some
authors) Aphrissa, Rhabdodryas and Prestonia are given in Figs. 5a, 5b and
5c. For a more accurate wavelength comparison, spectrophotometric plots of the
refectivity from the same region in the discal cell of each species is presented by
each graph. The absence of U-V reflectance in Phoebis sennae marcellina and
Aphrissa boisduvali males may be noted on both photographs and charts. This
fact, together with the occurrence of high intensity U-V reflection from females
of some species, i. e. philea, means that this type of sexual dimorphism is not uni-
form throughout the genus.
250 350 450 550 650
wavelength, nm wavelength, nn
Figure 5a: Ultraviolet reflectance patterns of the upperside of representative
males of the genus Phoebis and accompanying spectrophotometric records of
samples from comparable forewing areas.
The spectrophotometric measurements were made using a constant incidence
of reflection (100) so that structural color shifts as predicted and reported by
Huxley (1975) would be avoided. The 100 incidence of reflection should provide a
minimum color shift.
Structural color produced by scale ridges, for example, should exhibit a
variable intensity according to the angle of placement of the ridge with respect to
the plane of the light path. An examination of the spectrum of the blue iridescent
scales of Morpho menelaus exhibited this phenomenon in both the visible and
ultra-violet portions of the spectrum. The maximum reflectance occurred when the
ridge-line of the scales was in the same plane as the light path, and lower reflectance
was obtained when the ridge-line was at right angles to the light path.
250 350 450 550 650 250 350 450 550 650
wavelength, nm wavelength, nm
Figure 5b: Ultraviolet reflectance patterns of Phoebis agarithe, Prestonia
clarki, and the related genus Aphrissa. The accompanying graphs are the spec-
trophotometric records of samples from comparable forewing discal areas.
As noted above, we were cognizant of structural colors produced in the longi-
tudinal ridges of Pieridae, and we expected this in Phoebis. However, there was no
observable difference in intensity in the visible portion of the spectrum when the
wing was placed in various positions; only in the U-V part of the spectrum did
we obtain significantly different readings depending on the angle of the wing
to the plane of the light source. It seemed reasonable to conclude, therefore, that
the visible colors (yellow, orange, and red) stemmed from locations in the scale below
the ridges, or perhaps in the ridge sinuses, and that the ridges of these scales pro-
duced the U-V reflectance. The ultrastructure of ridges will be discussed below.
However, of some interest is the fact that there was no significant difference in
readings obtained between 90" and 270 and between 0 and 1800 to the plane of
the light source. This would indicate that the slant of the ridge shelves has no
apparent effect on the intensity of output, at the 100 incidence of reflection in
Details of the ultrastructure of scales were studied from 12 comparable wing
areas of the dorsal surface of male Phoebis philea. P. thalestris and P. avellaneda.
In addition, samples were taken from the undersurface, from the wings of females
and from specimens obtained from many different geographical areas, to insure
a broader estimate of parameters of variability.
A minimum of two categories of scales occur in all wing samples observed:
these were grouped by the horizontal layering of the supine scales into covering
scales and basal scales. In proximal regions of the wing (extending almost to
submarginal areas) the sockets of covering and basal scales are arranged in trans-
verse rows; in the marginal area adjacent socket rows lie side by side forming an
irregular, semi-tiered group of sockets. Each row is composed of alternate sockets
of cover and basal scales; however, the blade of the cover scale is so wide that
Figure 5c: Ultraviolet reflectance pattern of Rhabdodryas trite and accompany-
ing spectrophotometric record of a sample from the forewing discal area.
their lateral margins touch the margins of other cover scales in the same line, and
overlying the intervening basal scale. The latter are often shorter than the former,
so that in most basal and discal areas of Phoehis wings the surfaces of only cover
scales are visible when viewed dorsally. Sockets of specialized scales such as
basal "hair" scales and marginal (border) black scales may be included along and
between regular tiers.
Sugmarginal and marginal areas in Phoebis have less distinct layering,
and scales of intermediate position (the Mittelschuppen of Kuehn and Henke,
1929) are the norm; often they are indistinguishable in type from cover scales
above or basal scales below, but admixtures of several scale types are not uncom-
mon in this region.
