- Top of page
- Traditional and modern techniques in insect morphology
A brief account of the history of insect morphology is given. Different techniques and analytical methods used in current projects on insect morphology and phylogeny and their optimized combined application are described. These include fixation, dissection, maceration, histology (microtome sectioning), scanning electron microscopy (SEM), transmission electron microscopy (TEM), serial block-face scanning electron microscopy (SBFSEM), focused ion beam scanning electron microscopy (FIB/SEM), confocal laser scanning microscopy (CLSM), bleaching, micro-computed tomography (μCT), computer-based three-dimensional reconstruction, focus stacking of digital images, geometric morphometrics and the storage of morphological metadata. The role of insect morphology in the “age of phylogenomics” is discussed.
- Top of page
- Traditional and modern techniques in insect morphology
After a marked decline in the last decades of the last century, research in insect morphology and anatomy has enjoyed a remarkable renaissance in the last ten years. Innovative methods and new theoretical concepts have given new strong impulses to this discipline. An overview of more traditional and modern techniques and their optimal combination is the main topic of this contribution.
A detailed account of the history of insect morphology and anatomy was presented by Gupta (1994). Therefore the historical development will be outlined only briefly here. The different meanings of the terms morphology and anatomy were pointed out by Gupta (1994), but without mentioning that “Morphologie” was coined by the German poet Johann Wolfgang von Goethe in 1779, and later independently by the German anatomist Karl Friedrich Burdach in 1800. In contrast to anatomy (Greek: aná = on, tomé = cut; investigation of the shape structure, and position of organs or body parts) morphology (Greek: morphé = form, shape, lógos = word, teaching, ratio) always implies a comparative aspect and the concept of homology plays an important role.
We owe the earliest scientific entomological information to the famous Greek philosopher and naturalist Aristotle (384–322 BCE). Among other groups of animals, insects are also treated in his eminent work De Partibus Animalicum. However, the outstanding scholar pointed out that insects were so insignificant that they weren't worth of the types of investigations dedicated to vertebrates such as “fish”, “reptiles” and mammals. In Western Europe the roots of scientific entomology and insect morphology go back to the 17th century. In 1669 the Dutch naturalist, anatomist and microscopist Jan Swammerdam (1637–1680) published his remarkable book Historia Insectorum Generalis (“The Natural History of Insects”), focused on the development and metamorphosis of insects. Swammerdam's investigations were remarkably modern, already characterized by careful dissections, a comparative approach and the efficient use of microscopy. A prominent Italian researcher of this century was Marcello Malpighi (1628–1694) who discovered the excretory tubules named after him, the insect heart and the anastomosing tracheal system (Gupta 1994). A distinguished naturalist and entomologist of the 18th century was August Johann Rösel von Rosenhof who published numerous beautiful illustrations of insects and other arthropods including developmental stages and, in some cases, also anatomical details. His famous Insecten-Belustigung (“Insect Amusement”) was published in 1740. He already attempted to use a natural classification and can be considered as one of the founders of entomology in Germany. Many of Carl von Linné's later descriptions of insects are based on Rösel's work. Eminent entomologists of the 18th century were Johann Christian Fabricius and Pierre Latreille. However, their focus was more on classification than on morphology in a stricter sense. An outstanding insect anatomist of the 18th century was Pierre Lyonnet (1708–1789) who described the incredible number of 1647 muscles in the goat moth (Cossus cossus) and discovered the peritrophic membrane, imaginal discs and the prothoracic glands in the caterpillar (Tuxen 1973; Gupta 1994).
Carl Hermann Conrad Burmeister (1807–1892) was an exceptional entomologist, zoologist and paleontologist of the 18th century. He published the first volume of the Handbuch der Entomologie (Burmeister 1832), a remarkable contribution that was later translated into English (Gupta 1994). Charles Janet (1849–1932) was not only a dedicated entomologist and owner of approximately 40 000 fossils, but was also an engineer, inventor and company director. Aside from his studies on plant biology and evolution he excelled as a pioneer of insect histology. His remarkable treatment of the anatomy of social insects was mainly based on serial sections of stunning quality (Billen & Wilson 2008).
Two outstanding entomological works with simple and virtually identical titles were published in the first decade of the 20th century, Les Insectes by Louis Félix Henneguy (1904) and Gli Insetti by Antonio Berlese (1909). Other milestones in the study of insect structures were books published in the 1920s and 1930s by the American entomologists Augustus Daniel Imms (A General textbook of Entomology (Imms 1925) ) and Robert Evans Snodgrass (The Principles of Insect Morphology (Snodgrass 1935) ), and by the German morphologist Herrmann Weber (Lehrbuch der Entomologie (Weber 1933) and Grundriss der Insektenkunde (Weber 1938) ). A highly productive but controversial North American entomologist of the early 20th century was Guy Chester Crampton (e.g. Crampton 1918, 1928). Very important morphological contributions were made by another American insect morphologist, Gordon Floyd Ferris. The understanding of thoracic structures was greatly improved by his profound studies (e.g. Ferris 1940).
