Characterization of the ocelli of the migratory Bogong (Agrotis infusa) in comparison to the non-migratory Turnip moth (Agrotis segetum) Denis Vaduva Abstract The Bogong and Turnip are two of the genus Agrotis, similar in size, but different in life style. The Bogong moth is a long distance migratory species endemic to , whereas the Turnip moth has a wide spread in Africa, Europe and Asia where it is considered a pest. The ocelli are single-lens eyes of the camera , whose role is still unknown in many , including the studied species. The present study searched for a potential role of these visual organs in these species, by determining the physiological and optical characteristics of the ocelli. Because of their different life styles (migratory vs. non-migratory), we searched to see if any differences are found and if the roles of the ocelli may differ as well. In my thesis I have found that these species have two underfocused ocelli placed laterally on the vertex of the head, close to the dorsal margin of the compound eyes and posterior to the antennae. The lenses are smooth with no pronounced asymmetry and astigmatism. The projection fields from the Bogong ocellar neurons are located close to the posterior brain surface, adjacent to the esophageal hole, anterior to the mushroom body calyx and anterior of the central complex. Because of the optical and physiological characteristics of the ocelli, it is probable that they play a role in flight stabilization reflexes although a function as a regulator of the initiation and cessation of diurnal activities, cannot be excluded. migration occurs (Britton; Warrant et al., Introduction 2016). The Bogong (Agrotis infusa) and Turnip Agrotis segetum is slightly smaller with a (Agrotis segetum) moths have a similar body length of 18-22 mm and wing span size and appearance and belong to the of 34-45 mm. Although a common genus Agrotis. With a body length between European species, the Turnip moth has a 20 and 25 mm and wing span that is wide distribution in Europe, Asia and approximately 40-50 mm (Britton; Africa (The Cooperative Research Centre CSIRO; Warrant et al., 2016), the Bogong for National Plant Biosecurity) where it is moth is found in Australia and considered a polyphagous pest (Chumakov occasionally in and New and Kuznetsova, 2008). Zealand (Warrant et al., 2016). A nocturnal navigator, Agrotis infusa makes General morphology of the ocelli a migration of up to 1000 km from in insects their breeding sites in southern , western and northwestern As many insects, the Bogong and Turnip (NSW) and western moths have a two-pathway visual system: to their cool estivation caves in the compound eyes and the ocelli (Berry et the alpine regions of NSW and Victoria al., 2011). The ocelli are single-lens eyes where they “hibernate” over the of the camera type found in nymphal and months. In early , a reverse adult hemimetabolous insects and adult holometabolous insects (Dickens and Eaton, 1974; Dickens and Eaton, 1973;

1

Warrant et al., 2006). The ocellus has a All this led to the creation of the “single simple structure composed of a pigmented sensor model” (Stange et al., 2002) in cone-shaped extension of the cuticle, a which, each ocellus is a highly sensitive dome-like corneal lens and a retina optical sensor optimized for detecting separated from the lens by a layer of illumination levels, providing rapid corneagen cells (Caldwell et al., 2007; response through the ocellar large second Dickens and Eaton, 1974; Dow and Eaton, order neurons (L-neurons). Thus, together, 1976; Horie et al., 2008; Toh and the ocelli act as an autopilot system with Okamura, 2007). which the pitch, yaw and roll flight components are stabilized (Berry et al., Typical for visual sensors in general, the 2011). This is furthermore improved by number of ocelli varies between species the fact that many ocelli have two spectral (Vrsansky, 2008). Most commonly, such sensitivity peaks in the green and UV as in bees (Berry et al., 2011), wasps regions (Eaton, 1976; van Kleef et al., (Warrant et al., 2006), dragonflies (Berry 2005; Yamazaki and Yamashita, 1991) and et al., 2007a) and flies (Caldwell et al., within the UV sensitivity range, the world 2007; Chen and Hua, 2014), the ocelli is visible as only a bright sky and a dark occur as a triplet and are generally located ground (Wilson, 1978). on the apex of the head between the compound eyes (Berry et al., 2007) The function of the ocelli in flies whereas in locusts (Wilson, 1978) the median ocellus of the locust is directed The structure and the neural organization forwards (Stange, 1981). Other species of the ocellar system varies so much have only one ocellus, whereas species among insects (Mizunami, 1994; such as beetles lack them completely Mizunami, 1995) that a role as a single (Kalmus, 1946), and in some moth species sensor is likely not the only role the ocelli within the families Sphingidae and may fulfill. In insects such as the fly Saturniidae, internal ocelli have been Drosophila, the ocelli contribute to discovered (Dickens and Eaton, 1974; positive phototactive orientation mediated Dickens and Eaton, 1973; Dow and Eaton, by the compound eyes, by adjusting their 1976; Eaton, 1971) which may monitor sensitivity (Hu and Stark, 1980; light spread inside the head capsule Mizunami, 1994). The ocelli also have a (Eaton, 1971; Pappas and Eaton, 1977). guiding role in houseflies, which are able Several moth species, have two ocelli to walk toward edges and relatively small positioned laterally on the vertex of the bright objects placed in the frontal head, close to the dorsal margin of the equatorial part of the visual field with their compound eyes and posterior to the eyes completely occluded (Wehrhahn, antennae (Grunewald and Wunderer, 1996; 1984). Pappas and Eaton, 1977). The role of the ocelli as polarized Proposed functions of the ocelli light detectors Despite the intense research and their Many species such as honeybees, apparent simplicity, a singular explanation desert ants, monarch butterflies, dung for the function of the ocelli has remained beetles and many others are able to orient elusive (Berry et al., 2007b; Warrant et al., by using specialized ommatidia in the 2006; Wilson, 1978). Generally, the ocelli dorsal rim areas of the compound eyes to have broad visual fields that partially detect the celestial patterns of light overlap (Wilson, 1978; Taylor et al., 2016) polarization (Wehner and Strasser, 1985; and the images produced are underfocused, el Jundi et al., 2014). The western as the image plane appears to lie well bumblebees, Bombus occidentalis, are able behind the retinal layer (Mizunami, 1994). to use their ocelli alone or in conjunction 2 with the tops of its compound eyes to emergence, but becomes less effective detect celestial polarized light and use it to with time. However, ocellar ablation prolong its foraging period at twilight produces a permanent flight initiation (Wellingt. WG, 1974). delay (Eaton et al., 1983). The ocelli also regulate the intensity of the flight activity The desert ant, Cataglyphis bicolor, is also as anocellate males flew significantly less able to do this. By using only their ocelli than control and sham-operated males. and the patterns of polarized light, These males also had difficulties in individuals of this species are able to find adjusting to changes in the time of sunset their way back home (Fent and Wehner, as they performed worse than the males 1985). Similarly, the ocelli of the with intact ocelli (Sprint and Eaton, 1987). Australian desert ant Melophorus bagoti, The ocelli of the Arctiid moth, provide celestial compass information for Creatonotos transiens, also play a role in directional orientation, although, by diurnal activities. Thus, when the control themselves they are unable to mediate path of the mating system of this species was integration. It is still unknown whether the studied, the occlusion of the ocelli, directional information comes from produced a significant delay in the onset of polarized skylight, the sun’s position or the the luring activity. However, as this color gradient of the sky (Schwarz et al., activity is only retarded, the compound 2011a;Schwarz et al., 2011b). eyes may be involved too (Wunderer and The controlling role in diurnal Dekramer, 1989). activities of the ocelli in moths The aim of the present study In insects, light intensity levels often Despite the recent research, the role of the control the initiation and cessation of ocelli still remains unknown in many diurnal actives. This makes it species. The purpose of the present study advantageous to be able to detect minute is to determine the physiological and changes in illumination. Thus, the high optical characteristics of the ocelli in the sensitivity of the ocelli, higher than that of Bogong moth Agrotis infusa, and the the compound eyes, makes them perfectly Turnip moth Agrotis segetum. These were suited as circadian controllers of diurnal studied in order to find a potential role of activities (Mizunami, 1994). In bees, these visual organs in these species and if occlusion of the ocelli interferes with the any differences between them exist. These timing of the first and last foraging flights, two species were chosen because of their as the light intensity required increases close evolutionary relationship and with the number of non-functional ocelli different lifestyles. It is still unknown what (Lindauer and Schricker, 1963; Schricker, role is played by the ocelli during the 1965). migration of the bogong moths, if any. The The ocelli of the field crickets Teleogryllus Turnip moth, with its nonmigratory commodus, and house crickets, Acheta lifestyle, serves as a comparison species, domesticus, have a modulatory function, where, most probably, any differences are augmenting the sensitivity of the caused by differences in their lifestyles compound eyes to better perceive photic (migratory vs. nonmigratory). entrainment signals (Rence et al., 1988) Material and methods The ocelli perform an important function in the regulation of diurnal activities in the Studied specimens looper moths, Trichoplusia ni, as well. Ocellar occlusion delays the flight Turnip moth individuals were obtained initiation on the first day following from a laboratory culture maintained at the Department of Ecology in Lund, . 3