100 Yw i
Figures 6-9: Obverse surfaces of wing scales of male Phoebis. 6, sinuous and
elevated ridges of U-V reflective scales from discal cell showing shelves on lateral
ridge surfaces (6500x). 7, U-V reflective scale showing ovoid bodies beneath cross-
ribs. (6500x). 8, lateral view of U-V reflective ridges (19600x). 9, basal scale from
U-V reflective discal cell (6500x).
Pigment scales (Figs. 9, 10). Yellow-colored scales, and variations (light
yellow, lemon-yellow and yellow-range) occupy the greater part of the wing surface
and give each species its characteristic color. These scales show considerable
variation in their length and width, and more markedly, in characters of their
apices, whereas the apices of adjacent yellow scales may not alter abruptly, every
individual examined has transitions from an obtuse, gently rounded apex, to
the serrate condition with a varying number of incisions and denticles. Other
studies (see Downey and Allyn, 1975) indicate that environmental factors may
effect scale shape and the nature of the apex. However, in spite of this superficial
difference, these scales do not show fundamental differences in their ultra-
structure as observed with the SEM. We have not as yet been able to distinguish
significant differences in the pigment bodies of these scales (see below).
i' i; j
Figures 10-13: Obverse surfaces of wing scales of male Phoebis. 10, cover
and basal scales of U-V non-reflective areas. (SOn6500x 11, scale forms in lightly
U-V reflective areas. (200x). 12, ultrastructure of elongate scale (Fig. Ila) (6500x).
13, ultrastructure of spatulate scale (Fig. lib) (6500x).
1;3, ultrastructure of spatulate scale (Figs. 11b) (6,500x).
Pigment scales, with cross ribs showing a ladder-like arrangement (the
.ritert.'\pus scale of Suffert, 1924) are shown in Figure 9. They formed the cover
and the basal scales (even though different in outline) from sample areas 1, 4, 7,
8, 9, 10 and 11, in philea, thalestris and avellaneda. Some sample areas in the
three species had different scale types; we had anticipated greater differences
than observed. Area 3 dilfered in all 3 species, as suggested in Figures la, 2a,
and 3a, and in philea the cover scales in this area were of the ladder-like, pigmented
type. Area 6 was the only other sample site with major differences in scale types
for the three species above; the thalestris sample 6 underscale was of the ladder-
Figures 14 17. Obverse surfaces of wing scales of male Phoebis. 14, yellow
or clear edge scale (6500x). 15, black edge scale (6500x). 16, basal area of typical
non U V reflective scale showing various window forms (6500.\ 17, typical scale
with the ridge surface removed showing trabecular fragments and displaced
ovoid bodies on the inner surface of the ventral side of the scale (6500x).
Pigment scales were the principal scale types and formed the cover scales of
the male underside. The cover scales in females of the three species, even in the
area of the discal cell corresponding to the male U-V reflectant spot, were of this
The most prominent morphological feature of U-V reflection scales (Figs. 6,
7, 8) is the great height of the elevated longitudinal ridges; in Phoebis philea
they average 1.8 microns elevation above the surface of the obverse membrane,
compared to a height of 0.7 microns for the longitudinal ridges of the non-reflec-
tant, ladder-like scales. The ridges of the reflectant scales are also closer together
(a greater number per unit width) averaging 0.85 microns apart as measured
crest to crest, compared with 2.13 microns between the crests of adjacent ridges
in non-reflectant scales. Put in other terms, there are more than twice as many
ridges per unit scale width in highly U-V reflectant Phoebis cover scales as occur
in non-reflectant, or basal scales. The ridges in the former are twice as high as
the width of their inter-ridge channels.
Ridge shelves are very obvious on the lateral margins of the longitudinal
ridges (Fig. 6) where they form conspicuous, parallel ledges. Each shelf is set at
a slight upward angle (1-30) to the plane of the scale, and if traced distally, termi-
nate in a scute or scaly process on the crest of the ridge. The scutes are imbricate
and their apices project from under the apices of proximal scutes. Scutes are also
prominent features of other scales (see Figs. 10, 12, 13), but there are not as many
shelves beneath them.
These shelves and the air spaces between them constitute a thin-film, inter-
ference-reflectance system. The precision with which the superimposed surfaces
must be arranged in order to reflect certain wavelengths of light and the refractive
index of the material in the shelves, determines the color reflected and its intensity.