Outstanding morphological studies were published by Weber and his students at the University of Tübingen (e.g. Wenk 1953; Weber 1955, 1960). The exceptional beauty and detail-richness of the illustrations in Weber's posthumously published monograph on the elephant louse (Weber 1969) are still unsurpassed, and a similar degree of perfection was reached by other members of his group (e.g. Bierbrodt 1942). The tradition of the Weber school was continued by Gerhard Mickoleit (e.g. Mickoleit 1961, 1963) and students under his supervision (e.g. Burmeister 1976; Rieger 1976). Another outstanding school of insect morphologists was the group of Jean Chaudonneret at the Department d'Entmologie of the University of Dijon, France. Brilliant contributions were published by Chaudonneret himself (e.g. Chaudonneret 1948, 1950–51) but also by some of his co-workers (e.g. Bitsch 1966). A fascinating and extremely detailed anatomical study on the alder fly Chauliodes formosanus was carried out by Takadi Maki at the former Imperial Taihoku University (Maki 1936). Another remarkable contribution by a Japanese insect morphologist was the three-volume work on the evolution of the insect head, thorax and abdomen by Ryuichi Matsuda (Matsuda 1965, 1970, 1976). The work was mainly based on the results of earlier morphological studies but extremely useful as a reference work.
The classical tradition of insect morphology was upheld at a very high level by scientists at the Zoologisk Museum in Copenhagen, notably by the eminent insect systematist Niels Peder Kristensen. He published not only outstanding morphological treatments of lepidopteran key taxa (e.g. Kristensen 1968, 1984) and an entire series of profound reviews of insect phylogeny (e.g. Kristensen 1975, 1991), but also landmark volumes on systematics and morphology of Lepidoptera in the Handbook of Zoology series (Kristensen 1997, 2003).
Morphology-based insect systematics arguably reached a peak with the publication of the groundbreaking works of the German dipterist Willy Hennig (e.g. Hennig 1969), who also revolutionized phylogenetic reconstruction (Hennig 1950). In the last decades of the century, the detailed study of morphological features of insects, especially internal structures, became less and less popular. This development was probably partly correlated with the rise of molecular systematics. Morphology was considered as second-rate information by some systematists. In some primarily molecular studies, long lists with morphological characters were included (e.g. Wheeler et al. 2001). However, in some cases the characters were uncritically extracted from the literature and inadequately coded. Today, most systematists agree that the morphological characters used in morphology-based or combined analyses should be well documented (e.g. Beetle Tree of Life project (BToL, http://insects.oeb.harvard.edu/ATOL/description.htm); Lawrence et al. 2011). This attitude and new technological developments have given new strong impulses to the structural investigation of insects. Innovative approaches such as micro-computer tomography (e.g. Hörnschemeyer et al. 2002) and computer-based reconstruction (e.g. Beutel & Haas 1998), an optimized combined application of different techniques (e.g. Beutel et al. 2011) and the concept of evolutionary morphology (e.g. Wirkner & Richter 2010) have led to a remarkable renaissance in insect morphology in the last decade.
The purpose of this study, which is based on the experience of several research projects, is to outline recent developments in insect morphology and to give a brief overview over traditional and more recently introduced techniques and morphological analysis. This includes practical recommendations for structural investigation and analysis, and for an optimized workflow leading to an increased efficiency in the acquisition of high-quality anatomical data. Although we mainly focus on practical aspects of structural investigations, we decided to include geometric morphometrics, which is an analytical procedure. High-quality visualized data obtained with some of the techniques described here are an excellent basis for this approach, which becomes increasingly important in the context of functional and evolutionary studies.
The main aim of this review is to help students and researchers interested in insect morphology to carry out their projects. Since many techniques were developed in Germany and other European countries, this practical review is mainly based on this knowledge and research experience acquired at German institutions. However, even though the situation in Japan is quite different and the tradition of insect morphology distinctly younger, some technical approaches developed by Japanese researchers are also considered here. Although most of the techniques reviewed here are not widely used by Japanese scientists, the provided information on the specific situation in Japan may build bridges between research groups and accelerate morphological investigations on insects and other groups of organisms.
Traditional and modern techniques in insect morphology
- Top of page
- Traditional and modern techniques in insect morphology
In the following sections, different technical approaches are outlined to encourage and facilitate morphological investigations of insects. Even though this is explicitly a practical introduction, covering all technical details would go far beyond the scope of this contribution. Therefore the study of original literature and detailed technical manuals is highly recommended. A similar outline of technical approaches is also presented in the textbook “Insect Morphology and Phylogeny” (Beutel et al. 2013).
Features of techniques are summarized in Table 1.