The culture was established 37 years ago Scanning electron microscopy from field-collected moths from southern (SEM) Sweden and Denmark. The individuals were kept in a L17:D7 photoperiod at The external morphology of the Bogong 22°C and 60% RH (relative humidity). ocelli was further studied with the help of scanning electron microscopy. Two series Bogong moth individuals were collected of SEM scanning occurred. First, four by Eric Warrant, Stanley Heinze, Kristina already dead individuals (two males and Brauburger, and Anna Stoeckl from an two females) were taken from the Bogong alpine cave in Mt. Kosciusko National moth cage. Out of these, the scales on the Park, close to South Ramshead in early head of one male and one female were January 2013 and 2014. These individuals manually removed in order to make the were kept in an incubator at Lund ocelli more visible. For the second session, University in simulated cave conditions: two freshly killed males and one already lights on: 8pm to 12am (L16:D8), with 6 dead female were used. The female was degrees at night and 10 degrees during the used previously in the stereomicroscopy day and RH ca 75%. study and was kept in the refrigerator at 4⁰C. All three individuals had the scales Characterization of external completely removed. morphology All the individuals were fixed in a solution Stereomicroscopy composed of 2% paraformaldehyde and 2.5% glutalaldehyde in 0.1 M sodium The external morphology of the ocelli was Cacodylate buffer. The samples were then studied using a Nikon SMZ18 dehydrated in a graded ethanol series and stereomicroscope, by taking multiple Z- critical point dried, prior to being sputter- stacks which were then processed in NIS- coated with gold. Elements software, in order to obtain images with a large depth of field. The Characterization of internal resulting images have all the different regions of the ocelli in focus. morphology Bogong and Turnip moth individuals were Paraffin embedding and mounted on a holder using electrical tape sectioning and wax. In order to provide better access Prior to paraffin embedding, cryo- to the ocelli, the majority of the scales on sectioning and plastic embedding, the head were manually removed. Both microscope slides were coated in a males and females were used for this kromalungelatine solution (1g gelatin, 0,1g procedure. kromalun CrK(SO4)2*12H2O, 200ml In order to anesthetize the individuals prior distilled H2O) and dried overnight in the to mounting, a male and a female Bogong oven at 60⁰C. moth were placed in the freezer for 20-30 A total of seven Bogong moth heads with minutes. These individuals were then kept the antennae, mouth pieces and scales further in the refrigerator for three days, removed, were fixated overnight in two over the weekend. After this period solutions: three males and one female in additional imaging sessions occurred. Two AFA II (alcohol-formaldehyde-acetic acid: additional males were anesthetized using 75 ml ethanol of 96% concentration, 20 ml carbon dioxide from a Soda Stream of concentrated formaldehyde 37%, 5 ml machine, prior to mounting and of glacial acetic acid) and two females and photographing. one male in 10% neutral buffered formaldehyde (37% formaldehyde diluted 4 ten times in phosphate-buffered saline – to differentiate the tissue, 96% ethanol 1-2 PBS). min, 100% ethanol 2x5 min) and xylene (2x5 min) finishing with the mounting of a The fixated samples were rinsed in coverslip with New Entellan. distilled water (3 x 10 minutes) followed by a dehydration in an ascending ethanol Cryo sectioning series (70% 2x15 min, 90% 2x15 min – AFI samples only, 96% 2x10 min and A total of three Bogong moth heads with 100% 2x10 min). The dehydrated samples the antennae, mouth pieces and scales were incubated in xylene (2 x 10 min) and removed, were fixated overnight in two xylene/ paraffin (1:1 30 min) in the oven, solutions: two males in AFA II (75% followed by 3 changes of infiltration with ethanol of 96% concentration, 20% of fresh paraffin in (I x 1 h, II x overnight, III concentrated formaldehyde 37%, 5% of x 1h ) in the oven at 60⁰C. These samples glacial acetic acid) and 1 female in 10% were in turn embedded in a paraffin block neutral buffered formaldehyde (37% inside a handmade paper form with a pre- formaldehyde diluted ten times in established orientation (both horizontal phosphate-buffered saline – PBS). and vertical planes of the ocelli were After the fixation was completed, the AFA sliced). The samples were left in room II individuals were rinsed in a descending temperature to harden. ethanol series (90% 3 x 10 min, 70% 2 x The excess paraffin was removed from the 10 min, 50% 2 x 10 min) finishing with a blocks containing the samples so that two rinsing in distilled water (2 x 10 min parallel sides were obtained. These, were minimum). The rinsed samples were sectioned using a rotation microtome in infiltrated overnight in the refrigerator in slices with a thickness of 5 μm. The result cryoprotectant (25% sucrose in PBS). The is a stripe of paraffin sections, which stick formaldehyde fixated individual was to each other. These in turn need to be rinsed in distilled water (at least 3x10min) “stretched out” as the samples are slightly and then infiltrated in cryoprotectant as compressed, by placing the sections in well. water drops on a kromalungelatine coated The samples were embedded in a drop of microscope slide placed on a warm plate. mounting medium (Neg-50) on an This water must be warm enough so that uncoated microscope slide placed on the the paraffin stripes stretch out, but not too freezing plate (-60⁰C) of the cryostat with warm so that the paraffin will melt. After a pre-established orientation (mouth pieces this is achieved, the excess water is toward the side and the ocelli upward). removed and the samples are left to dry The mounted sample had the bottom end overnight. The sectioned samples were cut so that a straight edge was created. The then deparaffinized by going through a sample was them removed from the slide series of xylene (2 x 5 min) and a and mounted vertically on the straight edge descending ethanol series (100% 2 x 3 in the same medium on a specimen holder. min, 95% 2 x 3 min, 70% 1 x 3 min, 50% This in turn was removed from the 1 x 3 min) finishing with a rinsing in freezing plate and left inside the cryostat distilled water (1 x 3-5 min). for 15 minutes to slightly warm up in order The deparaffinized samples were stained to reach the optimal sectioning for at least 1 minute in a toluidine blue temperature. The sections were manually solution (Toluidine 0.5g, Sodium Borate – placed on a kromalungelatine microscope Borax 1.91g, distilled water 200 ml). slide. The slide was in turn warmed up so Following the staining process, the that the sections will melt and thus avoid samples were rinsed in an ascending freeze-drying of the tissue. ethanol series (70% ethanol – quick rinse 5