Ghiradella et al (1972: 1216) state that the average thickness of the lamellae in
Eurema lisa is 550 t 35 Angstroms, and the air spaces between shelves is 826
42 Angstroms. From the known refractive index of air (1.0) and the inferred approx-
imate refractive index of shelf cuticle (1.60, see discussion in chitin below), Ghira-
della et al (loc. cit.) assumed that the shelves and the inter-shelf grooves had similar
optical thicknesses (shelves: 880 56 Angstroms, grooves 826 42 Angstroms).
Considering preparation distortions and measurement errors with the electron
microscope, these authors were convinced that any disparity between the optical
thicknesses of the two regions would be insignificant, and the common thickness
of 858 + 51 Angstroms assumed. They thus claimed that in Eurema lisa. the
superimposed layers would function in the manner of a "quarter-wavelength
interference reflection filter", which should reflect maximally at 343 nm. They
confirmed this with reflection spectroscopy where the maximum ultraviolet
reflection (corrected for tilt of the ridges on the longitudinal ridges) was 348
+ 2 nm.
During the present study we duplicated work on Eurema, and while our mea-
surements on the scales were similar to those of Ghiradella et al (loc. cit.), sufficient
significant differences were encountered to warrant further study. We anticipate
more detailed investigation and separate publication on this question at an
Since the ultraviolet reflectance of Phoebis and Eurema scales is produced by
optical interference, the refractive index of the shelves of these ridges may vary
depending on the presence or absence of chitin; the intervening intershelf grooves
would also have to be correspondingly larger or smaller (depending on the refrac-
tive index of the shelf) in order for reflected light waves to reinforce rather than
cancel one another.
This suggests that the inclusion of chitin or other chemicals in the plasticized
shelf system needs also to be related to the structural arrangement in order to
produce the consistency of color reflection/absorption usually noted in wing
scales. In tests by Richards (1947) for chitin in butterfly wings, the alkaline treat-
ment usually destroyed the iridescent qualities of scales. The iridescent scales of
Doxocopa (= Chlorippe) seraphina were an exception to this treatment, however,
retaining their iridescence; this indicates a basic difference in the cause of the
iridescence in this species, and perhaps others.
We were also able to observe intra-shelf sinuses inside the longitudinal ridges
of Phoehis. There appeared to be seven or eight such cavities in each ridge. Since
they follow the downward angle of the shelves, we assume they communicate
with the air space in the lumen of the scale, near the point of junction of the shelf-
line with the obverse surface. The scute on the crest of the ridge lacks the sinus
towards its apex.
The occurrence of ridge sinuses indicates that the optical properties of the
ridge system corresponds to those reported by Ghiradella et al (1972) in Eurema,
and by Ghiradella (1974) in Pieris. It would appear that the gradually enlarging
sinuses (moving dorso-ventrally in a transverse section, or from the distal apex
of the sinus near the scute to its proximal union with the lumen) maintain a fairly
uniform light refracting or reflecting qualities at any point on the ridge.
Beneath the lowest two or three shelves is the lumen of the scale, which pro-
jects upward, forming an inverted "V"-shaped cavity beneath each ridge.
The U-V reflecting scales in Phoebis always contain ovoid bodies (=pterino-
somes) which can be noted between the ridges in Fig. 7. They are attached to cross
ribs and the trabecular braces beneath ridges. They may contribute to the visible
colors noted in these scales as will be discussed below.
U-V reflectant scales with high longitudinal ridges were the cover scales in
sample area 2. In addition, sample area 1 of philea and area 3 of avellaneda had
cover scales of this type, as was predictable from the U-V photographs (Fig. la,
and 3a). In all cases, basal scales in these areas were small (hidden), ladder-like,
pigment bearing, yellow-type scales. The correlation of scale type with U-V pro-
perties was further demonstrated by SEM observations in other Phoebis species.
P. argante (Fig. 5a) had high-ridged scales in sample area 4, and proximal 5, as
well as area 1 and 2.
Directional Reflectance. Various authors have called attention to the "on-
off" directional reflectance of the ultraviolet in some pierids; oblique illumina-
tion from one side of the specimen causes only the scale-patch on the opposite
side to "light-up". The wings or patch on the same side as the source of illumination
absorbs, rather than reflects in the U-V wavelengths. These authors include:
Nekrutenko (9I;bih in Gonepteryx rhamni where the phenomenon was described
as a gynandromorphicc" effect and a "dip angle"; Eisner et al (1969) photo-
graphed in Phoebis rurina and indicated for other genera and described as "direc-
tional iridescence"; Ghiradella et al (1972) in Eurema lisa and Colias eurytheme;
and Ferris (1973) in Colias alexandra.