Table 1. Summary of features of morphological techniques
|Preferable fixation||e.g. 70% ethanol, FAE||e.g. FAE||–||Primary and secondary fixation (see text)||Same as TEM||Same as TEM||e.g. 70% ethanol, FAE; bleaching is useful||e.g. FAE (also immobilized living organisms)|
|Data acquisition speed||Usually fast, depending on object and structures||Time-consuming||Fast||Very time consuming||Slow but less than TEM||Slower than SEM, faster than SBFSEM||Fast||Fast (depending on equipment)|
|Alignment||Perfect||Problematic||3D images can be obtained by focal stacking of micrographs (see text)||Problematic, loss or deformation of sections||Perfect||Perfect||Perfect||Perfect|
|Sample size||Minimum about 2 mm (depending on aim)||<1 mm to >100 mm (depends on embedding medium and microtome)||Less than 10 cm, limited by chamber size and specimen stage of SEM||Approximately cell organelle level (nm)||<1 mm||Up to a few mm; depends on setup (e.g. anatomy of sensilla)||Usually <3 mm, also very flat larger samples||Diameter about 0.5–50 mm (depends on equipment)|
|Color information||Present, usually modified by fixation||Artificial staining||Grey scale or false color representation||Grey scale or false color representation||Grey scale or false color representation||Grey scale or false color representation||Artificial, depending on laser wavelength||Grey scale or false color representation|
|Voucher||Samples||Slides||Samples||Usually digital images and section series||Digital images||Digital images||Samples||Samples|
|Data size||–||–||Small to medium (MB)||Medium (some MB)||Large (tens of GB/data set)||Large (up to tens of GB/data set)||Small to large (MB to GB)||Large (several tens of GB)|
|Maximum resolutiona||Low||High||Medium to high||Very high||Very high||Very high||Low to medium||Medium to high|
Pinning and drying
A traditional and still widely used method of killing insects is the use of ethyl acetate. Subsequently the specimens are dried, pinned and labeled. The long established procedure is generally used by amateur collectors, but also frequently by researchers, for instance in the framework of biodiversity projects. To store the pinned specimens in entomological boxes or insect cases has the advantages of saving space and of easy accessibility. Dried material is useful for taxonomic investigations and for studying exo- and endoskeletal features including the genitalia. Their use for detailed anatomical investigations and extraction of DNA is limited but in principle they can also be used for molecular phylogeny (e.g. Ohshima & Yoshizawa 2006: see practical information in Yoshizawa & Ohshima 2003; Ohshima & Yoshizawa 2012). In rapidly dried insects soft tissues (e.g. muscles) and DNA are preserved to a certain degree. However, this is not the case if specimens are kept in vials with ethyl acetate for a longer time. Killing insects (especially larvae) in boiling water has the advantage that the external shape is well preserved without shrinking artifacts. However, the material is useless for anatomical study.
A simple and useful preservative widely used by insect morphologists is 70–80% ethanol. Soft tissues are reasonably well preserved and specimens remain relatively flexible after fixation. Material preserved in ethanol can be used for investigating skeletal features and can be stored for a long time provided that drying out is prevented (Fig. 2a). Usually it is also suitable for anatomical study (e.g. histological sectioning). However, the quality of the tissue preservation is lower compared to material treated with fixatives (e.g. Bouin's fluid, see below). In specimens kept for several years, muscles often become detached from their sites of origin and insertion. Usually specimens lose or change their coloration and become more fragile.
For the extraction of DNA specimens are usually fixed in 98–100% ethanol (preferably not denatured). This material is suitable for histological section series. However, artifacts are often caused by dehydration and the specimens become very inflexible and brittle.
This excellent fixative is a saturated aqueous solution of picric acid (15 parts), concentrated formaldehyde (5 parts) and glacial acetic acid (1 part). It should be prepared before use and stored in a refrigerator. Fixation between two hours to several days is required depending on the size of the specimen. Subsequently, specimens should be washed several times in 70% ethanol until the yellow coloration (picric acid) vanishes from dipping ethanol. Contact with the skin should be strictly avoided.
Duboscq Brazil is a slightly modified alcoholic version. Nine parts of saturated alcoholic solution of picric acid (10 g in 100 mL absolute ethanol) are mixed with four parts concentrated formaldehyde and one part glacial acetic acid. Duboscq Brazil is applied and stored like Bouin's fluid.
Formaldehyde – acetic acid – ethanol
Formaldehyde – acetic acid – ethanol (FAE (FAA)) solution, which is composed of 6 parts concentrated formaldehyde, 10 parts absolute ethanol and 1 part glacial acetic acid, is relatively easy to prepare and very useful for insect morphology. It should be stored in a refrigerator. Landowsky's fluid is similar but contains a higher portion of ethanol (17 parts). Specimens should be fixed for 1–24 h. Both solutions are less toxic than fixatives containing picric acid and therefore are more convenient in their application. They are good fixatives for anatomical work based on dissection, histological section series and micro computed tomography (μCT).
Kahle's fluid is a modified version composed of 15 parts ethanol, 5 parts concentrated formaldehyde, 1 part glacial acetic acid and 30 parts distilled water. It is mainly used by North American insect morphologists.
Due to the high alcohol content this mixture of six parts absolute ethanol, three parts chloroform and one part glacial acetic acid penetrates specimens very fast, but causes strong dehydration. This can result in shrinking and hardening of tissues if applied too long. Fixation times of 1–5 hours are appropriate, and after this specimens have to be extensively washed in absolute ethanol.