Fixation, plastic embedding and the measured thickness of the visible part sectioning of the lens, an approximation of the total thickness of the lens was obtained. The For the plastic embedding, two Turnip and perpendicular diameters passing through two Bogong moths were used. The scales the center of the lens were measured on the mouthpieces, antennae as well as large images in which the ocellus was portions of the eyes were removed in order photographed directly from above. These to improve fixative penetration. The represent the width of the ocellus. samples were fixated overnight in the refrigerator in a modified Karnovsky The plastic sections provided solution (2.5% Glutaraldehyde + 2% measurements of the lens thickness and paraformaldehyde in 0,1 M sodium width, as well as the total length of the cacodylate buffer of pH ca 7,4). After the ocellus and the distance to the fixation was completed, the samples were photoreceptor layer. For the Turnip moth, rinsed in in 0.1 M sodium cacodylate the measurements come from different buffer and dehydrated in an ethanol series sections of the same ocellus, whereas for (70% etoh 2x10 min, 96% etoh 2x10 min, the Turnip moth, the sections came from 100% etoh 2x10 min). The embedding of two different ocelli. the samples occurred according to the The optical properties of the ocelli following series: acetone 2x15 min, aceton/epon 2:1 (2+1) 30 min, aceton/epon The hanging-drop method used by Warrant 1.2 (1+2) overnight, epon 6h and finished et al (2006) was employed to measure the in an embedding in new epon molds. The back focal distances (BFD) and focal molds were in turn polymerized in the lengths (f) of the ocellar lenses. The ocelli oven at 60⁰C for 48h. lens together with a small portion of the surrounding capsule were carefully The ocelli were sectioned perpendicular to dissected and placed in a petri dish of the height of the ocellus. These three water. By using an eyelash glued to a micrometer sections were placed in water toothpick, the lens was cleaned from any drops on a microscope slide coated with tissue and pigment. An o-ring was waxed kromalungelatine and stretched on a warm to a conventional microscope glass and its plate. These were stained in a upper surface was lightly greased with Richardssons Metylenblue for ten seconds, Vaseline. The cleaned ocellus was placed rinsed in distilled water and dried on the with its external side outwards in a tiny warm plate. The process was completed drop of water (refractive index = 1.34) on with the mounting of a coverslip with New the center of a microscope cover slip, Entellan. which in turn was placed upside down Image analysis onto the greased o-ring. Thus, an air-tight chamber was created. The microscope The obtained images from the slide was mounted on the stage of a stereomicroscopy, plastic sectioning and conventional light microscope (Leica) that hanging drop methods were performed had its condenser removed. Dark stripes of using the software Fiji version v.1.51a known size on translucent tracing paper of x64. were placed on the foot of the microscope, over the lamp aperture. These objects Lateral and perpendicular images of the focused by the ocellus were viewed with ocelli were obtained using the Nikon the 40x objective and photographed with a SMZ18 stereomicroscope. Thus, for the digital camera fitted to the microscope. lateral view images, the total height of the ocellus and the thickness of the lens above The following equation was used to the cuticula were measured. By doubling calculate the focal length f of each ocellus:

6

in order to expose the inside tissue. Neurobiotin (Vector Laboratories, Burlingame, UK) crystals were applied to the tip of a borosilicate micropipette, where the so is the distance between the striped object and the ocellus (127 mm), which was inserted into each exposed ocellar tissue. o is the spatial wavelength of the striped pattern (the distance between the center of The brain and ocellar optical nerves were one stripe and the center of the next: 4 dissected, cleaned of any remaining air mm) and i is the spatial wavelength of the sacks and fixed overnight at 4⁰C in a image of the striped pattern (mm). fixative containing 4% paraformaldehyde Multiple photographs were taken for each (PFA), 0.25% glutaraldehyde, and 2% dissected ocellus. Three measurements saturated picric acid (in 0.1 M phosphate were performed on each image in which buffer). the striped pattern was clearly in focus. The brains were washed 4 × 15 min in The distance from the back of the ocellar 0.1M PBS and then incubated with Cy3- lens to the plane of best focus (the optical conjugated streptavidin (1:1000; Jackson back focal distance, BFD) was measured ImmunoResearch, West Grove, PA, USA; by first focusing upon small particles of catalog number 016-160-084) for three debris attached to the back of the lens, days at 4 ◦ C. After incubation, brains were followed by focusing upwards until the rinsed 4 × 15 min in PBS + Triton-X (Tx) best image of the striped object was and 2 × 20 min in PBS, dehydrated in an obtained. The difference in micrometers ascending ethanol series (50, 70, 90, 95 between these two focus points was and 100%; 15 min each), treated with a 1:1 measured using a micrometer gauge mix of 100% ethanol and methyl salycilate attached to the microscope stage. The for 15 min, and eventually rinsed for at mean value was obtained from nine least 35 min in pure methyl salycitate. The consecutive measurements for each prepared brains were mounted in Permount ocellus, which in turn was corrected for the between two coverslips, using plastic refractive index of water, by multiplication spacers to prevent squeezing of the brains. by 1.34. Imaging Descriptive statistics of the obtained data as well as graphical illustrations were The whole-mount preparations were performed using IBM SPSS statistics v.22 imaged using a 564 nm Argon laser on a x64. confocal microscope (LSM 510 Meta, Zeiss, Jena, Germany) using a 25× Analysis of brain projections objective (LD LCI Plan-Apochromat from ocelli 25×/0.8 Imm Corr DIC; Zeiss). We scanned at a frame size of 1024 × 1024 Neurobiotin injections voxels with optical sections every 0.79 µm. The resulting voxel size was 0.4972 x During the experiment, five Turnip and 0.4972 x 0.7914 µm3 with a field of view four Bogong moths were prepared. Each covering 509 × 509 µm. individual was restrained by taping the thorax tightly to a plastic holder to prevent Image analysis and three- any movement. The head and thorax were dimensional reconstruction fixed with wax, and the majority of the scales on the head were manually The image stacks of the adjacent moth removed. The tip of each ocellus brains and the ocellar optical nerves containing the lens was carefully removed sections were aligned and merged in

7

Amira 5.3.3 software (FEI Visualization 3.54 µm; width 66.75 ± 12.97 µm; Figure Sciences Group, Mérignac Cedex, France; 1, Table 1). RRID: nif-0000–00262). Characterization of the The ocellar nerves and their projection fields, together with several other brain internal morphology of the structures were manually reconstructed ocelli using the tool Wrap in the Segmentation editor as well as the Amira plugin Paraffin, cryo and plastic sections Skeletonize (Beetz et al., 2015). A Three methods were employed to study the complete reconstructed brain model was internal morphology of the Bogong and provided by Stanley Heinze in order to Turnip moths. Both the paraffin embedded better illustrate the obtained results. and cryo-sectioned material suffered from severe fragmentation making it impossible Results to obtain any data. Characterization of external The plastic embedding method was morphology of the ocelli somewhat more successful, although even here a high degree of fragmentation was Scanning electron microscopy and present. Almost none of the connective stereomicroscopy tissue and photoreceptor layer were preserved in the samples. The ocelli lenses The Bogong and Turnip moths have two were partially fragmented as well (Figure ocelli placed laterally on the vertex of the 1). However, measurements of their head, close to the dorsal margin of the thickness and width were possible. compound eyes and posterior to the Because of the high degree of antennae. The SEM and stereomicroscopy fragmentations, all the measurements of parts of this study have shown that these these samples represent approximations. are cone shaped with a lens that is slightly oval. Generally, the size of the Bogong The image D4 is based on a toluidine blue moth ocelli are slightly smaller than those stained ocellus sectioned at a thickness of of the Turnip moth (Table 1). 0.8 micrometers. This was provided to us by Professor Willi Ribi from the In the SEM images, a partial collapse of Department of Neurobiology, Research the lens is present, which is present only in School of Biology, ANU, , the images B3 and B4 (Figure 1). These Australia. images come from the individuals that were anesthetized for 20 to 30 minutes in The measurements of the lens dimensions the freezer. This treatment, always caused are similar to those obtained from our own a partial collapse of the lens. This collapse ocelli. However, one difference has been would only partially affect the total height observed: the total height of the ocellus is of the ocellus, which was not measured on slightly smaller than those obtained from these images. The diameters of the lens are our own ocelli. the same, as they are measured from The measurements of the width (75.75 ± cuticle rim to cuticle rim as shown in 1.5 µm) of the lens and the height (138.5 ± image C1 (Fig. 1) In the SEM images in 15.16 µm) of the ocellus are similar in the Figure 1, originating from the same plastic sections to those obtained from the individual, the height of the ocellus and stereomicroscopy method. The lens width of the lens are between 9,7% and thickness is significantly larger in the 40% larger than those obtained from the stereomicroscopy method (92 µm vs. stereomicroscopy images (height 138.5 ±

8

Table 1. Size measurements of the ocelli and their lenses. A - D represents the Bogong moths and E - F the Turnip moths. The numbers and letters following, represent the image IDs from Figure 1. The measurements are in micrometers. The measured lens dimensions (thickness, width, diameters, height) are illustrated in Figure 1.

Species Lens Ocellus Distance to Image Thickness Width Diameter Diameter Height photoreceptor ID A B layer B1 92 141 B2 92 136 B3 59 53 B4 81 74 C1 89 98 C2 152 D1 75 75 145 D2 72 75 144 D3 70 78 149 D4 70 75 116 5 E1 78 122 E2 62 66 F1 89 97 135 F2 90 85 143