Of the species reported herein, only Phoebis argante and A. statira exhibited
this phenomena. However, in corroboration of earlier reports, we have found
numerous other species in other genera show this "blinking". Our preliminary
studies tend to confirm the shelf-system angle produces this effect as reported by
Ghiradella et at (1972) in Eurema Lisa. Our findings with respect to directional
reflectance will be reported in detail in a following paper.
Not all male pierids exhibit U-V reflectant properties. The fact that Aphrissa
hoisduvali (C. Feld.), Colias philodice Godt. and Phoebis sennae (Linn.) lack this
character, while their near relatives show scale reflectance, might permit some
evolutionary inferences. Cover scales appear to be absent in all these species,
and the remaining basal scales lack the high reflectant longitudinal ridges. The
ultraviolet reflectance in females of Colias chrysotheme reported and shown in
photographs by Silberglied and Taylor (1973) is very similar to the "mirror" re-
tlL'lIc obtained from wing membranes devoid of scales, or those whose cover
layer of scales has been removed by rubbing.
Ultraviolet-reflectance is inherited as a sex-linked recessive trait in Colias
ISilbrglied and Taylor, 1973: 408). This was determined from examination of
preserved specimens of Gerould's genetic broods at Yale University, and by experi-
mental crosses by Taylor (op. cit.). Ultraviolet absorbance is sex-linked dominant.
The mode of inheritance of this trait in Phoebis is unknown.
Perforated-type scales (Figs. 12, 13). This category of scale, which was dis-
tinguished by Suffert(1924) (=Lochreihentypus), has numerous "pores" or windows-
on-the-septum between the longitudinal ridges. An equally distinctive feature is
that of the transverse flutes. These are a continuation of the vertical or oblique
flutes or foldings on the lateral margins of the longitudinal ridges. It is presumed
these folds impart more strength to the surface from which they originate, but
their linear, parallel nature makes them obvious features, and imparts a pinnulate
or feather-like appearance to such a scale. This is particularly apparent in one
grouping of this type of scale (Fig. 13) in which the windows are reduced to small,
pore-like slits, appearing in irregular positions on the septum. Such a scale has
much more of a reflective obverse surface than the scale whose windows are large
and rectangular, where ovoid bodies in the scale lumen can be readily observed.
It is assumed that this "feather-like" scale type accounts for the minor mirror
type U-V reflectance in the marginal and submarginal areas as may be noted in
Fig. la, 2a and 3a.
The nature of the windows on this scale type enhances the disparate appear-
ance of extremes, and emphasizes scale differences rather than similarities. This
is demonstrated in comparing Figure 12 with Figure 13. The former has fairly
uniform, obtuse to rounded windows, with a degree of regularity to the width of
the cross ribs between openings. In the latter, one might assume that windows
are lacking, and that only an occasional opening or pore-like slit breaks the con-
tinuity of the surface. Occasional scales can be found which lack these openings.
The ultrastructure of such different appearing scales is remarkably similar, and
seems to vary only with the size of the window. In fact, they represent stages along
a presumed continuum which ends with the fully "opened" windows as may be
noted in the pigment scales discussed above. For purposes of calling attention to
their possible increased mirror reflectance, we are here separating the fully-opened
window condition, from all those with smaller openings, and an increased number
of transverse flutes.
While the basic ultrastructure of this particular scale type is fairly uniform,
the pattern of the longitudinal ridges and the sculpturing on the obverse mem-
brane shows some variation from place to place on a single scale. That is, the basal
area may show a transition between several conditions of window types, as shown
in Figure 16. Nearer the pedicel, the obverse membrane may lack openings, but in
traversing an inter-ridge channel to more distal positions the windows are at
first small pores, with much transverse fluting between, to irregular medium-
sized, round opening exposing about half of the linear distance between ridges,
to the fully opened rectangular windows observed in pigmented cover scales (above).
Another feature shared by most of these scale types is the occurrence of ovoid
bodies in the scale lumen. They may be clear, yellow, orange, or black (Fig 15)
should there be melanin deposited in the lamellar or other surfaces (see below).