Simple dissection often yields results rapidly, especially when larger insects are investigated. Razor blades, hand-sharpened pins and fine forceps are simple but useful tools. In well-preserved specimens (e.g. FAE) the attachment sites of muscles can be investigated and movements can be simulated by pulling muscles with forceps. Dissections can be carried out in glycerine or alcohol. Preparation in water is useful for visualizing tracheae, which are air-filled and thus appear silvery-white. The contrast between them and muscles can be enhanced by applying methylene blue solution. Specimens cut sagittally and stabilized by a small lump of Plasticine with a fitting concavity can provide a good impression of the spatial arrangement of structures. In critical-point-dried or freeze-dried specimens muscles can be removed subsequently without damaging the skeletal attachment sites. Steps of dissection can be documented by drawings or microscopic pictures.
Dissections can be efficient but have limitations. Minor structures can be easily overlooked, especially in small insects. Useful results can not be obtained for insects below 2 mm. The precise histological properties of tissues can not be assessed. Moreover, the specimens are more or less destroyed after the dissection.
K.-D. Klass (Museum of Zoology, Dresden) has developed a specific dissecting and drawing technique. In anatomical investigations at least two samples for each species are used: one is macerated in KOH to assess the three-dimensional configuration of sclerites and membranes. The second alcohol-preserved specimen is dissected on a silicon-embedded laboratory dish. The cuticle is extended and fixed with pins to expose the muscles. The cuticle is considered as a continuous plane that is extensively folded and includes areas that are more or less strongly invaginated or evaginated. The details can not be treated here but examples are found in Klass and Matushkina (2012) and Schneider and Klass (2013).
Dissection is the most popular method in Japan, but Japanese researchers tend to focus on sclerotized parts (for detailed explanation of methods see, e.g. Takagi 1970; Ôhara 2006; Ohshima 2013), especially for assessing diagnostic characters for species identification in a taxonomic framework. In this context the male intromittent organ plays a dominant role, recently also including the endophallus (=internal sac). Not only sclerites are used but also reconstructions of membranous swellings occurring during copulation (e.g. Dang 1993; Owada 1995; Imura 2007a,b). In Germany Hünefeld et al. (2013) recently developed a new appliance to reverse the endophallus for easy observation. This minimizes damage of the membranous region and reflects the natural shape during copulation.
A morphological concept, which is still unusual in Japan, was developed by T. Saigusa (Fukuoka, Kyushu) and co-workers. Similar to K.-D. Klass's approach, the cuticle including membranous and semi membranous areas is considered as a continuous sheet, and the body organization results from modeling this structural entity. An illustration method was developed to reflect complicated morphological configurations exactly and to define highly sclerotized parts and membranous areas following R. E. Snodgrass's drawing procedures (see examples in Shirôzu & Saigusa 1971; Ueda & Saigusa 1982; Sugimoto & Saigusa 2001; Yoshizawa & Saigusa 2001). Other examples of precise depictions of complex structures were provided by K. Yoshizawa (Hokkaido University) and K. Suzuki (Toyama) who intensively investigated the wing and wing base morphology in a phylogenetic context. Especially, Yoshizawa and co-workers (Yoshizawa & Ninomiya 2007; Ninomiya & Yoshizawa 2009; Yoshizawa 2011; Yoshizawa & Wagatsuma 2012) have investigated the homology and phylogenetic utility of the wing articulation. Aside from providing phylogenetic information on a higher systematic level, establishing the homology, shape and connection of each sclerite is crucial in a functional context (e.g. Wootton 1979). Another rarely used but valuable technique adopted by K. Suzuki is manual dissection of insects anesthetized with ether to observe anatomical details of internal reproductive organs in vivo. This apparently also reveals informative characters (Suzuki 1988). Using this approach, male and female reproductive organs of chrysomelid beetles were investigated intensively (Suzuki 1974, 1988; Suzuki & Hara 1975; Suzuki & Windsor 1999; Matsumura & Suzuki 2008). The observation of internal reproductive organs using dried specimens is possible in principle but has its limitations (Suzuki 1994). In any case, the simple classical technique of using binocular microscopes combined with a pair of forceps is still highly efficient for mentally reconstructing three-dimensional images of structures and movements.
Maceration is often used to dissolve internal soft parts for a detailed study of skeletal elements. The most frequently used agent is potassium hydroxide (KOH, approximately 5% aqueous solution), which macerates soft tissue efficiently, especially at higher temperatures. Specimens should be kept at 60°C for some hours depending on size. Prolonged maceration results in the loss of the original coloration and sclerites are rendered more or less transparent. The latter effect can be compensated by applying chlorazol black or pyrogallol staining, which renders sclerotized structures blackish. The staining is reversible and can be removed by washing with ethanol. The cleared specimens can be stored in glycerin or 70% ethanol or critical-point-dried and used for detailed SEM study.
Alternative substances are sodium hydroxide (NaOH), lactic acid (C3H6O3), lactophenol (lactic acid + distilled water + glycerin + phenol; 1:1:2:1) or diaphanol (chlorine dioxide – acetic acid). An alternative method is to let insects rot slowly in distilled water at room temperature. After some days or few weeks (depending on size) the decaying soft tissue can be removed using fine forceps. An advantage is that the cuticle retains its original coloration.