71.75 ± 2.36 µm; t test: t = -11.43, df = 4, μm) and I (248 ± 5.99 μm) having been the P < 0.001; Table 1). largest (Figure 3d). In the case of the Bogong moth, the plastic Because of these, the focal length data for embedded sample provided a lower all ocelli and pictures is not normally thickness for the lens compared to the distributed. The variation in the data of stereomicroscopy measurements, whereas each ocellus is smaller and normally for the Turnip moth, the opposite is true. distributed (Figure 3c, Table 2). The back Compared with each other, the Bogong focal distance data for all ocelli is not moth has a higher ocellus (t test: t = 0.16, normally distributed and neither is the data df = 7 P = 0.88; Table 1), but a thinner (t for ocelli D and F. The mean values vary test: t = -1.54, df = 7, P=0.168) and between a minimum of 15.3 ± 4.03 μm for narrower (t test: t = -2.52, df = 1, P=0.23) ocellus D and a maximum of 21.4 ± 3.96 lens, although these differences are not for ocellus F (Table 2). statistically significant. The focal length for all the Turnip moth The optical properties of the ocelli was measured. However, ocellus C had only one picture that was clear enough ocelli to permit any measurements (Figure 2 Out of the nine Bogong ocelli studied, Turnip moth). The mean f of the Turnip three were removed due to lens distortions moth ocelli (183 μm) is smaller than that or low picture quality (Figure 2 Bogong of the Bogong moth (198 μm ± 34.166) moth). For the first three ocelli, no back although not statistically different (Mann– focal distance (BFD) was obtained. Whitney U = 3262.5, n1 = 69 n2 = 111, P=0.095 two-tailed) (Table 2). The focal length f of the lens in the Contributing to this, is ocellus A (Figure Bogong moth ocelli was found to range 3b), which has values varying around 225 between 166 μm and 248 μm. There are μm. Except for this ocellus, the differences differences in the values of each Bogong moth ocellus, with ocelli C (221 ± 2.987

9

Figure 1. Images on which ocelli size measurements were performed from the stereomicroscopy, SEM and plastic embedding. Panel A: the position of the ocelli on the Bogong moth’s head with 1 – scales present, 2 – scales removed, 3 – fluorescent imaging. Panel B: Bogong moth ocelli used in measurements, image 2 – fluorescent imaging. Panel C: Bogong moth ocelli SEM imaging. Panel D: Bogong moth ocelli, plastic embedding images 1-3 are from the same individual. Panel E: Turnip moth ocelli, image 2 fluorescent imaging. Panel F: Turnip moth ocelli, plastic embedding, both images come from the same individual. Notations: O – ocellus, L – Lens, P – the photoreceptor layer of the retina, c – chitin sheath of the ocellus, h – measured height of the ocellus, dA&B – the measured diameters of the ocellar lens seen from above, t – the measured thickness of the lens, w – the measured width of the lens. Image D4 courtesy of Professor Willi Ribi of the Department of Neurobiology, Research School of Biology, ANU, Canberra, Australia.

10

Table 2. Size measurements of the optical properties of the ocelli (labeled A-I) for both species. f represents the focal distance and BFD back focal distance (p= p-value, N = number of measurements for each ocellus – three measurements/image).

Ocelli ID A B C D E F G H I Mean Optical value property Bogong f 188 ± 221 ± 168 ± 166 ± 169 ± 248 ± 197.7 ± 3.51 2.99 4.85 3.57 6.061 5.05 34.17 p=0.87 p=0.97 p=0.55 p=0.88 p=0.16 p=0.79 p<0.001 N=15 N=6 N=3 N=12 N=15 N=18 N=69 Bogong 15.3 ± 16.4 ± 21.4 ± 16.8 ± 19.1 ± 18 ± 17.8 ± BFD 4.03 2.83 3.96 1.91 3.59 3.08 3.74 p=0.01 p=0.23 p=0.011 p=0.25 p=0,23 p=0.21 p=0.002 3 N=9 N=9 N=9 N=9 N=9 N=54 N=9 Turnip f 220 ± 164 ± 176 ± 182 ± 179 ± 179 ± 181 ± 183 ± 6 2.09 4.01 7.78 4.53 4.66 2.6 16.77 p=0.51 p=0.83 p=0.76 p=0.09 p=0.76 p=0.08 p=0.48 p<0.001 N=15 N=18 N=3 N=24 N=15 N=15 N=21 N=111 Turnip 18.5 ± 20.6 ± 20.1 ± 23.2 ± 22 ± 21.6 ± 23.2 ± 21.3 ± BFD 3.95 3.61 2.84 2.84 2.13 3.7 2.12 3.37 p=0.24 p=0.74 p=0.39 p=0.30 p=0.81 p=0.18 p=0.91 p=0.044 N=9 N=9 N=9 N=9 N=9 N=9 N=9 N=63

Figure 2. Images on which optical measurements of the ocelli were performed for the Bogong and Turnip moth. between the values of the remaining ocelli The back focal distance values show no are smaller (Figure 3a). normal distribution whereas the values for each individual ocellus are normally There is a large overlap between the focal distributed (Table 2). The mean values for lengths of both species. In Figure 3e the the back focal distance vary between a values for ocellus A represent the top minimum of 18.5 ± 3.95 μm for ocellus A outliers for the Turnip moth and the and a maximum of 23.2 ± 2.84 μm for bottom outlier is a value from ocellus D. ocellus D and 23.2 ± 2.12 μm for ocellus G (Table 2). The back focal distance of the

11

Figure 3. Boxplot graphs illustrating the focal distance data. A, C show all the measurements for each image used for the Turnip and Bogong moth. B, D show all the measurements for each ocellus. E shows all the image data per species. Each outlier represents the measurement size in micrometers. turnip moth is significantly larger than that short fibers end, and long neuronal fibers of the bogong moth (Mann–Whitney U = start, four on the right side and three on the 752.5, n1 = 54 n2 = 63, p<0.001 two- left side. The long fibers from the right tailed). side divide in half and form one projection field on each side. Out of the three fibers Analysis of brain projections coming from the left side, only one of from ocelli them forms a projection field on the opposite side. Thus, each projection field The Neurobiotin injections led to the contains fibers coming from both sides. In imaging and reconstruction of both long this individual, it was impossible to and short fibers of the Bogong moth ocelli. differentiate between the upper and lower Thus, from the four prepared brains, two divisions of the central complex. Instead, have shown enough detail to enable the the overall approximate location of these reconstruction of both the ocellar neuron structures are represented together with the bundles as well as important brain parts. reconstruction of the nodules. In the first individual, both mushroom In the analysis of the second individual, it body calyxes as well as each pedunculus was possible to reconstruct only one of the were reconstructed (Figure 4 A2). The mushroom body calyxes together with its short fibers coming from the ocelli pedunculus. However, here it was possible photoreceptor layers follow closely the to reconstruct the two divisions of the posterior contours of the mushroom central complex. The path of the short bodies. Approximately one third of the fibers is dorsal the pedunculus and width of the mushroom body calyx the posterior the mushroom body, similar to