Wing samples from areas 5, 6 of the forewing, and 11, 12 of the hindwing,
contained these pinnulate scales. In these areas, pinulate scales with medium
sized, irregular-shaped windows (Fig. 12), tended to be in elongate scales with a
rather acute apex (Fig. 11a). The small-pored scale (Fig. 13) whose inter-ridge
septum appears composed largely of concentric transverse flutes (not infrequently
mistaken for cross ribs in light microscopy studies) represents the surface sculptur-
ing of shorter spatulate scales (Fig. 11b) with obtuse apices and attenuate basal
Black scales, transparent scales, and dentate yellow scales from area 5 resemble
one another very closely in surface ultrastructure. The black scales lack the great
quantity of ovoid bodies of the other two types, but the black scales do not differ
(at least superficially) in other morphological features. One way to account for
the color difference in the latter case is to assume the presence of melanin pigments
in both the longitudinal ridges and obverse membranes and cross ribs of the black
scales. This would mask any yellow color which might be caused by pigments
beneath. Visual proof of the occurrence of melanin in the ridges was obtained
during this study by color photography of the marginal black scales in both P.
philea and P. thalestris. At high magnification (200) the longitudinal ridges ap-
peared as black stripes against the lighter inter-ridge channel. Yellow scales of the
same size, shape and physical appearance except for color, occur in the same
marginal areas; presumably the only fundamental difference is in the internal
We cannot account for the transparent nature of some of these scales when they
contain numerous "so-called" pigment bodies. They do not appear to otherwise
differ from the yellow scales, and except for direct and careful observation, would
have been thus classified by SEM appearance only. Clearly, careful chemical
analysis is warranted.
Differing amounts and varying sites of melanin deposits within single
scales have been reported in other pierids (Hirata & Uehara, 1959; Hirata & Kubota,
1957; Yagi, 1954, 1955; Hidaka & Okada, 1970).
We have not been able to confirm in Phoebis the possible occurrence of a
second structural type of pigment body, or a type whose normal position is horizontal
in the lumen lying against the underside lamella, as noted by Hidaka and Okada
(1970) in Pieris. This included scale samples from the same relative position of the
hindwing as well as many other locations. The occasional ovoid body observed in a
horizontal position in the scale lumen was presumed to be accidentally dislodged
from its normal position attached to the obverse cross ribs, under-ridge, supportive
struts or trabeculae. This was confirmed in cases where scales were subjected to
harsher handling (chemical extraction and/or scale incision and fragmentation)
to obtain cross sectional views (see Fig. 17) where greater numbers of ovoid bodies
would be disturbed, and could be found in atypical positions.
Hopkins (1895) first proposed that the white and yellow pigments of the Pieridae
were compounds related to uric acid and were restricted to members of this family.
A host of other studies commencing with Wieland and Schopf (1925) and as recent
as Lafont (1975) and Descimon (1976) have detailed the chemistry and relationships
of the specific pigments involved. The complexity of the physiology of pigmentation
has made it much more difficult to understand the biological role of pterins in
wing pigments. However, it is generally acknowledged that they function in intra-
and inter-specific signals (communication) and have excretory utility.
While there are a limited number of wing pterins (12 or less) which have been
identified from various pierid wings, it has been assumed that most, if not all of
these have been confined to the wing scales, and specifically to the pterinosomes,
of "pigment" bodies. As indicated above, these bodies are round, elliptical or rod-
shaped structures located in the lumen (inter-trabecular sinus) of the scale. They
appear to be loosely attached to cross ribs as well as trabecular struts and braces,
and hang suspended in the lumen with the longitudinal axis (if any) pointed
downward as if by gravity. Other scale pigments, such as melanius and tryptophan
derivatives (ommochromes, papiliochromes) may be located in other structural
parts of the scale, which may hinder extraction by some methods.
Pronounced U-V reflectance patterns occur in males of most species in the genus
Phoebis. The brilliant U-V iridescence on the male forewing basal and discal areas
is produced structurally in the longitudinal ridges of the cover scales. This was
confirmed by spectroscopy and scan electron microscopy. Lesser intensities of
light in the 200-400 nm range, other than that produced in the longitudinal ridges,
is primarily caused by mirror reflectance. This is associated with the inter-ridge
septum of scales of the perforated type, which are primarily marginal and sub-
marginal in occurrence.
Not all species of Phoebis and relatives have males showing high intensity
U-V reflectance patches or patterns. In those males that show the trait, it is subject
to less variation than color types (visible and U-V) which may occur in the females.