Semi-thin sections are still widely used for investigating internal structures and histological features. Traditionally, specimens are embedded in paraffin or celloidin (see, e.g. Heddergott 1939). These substances are soft compared to exo- and endoskeletal structures of most insects and do not rigidly interconnect with the cuticle. The use of teneral (freshly-molted) specimens or of specific chemicals can reduce the effects, but the resulting sections are comparatively thick (5–50 μm) and show a high rate of artifacts, for instance deformation, loss of structures, or interfolding of parts of the sections. An advantage is that after the removal of the paraffin contrast-rich staining, protocols can be applied (e.g. methylene blue + basic fuchsine, Masson's trichrome stain). Recently traditional embedding media were more and more replaced by hard epoxy resins as already previously used in transmission electron microscopy. These new media yielded distinctly thinner sections (0.5–1.5 μm). Furthermore, deformations and the loss of sections were greatly reduced.
Resin encompasses a broad variety of substances of similar mechanical properties, such as methacrylate, epoxy or styrene resin. A specific epoxy resin, araldite, turned out to be highly suitable for sectioning even strongly sclerotized insects (and other arthropods). Presently it is the most popular embedding medium in entomology. However, other resins used for TEM are also suitable. The preparation is almost identical to the procedure described for TEM (see below). Specimens should be well-fixed using FAE or buffered glutaraldehyde, but samples directly stored in 70% ethanol are also suitable. For the latter, a post-fixation with FAE prior to embedding is advisable. In order to increase the permeability for fixatives and the resin, removing appendages (e.g. legs) is recommended. Specimens are gradually dehydrated in an ethanol series (up to pure ethanol) and transferred over several steps of ascending acetone–Araldite mixtures into pure resin. The infiltration can be facilitated by applying vacuum. Finally the samples are separately placed in silicon molds filled with resin. After a minimum of two days hardening at 60°C the resin blocks can be removed from the molds and trimmed as described for TEM.
Ultra-microtomes used for TEM sectioning are usually suitable for producing semi-thin sections (up to 1 μm thickness), but specific microtomes are used in most cases. Sections are prepared using glass- or diamond-knifes. The former are cheap, but the latter are more durable and produce fewer artifacts. After cutting, every section is transferred to a drop of water on a microscope slide with a thin needle or eyelash glued on a glass pipette. As the sections are often compressed by sectioning, they are placed on a heating plate (60°C) for stretching. The dried sections can be stained using basic solutions. In contrast to paraffin embedding, the resin is usually not removed, which impedes several staining protocols. However, a combination of toluidine blue, borax (each 4 parts of 1% watery solution) and pyronin G (1 part of 1% watery solution) applied for about one minute at 40–60°C is a good staining method. After drying, the sections are sealed using cover glasses mounted with agents as Pertex or Eukit. After this they can be stored for a long time and investigated also using oil immersion microscopy at high magnification (100×). A practical description for using other epoxy resins is available in Pernstich et al. (2003).
Until quite recently the microscopic study of hundreds of sections was time-consuming and the reconstruction of structures required strong imaginational skills. Today, motorized microscopes (e.g. slide scanner) semi-automatically digitize sections at high-resolution and in very short times. Computer-based three-dimensional alignment and visualization functions (see below) greatly facilitate fast analysis and documentation of anatomical data.
Despite considerable progress in histology the sections are never completely free of artifacts. The loss of terminal structures (e.g. tips of legs) is often unavoidable. Deformations almost always occur, even though they are minimized if proper resins are used. The production of high-quality serial sections requires special training and the entire process takes at least two weeks. This excludes the detailed screening of a large taxon sampling from the economic perspective. Data can be acquired much more efficiently using micro-computer tomography. Nevertheless the very high optical resolution and reliable and fast discrimination of tissues using semi-thin sections are still essential for investigating anatomical details. The combination of fast, artifact-free computed tomography (CT-scans) for broader sets of taxa combined with histological sections of selected specimens is the ideal solution.
In Japan, some researchers in ecology and evolutionary developmental biology (evo-devo) prefer paraffin (e.g. Ishikawa et al. 2008; Ishikawa & Miura 2009; Yao & Katagiri 2011). The aim is mainly to compare volumes of body elements and to verify the presence or absence of structures among species. In some contexts deformation artifacts typical for paraffin sectioning are more or less irrelevant and results can be obtained with lower costs and less time.
Machida (Sugadaira Montane Research Center, University of Tsukuba) and coworkers have greatly contributed to insect embryology and phylogeny (Fig. 1). To investigate the morphology and development of delicate embryos they used methacrylate resin (Technovit 7100; Kulzer, Frankfurt AM, Germany) and a tungsten carbide steel knife (Superhard Knife; Meiwa Co., Mie, Japan) (Fig. 1, e.g. Machida et al. 1990, 1994a; Tojo & Machida 1998; Ikeda & Machida 2001; Uchifune & Machida 2005; Jintsu et al. 2010; Mashimo et al. 2011). Additionally, Machida developed an automatic vacuum infiltrator which greatly improved the infiltration efficiency and minimized deformation (Machida et al. 1994b; R Machida, pers. comm., 2013).