12

Figure 4. Anterior (A1, B1) and posterior (A2, B2) views of several brain structures and stained Bogong moth ocellar L-neurons. The grey neuropils come from the model brain of this species (used with the permission of Stanley Heinze). The colored neuropils represent my own reconstructions of different brain structures (AL – antennal lobe; CA – calyx body; CCA – central complex area; CBU – central body upper division; CBL – central body lower division; LOP – lobula plate; LO – lobula; ME – medulla; NO – nodules; PB – protocerebral bridge; UR – undefined regions) as well as the short (SFa – neuropil reconstruction SFb – voltex reconstruction) and long fibers (LF) of the ocellar neurons together with a neuron soma (NS). the path in the first individual (Figure 4 least four long fibers, half of each B1&2). In this reconstruction, four long originating from the right ocellus and half fibers are present that form two projection from the left ocellus. fields on both the right and left side. These projection fields are anterior of the central No reconstruction was possible for the complex, posterior to the brain surface and Turnip moth due to either damage close to the esophageal hole. For this occurring during the dissection and individual it was possible to even fixation steps or due to low staining of the reconstruct the soma of a neuron (Figure 4 nerve cells themselves. B1&2 NS). Discussion The long fibers form rather large projection fields in both individuals that The purpose of the present study was to are located in the currently undefined investigate the external and internal regions of the brain. It is very probable morphology of the ocelli in Turnip and that each projection field is formed of at Bogong moths as well as the neural projections from these structures. This,

13 together with comparing these two species Turnip moths and Bogong moths are rather with each other, may shed light on the symmetrical. function of these sensory organs. Characterization of the internal Characterization of the external morphology of the ocelli morphology of the ocelli Unlike the external surface of the lens, the The external morphology of the ocellus is internal ocellar surface of the Bogong nearly identical in both of the studied moth (as seen in the plastic embedded species, where the Bogong moth has a sections) is less symmetrical and rather flat slightly higher ocellus. Both species had with no pronounced curvature like that abnormally large individuals (Bogong found in the honeybee ocelli (Hung and moth – A, Turnip moth – I, Figure 3a&c), Ibbotson, 2014), as seen in the plastic which affected the mean value of the embedded sections. This can be observed measurements and skewed the dataset. from our measurements as well, where the thickness is generally smaller than the Unlike honeybees (Ribi et al., 2011), diameter (Table 2). wasps (Warrant et al., 2006) and dragonflies (Berry et al., 2007), the studied The fact that the ocellar lens is not species have only two ocelli, placed perfectly spherical results from the laterally on the vertex of the head, close to differences between the two perpendicular the dorsal margin of the compound eyes diameters of the lens. This however, and posterior to the antennae. This location cannot be seen in the plastic sections, as is typical for moths species, but not unique these were performed in the longitudinal among flying insects (Mizunami, 1994). plane of the ocellus. In the studied species, the ocellar surface Because of the shape of the lens, the was completely smooth, unlike in retinal is also different from that of the honeybees, where between half and two honeybees (Ribi et al., 2011), which thirds of each lens has a rough surface presents a division into a dorsal and a (Ribi et al., 2011). In the present study, ventral surface. The vitreous body any lens surfaces that were not smooth underlying the retina forms a thin and were caused by partial collapses of the uniform layer, that closely follows the surface, due to the treatment of the sample. shape of the lens. Because of the high degree of fragmentation in our plastic This was observed in all the SEM prepared sections, no more information can be samples. The same effect can be seen in obtained about the internal structure of the the images obtained by Grunewald and Bogong moth ocelli. Information about the Wunderer in 1996. The measurements internal structure of the Turnip moth performed using this method are not ocelli, is missing as well, although there is entirely accurate due to distortions, not due a high probability that is very similar to to individual size variation. It can be seen that of the Bogong moth, due to the overall that the diameters of the lens were with similar external shape and dimensions. approximately 20 micrometers larger in the SEM sample compared with the What caused this high degree of stereomicroscopy samples (Table 1). If the fragmentation and poor fixative ocellar lenses of the honeybees (Hung and penetration in our preparations is still Ibbotson, 2014), orchid bees (Euglossa unknown. The fact that these samples had imperialis) (Taylor et al., 2016) and lateral large parts of their eyes removed as well, dragonflies (lateral ocellus: Berry et al., did not seem to improve the results. The 2007)ocelli are asymmetrical, the lenses of same large degree of fragmentation was present as in the paraffin and cryo sections.

14

The method used is common for insect dragonflies, which possess underfocused plastic embedding. Another possible ocelli as well (Stange et al., 2002). There explanation may be the thickness of the are however some exceptions such as the slice, as our slices were three micrometers orchid bees (Euglossa imperialis) (Taylor thick versus 0.8 micrometers (as used for et al., 2016) as well as the nocturnal sweat D4 sample. bee (Megalopta genalis), the nocturnal paper wasp (Apoica pallens) or the diurnal The dimensions of the lens are similar paper wasp (Polistes Occidentalis) between the D4 individuals and those (Warrant et al., 2006) in which the focal obtained by me. The height of the D4 plane of the lens is projected on the retina. ocellus was smaller than those the studied ocelli. However, this may be explained by Neural ocellar projections individual size variation or distortions in my sections. The number of neurons projecting from each ocellus in our preparations, was Why the cryo sectioning and paraffin found to be between three and four. embedding methods did not work, is still Typically, half of these form projections unknown, although poor fixative fields on the ipsilateral side of the brain penetration seems likely to be the probable (close to the esophageal), whereas the cause. other half project to the contralateral side of the esophageal hole. The diameters of The optical properties of the ocelli these fibers are very large and their The optical properties of the ocelli were numbers low. Comparison with other investigated by measuring the focal length species is difficult, as most studies have f and back focal distance BFD of the lens. been done on tri-ocellar species with The measurements obtained from both divided retinas (created by due to a strong species are rather similar. Thus, the curvature of the inner surface of the ocellar Bogong moths which have a thinner and lens). narrower lens, had a larger focal length It is worth mentioning that both the (197.7 µm), whereas the back focal descending ocellar neurons of both distance was smaller (17.8 µm) than the dragonflies and honeybees also form Turnip moth (f = 183 µm and BFD = 21.3 projection fields in the area around the µm, Table 2). esophageal hole. However, unlike our In our study, we found no astigmatism like findings, the neurons of these species often that present in honeybees (Ribi et al., descend lower into the suboesophageal 2011), the nocturnal sweat bee (Megalopta ganglion (Berry et al., 2006;Berry et al., genalis), the nocturnal paper wasp (Apoica 2007;Hung and Ibbotson, 2014). In our pallens) or the diurnal paper wasp species, the projection fields of the (Polistes occidentalis) (Warrant et al., descending large second order neurons 2006). Thus, the ocelli of our species (LD-neurons) are found in the currently havethe same focal plane for each axis of undefined regions of the brain, where no the lens, which in the case of the Bogong known brain structures have as yet been moth, lies well behind the retina whose defined. distal surface lies 5 microns behind the rear surface of the lens. Due to The role of the ocelli in Bogong unsuccessful plastic sectioning, it is still and Turnip moths unknown if the Turnip moth ocelli are also The ocelli of the Bogong moths are underfocused, although it is probable, due strongly underfocused and there is a high to the even higher back focal distance. Our probability that this is also true for the findings are similar to those for other Turnip moths as well. Because of this, it is insect species such as locusts, flies and 15 highly improbable that the poor spatial looper moth and in Arctiid moths (Eaton et resolution enables these two species to use al., 1983;Sprint and Eaton, 1987). A their ocelli for navigation using patterns highly light sensitive organ that can such as the spatial pattern of leaves in a accurately control the initiation and forest canopy. cessation of diurnal activities, is most certainly advantageous to a long distance Without a more detailed study of the migratory species such as the Bogong rhabdoms of these two species, it is moth, for which the timing of nocturnal impossible to say with any degree of migratory flight is important. certainty whether the ocelli are capable of detecting any pattern of skylight There are still many questions to be polarization. However, the brain projection answered, before a definite role for the fields found in the present study, make ocelli can be assigned in these two species. celestial polarization analysis improbable This can only be achieved by further as well. In other insects, polarization research on the structure of the retina and sensitive ocelli have LD-neurons that photoreceptors, the size and extent of the project into either the optic lobe, the visual fields of the lenses, and the anterior optic tubercle or the central responses of ocellar neurons to different complex (el Jundi et al., 2014). Our light stimuli. neurons project to neither of these polarized light processing structures.