We believe it is used as a sexual recognition signal.
A minimum of two types of scales occur in all wing samples taken: these may
be grouped by a horizontal arrangement of the supine scales into covering scales
and basal scales. In proximal regions of the wing (extending almost to submar-
ginal areas) the sockets of covering and basal scales are arranged in alternate
transverse rows which show various degrees of coalescence; in the marginal area,
adjacent socket rows lie side by side forming an irregular, semi-tiered group of
sockets. In some cases, the rows become superimposed, and the "single" row result-
ing will then be composed of alternate sockets of cover and basal scales. Sockets
of "specialized" scales such as basal "hair" scales and marginal (border) black
scales may be included along and between regular tiers as extras.
Submarginal and marginal areas in Phoebis have less distinct layering,
and scales of intermediate position (the Mittelschuppen of Kuehn and Henke,
1929) are the rule; often they are indistinguishable in type from cover scales above
or basal scales below, but admixtures of several scale types are not uncommon in
Several ultrastructural details were noted in wing scales of the genus Phoebis
for the first time:
1) Over twice as many longitudinal ridges per unit scale width occur on U-V
reflectant cover scales as occur on basal (or regular) pigment-bearing scales.
Ridges on the former are twice as high as the width of their inter-ridge chan-
nels; in basal or pigment-bearing scales in this genus, the longitudinal ridges
have a low profile, which seldom finds them over 1/3 as high as the width
between adjacent longitudinal ridges (measured crest to crest).
2) Ridge shelves on the lateral margins of the longitudinal ridges are in the
opposite (rather than laternate) position in Phoebis. While there is some vari-
ation in shelf number even in the same scale, some species show significant
differences (viz. philea shelf number 10-11; argante shelf number 6-7).
3) The absolute intensity of U-V reflectance in Phoebis males may vary slightly
between individuals of the same species, but more significantly, between
species. Some of the ultrastructural features observed which help account
for this include: height and sinuous nature of the longitudinal ridges; number
of longitudinal ridges per given width on the obverse surface; number of
shelves per ridge.
4) Wing-size of the specimen (giant, dwarf or average) did not appear to in-
fluence U-V reflectivity; this verifies the relationship between reflectivity
and ultrastructure established for this character.
5) Ridge sinuses are present, which together with ridge shelves in an apposite
position indicate that the optical properties of the ridge system corresponds
to those reported by Ghiradella et al (1972) in Eurema, and by Ghiradella
(1974) in Colias.
6) The lumen of the scale forms an inverted "V" shape at the base of each longi-
tudinal ridge, and extends vertically to occupy a position beneath the lowest
2 or 3 of the parallel ridge shelves. The intra-shelf sinuses occur between the
upper 7 or 8 shelves; the apical portions of the scute (or scaly process) on the
crest (summit) of the ridge usually lacks a sinus.
7) The U-V reflectant scales in Phoebis contain ovoid bodies, and show color
characteristics in the visible spectrum (yellow, orange, red) not dissimilar
from scales which absorb the U-V wavelengths.
8) Differences in pigment extractions of these scales and others suggests that
at least two chemical types of (yellow-orange) pigment are involved and/or
that they may be physically bound in a different manner within the scale,
and should be studied further. Bleaching or extraction of these pigments does
not change the U-V reflectance properties of the scale to a significant degree:
such scales appear colorless against both white and black backgrounds.
9) Scales with ovoid bodies may be colored, or be black, or transparent. Evi-
dence was present to show that melanins occur in the longitudinal ridges,
and elsewhere in the scale, and may mask other pigments within. The trans-
parency of some scales bearing ovoid bodies is as yet unexplained.
10) Scale samples from various areas of the same wing had fewer ultrastruc-
tural differences than expected. Scales from some wing areas, i.e.. discal
regions, had greater inter-specific diversity than did scale samples from other
comparable wing areas of different taxa.