Figure 1. Microtome and histological sectioning by R Machida (Sugadaira Montane Research Center, University of Tsukuba). Image by K Sekiya.
Download figure to PowerPoint
Figure 2. Preparation and preservation of samples for SEM and an example image. (a) Permanent preservation of samples in 70% ethanol, (b) rotatable specimen holder and sample attached to an insect pin, (c) preservation of samples in plastic cases, (d) SEM image taken using the holder; Diasemopsis comoroensis (Diptera: Diopsidae).
Download figure to PowerPoint
Scanning electron microscopy
Scanning electron microscopy (SEM) is an approach frequently used by entomologists since the 1970s. Samples for this technique are relatively easy to handle and yield excellent results for documenting surface structures at very high resolution. It is conceivable that the intensive use of SEM in the last decades has contributed to the decline of insect anatomy (e.g. musculature), as attractive results can be obtained with a limited amount of training. The visualization of internal soft parts with SEM is possible in principle but clearly limited. Sagittally sectioned specimens dried at the critical point can be examined but this approach yields only a general overview of internal structures. Even histological sections can be studied by SEM after the embedding resin has been removed (Tsutsumi & Machida 2004). SEM images are clearly insufficient for detailed anatomical reconstructions. Aside from this, coloration, degree of sclerotization, and structures below tissues such as membranes, which would be transparent in visible light, can not be visualized using SEM. However, in some cases structures just below the cuticle can be made visible by irradiation of high emission current, i.e. 20–30 kV (H Pohl, pers. obs., 2013).
Before scanning, specimens should be cleaned if the cuticle is covered with secretions, soil particles, food substrates or other materials. Especially in the mouthpart region, structural details can be obscured by adhering substrates. An efficient cleaning method is by ultrasonic sound. However, this can result in the loss of appendages or disintegration in delicate insects, especially larvae or in insufficiently fixed specimens. Surface contamination of strongly encrusted specimens (e.g. larvae using substrates for camouflage) can be removed using KOH (about 1–5%), cleaning agents (e.g. Triton) or diluted sodium hypochlorite solution (4% solution, i.e. antiformin, commercially available) for a few minutes prior to ultrasonic treatment.
After cleaning, specimens are dehydrated in an ethanol series (absolute ethanol, acetone or isoamyl acetate as final medium). Air drying can be sufficient in the case of strongly sclerotized insects (e.g. beetles), but drying at the critical point or freeze-drying using t-butyl alcohol is usually necessary to minimize shrinkage artifacts. Good results can also be obtained by gradually transferring specimens to distilled water and then freeze-drying them. Alternatively, chemical drying with hexamethyldisilazane (HMDS) yields good results in short time (Brown 1994). Specimens are dehydrated up to pure ethanol and kept in pure HMDS two times for 30 min. Finally they are placed in an open Petri dish under a laboratory hood, allowing the fluid to vaporize. Drying at the critical point is comparatively time-consuming, but in most cases yields the best results.
After drying, specimens are mounted on a specimen holder (e.g. a thin wire) with glue or nail polish or on a stub using carbon adhesive pads. A very useful tool is a rotatable specimen holder developed by Pohl (2010) (Fig. 2b). The insect is fixed on a pin on the rotatable arm of this device. The precisely oriented insect can be rotated through 360°, which makes it possible to obtain all standard views (usually dorsal, lateral and ventral views) with a single specimen. Moreover, the hollow basal part of the brass specimen holder absorbs electrons passing the specimen. This results in a homogenous black background and distinctly reduces charging, which is often a problem, especially when a dense vestiture of setae is present (e.g. with Hymenoptera). Usually the dried specimens are sputter coated with a very thin layer of a conductive element (e.g. gold, platinum, carbonate).
A novel approach sometimes used in entomology is environmental scanning electron microscopy (ESEM). This technique makes it possible to examine samples under environmental pressure using specialized electron detectors and pumping systems differing from those of the normal high vacuum SEM mode. This has the advantage that moist or even live insects can be examined with low vacuum without conductive coating. The resolution is lower than in high vacuum but charging of the unprocessed specimens is minimized. The functional principle and merits vs demerits of ESEM compared to standard SEM are discussed in Kirk et al. (2009). ESEM also allows examining museum specimens (including type material), as the surface is not affected.
SEM is popular among Japanese taxonomists and systematists, especially those associated with a university or museum. In most cases it is used to investigate surface structures. Adhesive sheets sometimes used for mounting specimens can emit small particles, which are reflected by the electron beam and can cause a contamination of the detector. This can be avoided with the holder established by Pohl (2010), aside from other advantages (see above). Samples that have to be spread out (e.g. membranous parts of genitalia) can be placed on a small piece of cover glass, which is then mounted on a stub using carbon adhesive pads. Gaps between the piece of the glass and the carbon pad should be covered with an electrically conductive adhesive to minimize electronic charging.
Recently Takaku et al. (2013) established a new method for observing live animals with a high vacuum SEM setup called “nano-suite”. This is used to investigate physical interactions between small organisms or between organisms and substances. It is a new approach to interpret functions of organisms and underlines the future potential of the “classical” SEM.