The size and number of the LD-neurons, reveal that they have a high information Acknowledgements convergence as well as very fast signal I thank my supervisors Eric Warrant and propagation. These characteristics, may Stanley Heinze for offering me the point to a role of the ocelli in flight opportunity to conduct this study, for stabilization reflexes, by directly providing the necessary equipment and for connecting the ocelli to the motor centers their invaluable guidance. I also thank of the brain. This of course may prove Carina Rasmussen and Ola Gustafsson for beneficial for both a migratory as well as a their help with the laboratory methods, non-migratory species. plastic embedding and sectioning and the The two laterally positioned ocelli of SEM preparation and analysis of the Cabbage and Arctiid moths (Grunewald samples. I also thank Willi Ribi for and Wunderer, 1996; Pappas and Eaton, allowing me to use one of his images of 1977) have been shown to perform an the plastic sections of the ocellus of the important function in the circadian Bogong moth. regulation of diurnal activities (Eaton et al., 1983; Sprint and Eaton, 1987; References Yamazaki and Yamashita, 1991). Typical for the ocelli of many species, the ocellar Beetz, M. J., el Jundi, B., Heinze, S. and lens of the Arctiid moths creates an image Homberg, U. (2015). Topographic plane that lies behind the retinal layer organization and possible function of the (Grunewald and Wunderer, 1996; posterior optic tubercles in the brain of the Mizunami, 1994). desert locust Schistocerca gregaria. J. Comp. Neurol. 523, 1589-1607. These findings are similar to those in the Berry, R. P., Stange, G. and Warrant, E. two species studied here. Thus, another J. (2007a). Form vision in the insect dorsal role for the ocelli may be to regulate the ocelli: An anatomical and optical analysis initiation and cessation of diurnal activities of the dragonfly median ocellus. Vision such as flight, as found in the Cabbage Res. 47, 1394-1409.

16

Berry, R. P., Warrant, E. J. and Stange, Dow, M. and Eaton, J. (1976). Fine- G. (2007b). Form vision in the insect Structure of Ocellus of Cabbage-Looper dorsal ocelli: An anatomical and optical Moth (Trichoplusia-Ni). Cell Tissue Res. analysis of the Locust Ocelli. Vision Res. 171, 523-533. 47, 1382-1393. Eaton, J. (1971). Insect Photoreceptor - Berry, R. P., Wcislo, W. T. and Internal Ocellus is Present in Sphinx Warrant, E. J. (2011). Ocellar Moths. Science 173, 822-&. adaptations for dim light vision in a Eaton, J. (1976). Spectral Sensitivity of nocturnal bee. J. Exp. Biol. 214, 1283- Ocelli of Adult Cabbage-Looper Moth, 1293. Trichoplusia-Ni. Journal of Comparative Berry, R., Stange, G., Olberg, R. and Physiology 109, 17-24. van Kleef, J. (2006). The mapping of Eaton, J., Tignor, K. and Holtzman, G. visual space by identified large second- (1983). Role of Moth Ocelli in Timing order neurons in the dragonfly median Flight Initiation at Dusk. Physiol. Entomol. ocellus. J. Comp. Physiol. A -Neuroethol. 8, 371-375. Sens. Neural Behav. Physiol. 192, 1105- el Jundi, B., Pfeiffer, K., Heinze, S. and 1123. Homberg, U. (2014). Integration of Berry, R., van Kleef, J. and Stange, G. polarization and chromatic cues in the (2007). The mapping of visual space by insect sky compass. J. Comp. Physiol. A - dragonfly lateral ocelli. J. Comp. Physiol. Neuroethol. Sens. Neural Behav. Physiol. A -Neuroethol. Sens. Neural Behav. 200, 575-589. Physiol. 193, 495-513. Fent, K. and Wehner, R. (1985). Ocelli - Britton, D. Agrotis infusa Bogong Moth a Celestial Compass in the Desert Ant (Boisduval, 1832) . Atlas Of Living Cataglyphis. Science 228, 192-194. Australia 2016,. Grunewald, B. and Wunderer, H. Caldwell, J. C., Fineberg, S. K. and (1996). The ocelli of arctiid moths: Eberl, D. F. (2007). Reduced ocelli Ultrastructure of the retina during light and encodes the leucine rich repeat protein dark adaptation. Tissue Cell 28, 267-277. Pray For Elves in Drosophila Horie, T., Sakurai, D., Ohtsuki, H., melanogaster. Fly 1, 146-152. Terakita, A., Shichida, Y., Usukura, J., Chen, Q. and Hua, B. (2014). Kusakabe, T. and Tsuda, M. (2008). Ultrastructure of dorsal ocelli of the short- Pigmented and nonpigmented ocelli in the faced scorpionfly Panorpodes brain vesicle of the ascidian larva. J. kuandianensis (Mecoptera: Panorpodidae). Comp. Neurol. 509, 88-102. Micron 59, 8-16. Hu, K. and Stark, W. (1980). Roles of Chumakov, M. A. and Kuznetsova, T. Drosophila Ocelli and Compound Eyes in L. (2008). Pests Agrotis segetum (Denis & Phototaxis. Journal of Comparative Schiffermueller) - Turnip Moth. Physiology 135, 85-95. Interactive Agricultural Ecological Atlas Hung, Y. and Ibbotson, M. R. (2014). of Russia and Neighboring Countries. Ocellar structure and neural innervation in Economic Plants and their Diseases, Pests the honeybee. Front. Neuroanat. 8, 6. and Weeds 2016,. Kalmus, H. (1946). Correlations between CSIRO. Bogong Moth Agrotis infusa. flight and vision, and particularly between Museum 2016,. wings and ocelli, in insects. Proceedings Dickens, J. C. and Eaton, J. L. (1974). of the Royal Entomological Society of Fine-Structure of Ocelli in Sphinx Moths. London (A) 20, .84-96. Tissue Cell 6, 463-470. Lindauer, M. and Schricker, B. (1963). Dickens, J. and Eaton, J. (1973). On the function of ocelli in twilight flight External Ocelli in Previously of the honey bee. Biol. Zentralbl. 82, 721- Considered to be Anocellate. Nature 242, 725. 205-206. 17