11) Male U-V reflectance is reported for the first time in the following Phoebis
species or near relatives: Aphrissa statira (Cram.); P. agarithe (Bdv.); P.
avellaneda (H.-SchaeffT, P. neocypris (Hbn.); P. thalestris (Illiger); Prestonia
clarki Schaus; Rhabdodryas trite (Linn). The males of Aphrissa boisduvali
(C. Feld.) and Phoebis sennae (Linn.) lack marked U-V rcllectant properties,
which parallels the case in Colias where male C. philodice lack U-V reflecting
scales while most other Colias species possess such characters. The variable
occurrence and intensity of U-V reflecting scales in the female also attests to
the lack of uniformity of this type of sexual dimorphism throughout the
Our thanks to our friend and colleague, Dr. Lee ). Miller, for taxonomic
opinions and comments as well as a manuscript review. We are also grateful
to Dr. C. G. McCollum and Mr. A. C. Haman for review and criticism of the manu-
script. The assistance of Dr. Tyler Estler in advising us concerning the physics
aspects of this study was most valuable.
Brown, F. Martin. 1929. A revision of the genus Phoebis (Lepidoptera). Amer.
Mus. Novit., 368: 1-22.
---. 1931. A revision of the genus Aphrissa. Amer. Mus. Novit., 454: 1-14.
Cockayne, E. A. 1924. The distribution of flourescent pigments in Lepidoptera.
Trans. Ent. Soc. London, 1924 (1-11): 1-19.
Descimon, Henri. 1965. Ultrastructure et pigmentation des ecailles des Lkpidopteres.
Jour. Microscopie, 4: 130.
--.-. 1966a. Action du 5-fluoro uracile sur la genBse des ecailles de Lepidopteres
Jour. Microscopie, 5:41.
---. 1966b. Developpement des ecailles au course de la metamorphose chez
Colias croceus F. (Lepidoptera Pieridae). Jour. Microscopie, 5: 41-42.
-- 1969. Contribution a l'6tude de la pigmentation pterinique des Pieridae.
These de octorate d'6tatd es sciences naturelles, Paris.
--. 1971. Les pterines des Pieridae et leur biosynth6se I. identification des
principles pterines de Colias croceus iFourcroy et de quelques. autres espkces
de Pieridae Biochimie, 53: 407-418.
-- 1976. Biology of pigmentation in Pieridae butterflies. In de Gruyter,
Walter. Chemistry and Biology of Pteridines. Pp. 805-840.
Downey, J. C. and A. C. Allyn. 1975. Wing-scale morphology and nomenclature.
Bull. Allyn Mus., No. 31: 1-32.
Eisner, T., R. E. Silberglied, D. Aneshansley, J. E. Carrel and H. C. Howland.
1969. Ultraviolet video-viewing: The television camera as an insect eye. Science,
Ferris, C. D. 1972. Ultraviolet photography as an adjunct to taxonomy. Jour.
Lepid. Soc., 26: 210-215.
1973. A revision of the Colias alexandra complex (Pieridae) aided by
ultraviolet reflectance photography with designation of a new subspecies.
Jour. Lepid. Soc., 27 (1): 57-73.
Ghiradella, Helen. 1974. Development of ultraviolet-reflecting butterfly scales:
how to make an interference filter. Jour. Morph., 142 (4): 395-409.
Ghiradella, H., D. Aneshansley, T. Eisner, R. E. Silberglied and H. E. Hinton.
1972. Ultraviolet reflection of a male butterfly: interference color caused by
thin-layer elaboration of wing scales. Science, 178 (4066): 1214-1217.
Hidaka, T. and M. Okada. 1970. Sexual difference in wing scales of the white
cabbage butterfly, Pieris rapae crucivora. as observed under a scanning electron
microscope. Zool. Mag., 79: 181-184.
Hirata, K. and T. Kubota. 1957. Microstructure of the scale of the butterfly,
Colias rate poliographus Motschulsky, elucidated by electron microscopy.
Sci. Rep. Kagoshima Univ., 6: 151-167.
Hirata, K. and N. Ohsako. 1960. Studies on the structure of scales and hairs of
insects. III Micro-structure of hairs of the butterfly, Colias rate poliographus
Motschulsky. Sci. Rep. Kagoshima Univ., 9: 123-135.
Hirata, K. and J. Uehara. 1959. Studies on the structure of scales and hairs of
insects. II Microstructure of scales of the blank female of the butterfly, Colias
rate poliographus Motschulsky. Sci. Rep. Kagoshima Univ., 8: 155-174.
Hopkins, F. G. 1895. The pigments of the Pieridae: A contribution to the study of
excretory substances which function in ornament. Phil. Tran. Roy. Soc. London,
B. 186: 661-682.
Huxley, John. 1975. The basis of structural color variation in two species of
Papilio. Jour. Ent. (A), 50: 9-22.