An attractive project using SEM is presently carried out by M. Ôhara (Hokkaido University Museum) in collaboration with engineers and physiologists (e.g. M. Shimomura, Tohoku University and T Shimozawa, Hokkaido University). It is based on an extensive collection of SEM micrographs of insects deposited in the museum. In regular meetings the images are screened, analyzed and discussed with respect to possible functions of structures of the scanned insects (Fig. 3). It is planned to establish an online database in the near future. This interdisciplinary project may lead to important new perspectives in insect morphology, functional morphology and biomechanics.
Figure 3. (a,b) Regular meetings held by M. Ôhara (Hokkaido University Museum) to analyze and discuss possible functions of structures of the scanned insects based on SEM images.
Download figure to PowerPoint
Transmission electron microscopy
Transmission electron microscopy (TEM) (Fig. 4) is essential for studies at the ultrastructural level. It has played a crucial role in the investigation of the structure and functions of cells and tissues (Stirling & Woods 2002). Even though its role in insect systematics is limited, it is indispensable for the study of specific character systems such as sperm ultrastructure (e.g. Dallai et al. 2012).
Figure 4. TEM in R Machida's laboratory (Sugadaira Montane Research Center, University of Tsukuba). Image by K Sekiya.
Download figure to PowerPoint
The crucial advantage of TEM is the tremendously increased resolution compared to light microscopy (Woods & Stirling 2002). A disadvantage is the complicated and time-consuming preparation of specimens, which also requires the use of toxic substances. A brief description of the fixation and preparation is given below. Specific literature should be used by researchers intending to use TEM; and satisfying results can not be obtained without appropriate training.
As penetration of the material is essential for its efficient fixation, the structures to be examined (e.g. gland tissue) should be isolated. Small insects can be fixed as a whole, but opening the body cavity (e.g. by removing appendages) is necessary to improve the penetration. Glutaraldehyde (2.5%) is usually applied as primary fixative to cross-link proteins. For larger specimens paraformaldehyde can be used. Both are prepared in a phosphate or sodium cacodylate buffer to maintain the pH and, if necessary, with various additives (e.g. sucrose) to control osmolarity. After one to four hours of primary fixation the specimens are washed overnight in buffer. This is followed by secondary fixation in osmium tetroxide (2%) for one to four hours to preserve and stain lipids (Hayat 1981) and finally washing in buffer for a few minutes.
Epoxy resins (e.g. Spurr, Araldite, Epon) are used as embedding materials. As they are immiscible with water, careful dehydration is necessary. A graded ethanol series is used but the time of dehydration should be kept as brief as possible to minimize the risk of extracting cellular constituents (Woods & Stirling 2002). Gradual infiltration of the resin is required, beginning with a 50 : 50 mixture of resin and solvent. Finally the specimen is transferred into fresh pure resin in a separate silicon mold. Polymerization takes place in a heating cabinet.
The polymerized resin block is tightly trimmed as a pyramid around the black object. For sectioning the specimen is fixed in an ultra-microtome. Glass knives or diamond knives are used for ultrathin sectioning. The former have to be replaced frequently, whereas the latter are very durable and therefore much better suited for continuous sectioning. The sections of about 70 nm thickness are transferred to small round cupric holders. Mesh grids provide the highest stability, but the bars of the mesh may hide structures of interest. A slot grid has a large single opening, which allows examining the whole section. The slot is coated with an extremely thin and fragile foil (e.g. formvar, pioloform) to carry the sections.
To enhance the contrast in the electron beam, specimens are treated with two heavy metal agents (see Reynolds 1963). Uranyl acetate is applied to stain proteins and nucleic acids, and lead citrate is used to contrast cytoplasm, membranes and glycogen-rich structures. After drying the sections can be examined.
The three-dimensional applicability of the technique is limited. Slot grids have to be used and the preparation of continuous section series is difficult. Nevertheless, there is no alternative for ultrastructural investigations below the cellular level. Down to the cellular level serial block-face scanning electron microscopy (SBFSEM) can be used as an alternative (see next section).
Even though most Japanese researchers are familiar with TEM, it is only rarely used in entomology. Notable exceptions are the laboratories of R. Machida (University of Tsukuba), T. Tsutsumi (Fukushima University) and S. Niitsu (Tokyo Metropolitan University), where TEM is used routinely (e.g. Machida et al. 1990, 1994a; Ikeda & Machida 2001; Niitsu 2001; Tsutsumi et al. 2005; Uchifune & Machida 2005; Niitsu & Kobayashi 2008; Jintsu & Machida 2009; Mashimo et al. 2011; Niitsu et al. 2011). Obviously the application of this demanding technique depends on the aim of the research. It is only recommended if extremely small objects are under investigation.
Serial block-face scanning electron microscopy
This innovative technique combines SEM surface analysis with ultra-microtomy similar to that used for TEM (Denk & Horstmann 2004). Ultrathin serial sections perfectly suitable for 3D reconstruction are produced (Zankel et al. 2009). Advantages are minimal artifacts, very high 3D resolution and almost automatic data acquisition. The maximum magnification is lower than that of TEM. A disadvantage is the limitation to very small specimens (every dimension below 1 mm; Hörnschemeyer et al. 2012).