Mizunami, M. (1994). Information- Sprint, M. and Eaton, J. (1987). Flight Processing in the Insect Ocellar System - Behavior of Normal and Anocellate Comparative Approaches to the Evolution Cabbage Loopers (Lepidoptera, of Visual Processing and Neural Circuits. ). Ann. Entomol. Soc. Am. 80, Advances in Insect Physiology, Vol 25 25, 468-471. 151-265. Stange, G. (1981). The Ocellar Mizunami, M. (1995). Functional Component of Flight Equilibrium Control Diversity of Neural Organization in Insect in Dragonflies. Journal of Comparative Ocellar Systems. Vision Res. 35, 443-452. Physiology 141, 335-347. Pappas, L. G. and Eaton, J. L. (1977). Stange, G., Stowe, S., Chahl, J. and The internal ocellus of Manduca sexta: Massaro, A. (2002). Anisotropic imaging electroretinogram and spectral sensitivity. in the dragonfly median ocellus: a matched J. Insect Physiol. 23, 1355-1358. filter for horizon detection. J. Comp. Pappas, L. and Eaton, J. (1977). Large Physiol. A -Neuroethol. Sens. Neural Ocellar Interneurons in Brain of Cabbage- Behav. Physiol. 188, 455-467. Looper Moth Trichoplusia-Ni Taylor, G. J., Ribi, W., Bech, M., Bodey, (Lepidoptera). Zoomorphologie 87, 237- A. J., Rau, C., Steuwer, A., Warrant, E. 246. J. and Baird, E. (2016). The Dual Rence, B., Lisy, M., Garves, B. and Function of Orchid Bee Ocelli as Revealed Quinlan, B. (1988). The Role of Ocelli in by X-Ray Microtomography. Curr. Biol. Circadian Singing Rhythms of Crickets. 26, 1319-1324. Physiol. Entomol. 13, 201-212. The Cooperative Research Centre for Ribi, W., Warrant, E. and Zeil, J. National Plant Biosecurity. Turnip moth (2011). The organization of honeybee (Agrotis segetum) Threat Data. ocelli: Regional specializations and CRCPlantbiosecurity. rhabdom arrangements. Struct. Toh, Y. and Okamura, J. (2007). Dev. 40, 509-520. Morphological and optical properties of Schmitt, F., Stieb, S. M., Wehner, R. the corneal lens and retinal structure in the and Roessler, W. (2016). Experience- posterior large sternma of the tiger beetle related reorganization of giant synapses in larva. Vision Res. 47, 1756-1768. the lateral complex: Potential role in van Kleef, J., James, A. and Stange, G. plasticity of the sky-compass pathway in (2005). A spatiotemporal white noise the desert ant Cataglyphis fortis. Dev. analysis of photoreceptor responses to UV Neurobiol. 76, 390-404. and green light in the dragonfly median Schricker, B. U. R. K. H. A. R. D. ocellus. J. Gen. Physiol. 126, 481-497. (1965). The orientation of the honey bee at Vrsansky, P. (2008). Central ocellus of dawn. And at the same time a contribution extinct cockroaches (Blattida: to the question of ocelli function in bees Caloblattinidae). Zootaxa, 41-50. [Engl. summ.]. Z Verglei Chende Physiol Warrant, E. J., Kelber, A., Wallen, R. 49, 420-458. and Wcislo, W. T. (2006). Ocellar optics Schwarz, S., Albert, L., Wystrach, A. in nocturnal and diurnal bees and wasps. and Cheng, K. (2011a). Ocelli contribute Arthropod Struct. Dev. 35, 293-305. to the encoding of celestial compass Warrant, E., Frost, B., Green, K., information in the Australian desert ant Mouritsen, H., Dreyer, D., Adden, A., Melophorus bagoti. J. Exp. Biol. 214, 901- Brauburger, K. and Heinze, S. (2016). 906. The Australian Bogong Moth Agrotis Schwarz, S., Wystrach, A. and Cheng, infusa: A Long-Distance Nocturnal K. (2011b). A new navigational Navigator. Front. Behav. Neurosci. 10, 77. mechanism mediated by ant ocelli. Biol. Wehner, R. and Strasser, S. (1985). The Lett. 7, 856-858. Pol Area of the Honey Bees Eye -

18

Behavioral Evidence. Physiol. Entomol. Wunderer, H. and Dekramer, J. (1989). 10, 337-349. Dorsal Ocelli and Light-Induced Diurnal Wehrhahn, C. (1984). Ocellar Vision and Activity Patterns in the Arctiid Moth Orientation in Flies. Proceedings of the Creatonotos-Transiens. J. Insect Physiol. Royal Society Series B-Biological Sciences 35, 87-95. 222, 409-411. Yamazaki, S. and Yamashita, S. (1991). Wellingt.WG. (1974). Bumblebee Ocelli Efferent Control in the Ocellus of a and Navigation at Dusk. Science 183, 550- Noctuid Moth. J. Comp. Physiol. A-Sens. 551. Neural Behav. Physiol. 169, 647-652. Wilson, M. (1978). Functional Organization of Locust Ocelli. Journal of Comparative Physiology 124, 297-316.

19