Koehler, Wilhelm. 1941. Experimentelle Untersuchungen uber die Determination
des Zeichnungsmusters bei der Mehlmotte Ephestia Kuhniella Zeller. Vjschr.
Naturf. Ges. Zurich, 86: 77-151.
Kuehn, A. and K. Henke. 1929-36. Genetische und entwicklungsphysiologishe
Untersuchungen an der Mehlmotte Ephestia Kuehniella Zeller. Abhandle. Ges.
Wiss., Goettingen, 15: 1-272.
Lafont, R. 1975. Aspects biochimiques de la differenciation des disques imaginaux
alaires des LIpidopthres. Ecole Normale Superieure Publications du Labora-
toire de Zoologie. No. 5. Pp. iv + 157. Paris.
Mazokhin-Porshnyakov, G. A. 1954. U-V rays from the sun as a factor in insect
habitats. Zh. Obshch. Biol 15: 362-367.
1957. Reflecting properties of butterfly wings and role of ultraviolet rays
in the vision of insects. (translation). Biophysics, 2: 352-362.
---- 1969. Insect vision. Pp. xiv + 306. Plenum Press, New York.
Mottram, J. C. and E. A. Cockayne. 1920. Fluorescence in Lepidoptera. Proc.
Ent. Soc. London, 1920: 36-39.
Nekrutenko, Yuri P. 1965a. The hidden wing-pattern of some palearctic species
of Gonepteryx and its taxonomic value. Jour. Res. Lepid., 1964, 3 (2): 65-68.
---. 1965b. "Gynandromorphic effect" and the optical nature of hidden wing-
pattern in Gonepteryx rhamni L. (Lepidoptera, Pieridae). Nature 205: 417-418.
-----. 1966. Three cases of gynandromorphism in Gonepteryx. Jour. Res. Lepid.,
1965, 4 (2): 103-108.
--.. 1970a. Comments on forms of Gonepteryx aspasia (Pieridae) described by
Shu-iti Murayama. Jour. Lepid. Soc., 24 (3): 213-217.
------. 1970b. A new subspecies of Gonepteryx rhamni from Tian-shan Mountains,
U.S.S.R. Jour. Lepid. Soc., 24 (3): 218-220.
.-----. 1973. A new subspecies of Gonepteryx amintha (Pieridae) from Yunnan,
Mainland China, with comparative notes. Jour. Res. Lepid., 1972, 11 (4):
Obara, Y. and T. Hidaka. 1968. Recognition of the female by the male, on the basis
of ultraviolet reflection, in the white cabbage butterfly, Pieris rapae crucivora
Bdv. Proc. Japan Acad., 44: 829-832.
Richards, A. Glenn. 1947. Studies on arthropod cuticle. I. The distribution of chitin
in lepidopterous scales, and its bearing on the interpretation of arthropod
cuticle. Ann. Ent. Soc. Amer., 40: 227-240.
Scott, James A. 1973. Survey of ultraviolet reflectance of nearctic butterflies.
Jour. Res. Lepid., 12 (3): 151-160.
Sellier, Robert. 1971. Donnees sur les apports de la microscope electronique
a balayage, pour 1' etude ultrastructurale des 6cailles alaires chez les Ikpidoptires
diurnes. Comp. Rend. Acad. Sci. Paris, Ser. D, 273: 2097-2100.
Silberglied, R. E. and 0. R. Taylor. 1973. Ultraviolet differences between the sulfur
butterflies Colias eurytheme and C. philodice. and a possible isolating mech-
anism. Nature, 241: 406-408.
Suffert, Fritz. 1924. Morphologie und Optik der Schmetterlingsschuppen
insbesondere die Schillerfarben der Schmetterlinge. Zeitsch. Morph. Okol.
Tiere, 1 (2): 171-308. taf. v-ix.
Wieland, H. and C. Schopf. 1925. Ueber den gelben Fluegelfarbstoff des Citronen-
falters (Gonepteryx rhami). Ber. Dt. Chem. Ges., 58: 2178-2183.
Yagi, Nobumassa. 1954. Note of electron microscopic research on pterin pigments
in the scales of pierid butterflies. Annot. Zool. Japonenses, 27 (3): 113-114.
.------- 1955. Electron microscopical studies on the form of pterin pigments in
the scale of Pieridae with reference to their bearing on avs'tematic.i Jour.
New Entomologist, 4 (1-2): 1-20.