Specimen preparation is similar to what is described for TEM (see section above) but contrasting has to be applied prior to resin embedding (block contrasting). Primary fixation in glutaraldehyde (2.5%) is followed by a secondary treatment with osmium tetroxide (1–2 h). As for TEM, uranyl acetate and lead citrate are used for contrast but are applied longer to generate high contents of heavy metals. Additionally substance binding especially to membranes (like thiocarbohydrazide (TCH)) should be applied. Since all the contrasting is done on a comparatively large piece of tissue the ability of contrasting agents to penetrate the tissue is an extremely important factor. To get homogenous contrast throughout the specimen, extreme care should be taken to cut it as small and early as possible in the contrasting process (2 mm in every direction should be the maximum). Heavily sclerotized parts of arthropods are especially problematic, because most contrasting agents penetrate sclerotized cuticle only very slowly or not at all. In such cases it is inevitable to cut away as much of the cuticle as possible, without damaging the structures that are to be investigated.
Careful contrasting with heavy metals is not only necessary to distinguish between different types of tissues but also to reduce charging artifacts during the scanning procedure (Hörnschemeyer et al. 2012). The contrasted specimens are dehydrated in an ethanol series, transferred to two stages of pure acetone and finally embedded in resin. Durcupan has been shown to be most resistant to the electron beam, but other resins such as Araldite can also be used. However, softening of the resin during high-resolution scans may result in artifacts.
The polymerized resin blocks are trimmed to fit into the ultra-microtome in the SEM chamber. In an automatic process the microtome cuts ultrathin sections from 25 to 100 nm, which are discarded. After every section an image of the surface of the block is automatically recorded (Hörnschemeyer et al. 2012). The fully aligned image stack is perfectly suitable for 3D reconstruction.
This novel technique has not yet been applied in insect morphology and phylogeny in Japan. Cardona et al. (2010) emphasized using it to visualize the 3D anatomical configuration of the brain of Drosophila on the level of individual neuronal processes and synapses. Although this technique has strong size-related limitations (Table 1), it has a great potential to investigate anatomy of very small insects (Fig. 5c–e; e.g. first instar larvae of strepsipterans; H Pohl, pers. obs., 2013) and other organisms with a resolution distinctly surpassing that of classical histology.
Figure 5. SBFSEM and images example pictures. (a,b) SBFSEM, Institut für Zoologie und Anthropologie der Universität Göttingen. (c–e) Stylops sp., first instar larva (Strepsiptera: Stylopidae); rendered pictures of a head and a part of prothorax in lateral view (sagittal plane shown in (c) and entire view shown in (d)); (e) SEM image of the surface of the block; the region is shown in (c).
Download figure to PowerPoint
Focused ion beam/scanning electron microscopy
Focused ion beam/scanning electron microscopy (FIB/SEM) is mainly used in material sciences but was also applied in insect morphology recently (e.g. Schmitz et al. 2007; Di Giulio et al. 2012). Dual beam SEMs are equipped with an additional, strongly focused ion beam (e.g. helium, gallium), which allows removing precisely defined parts of the specimen to examine structures below the surface (Knott et al. 2008) or to open small compartments (e.g. surface receptors) (e.g. Stavenga et al. 2004). FIB/SEM samples must be embedded in resin, and specimen preparation is similar to that described for TEM. Due to the highly precise milling process structures can be removed layer by layer (Knott et al. 2008; Di Giulio et al. 2012). The freshly created surfaces are recorded using the electron beam.
When FIB/SEM is used for milling specific structures for observing materials whose covering structure was removed (e.g., Stavenga et al. 2004), specimen preparation is identical to standard SEM-based investigations. Proper fixation (e.g. glutaraldehyde) and an ultrastructure-preserving drying process (e.g. critical point drying) are necessary.
K. Ohta (Kurume University School of Medicine) and co-workers introduced a FIB/SEM facility in Japan. They greatly improved the technique to produce high-resolution images for biological tissues, equivalent to a resolution obtained with TEM (Ohta et al. 2012). An introduction in Japanese is available at: http://www.med.kurume-u.ac.jp/med/anat2/. As FIB/SEM can provide a larger volume of information than TEM, it is foreseeable that it will be more widely used in our discipline in the future.
Confocal laser scanning microscopy
Even though this technique allows for a very efficient investigation of structural details, confocal laser scanning microscopy (CLSM) (Fig. 6) is rarely used in a phylogenetic context. Without staining (e.g. immunostaining), results can be obtained using the autofluorescence of the cuticle and unsclerotized body parts (e.g. muscles). No specific fixation is necessary. Specimens preserved in 70% ethanol are suitable. Using combined lasers with different excitation wavelengths, soft and hard parts can be easily differentiated. Even the degree of sclerotization and the resilin content can be visualized (e.g. Michels & Gorb 2012).
Very small or flat insects and small isolated parts (e.g. genitalia) can be visualized completely (e.g. Schawaroch et al. 2005; Klaus & Schawaroch 2006; Michels 2007). Deeper layers of larger specimens can not be detected without clearing the cuticle. Suitable agents for bleaching are hydrogen peroxide (35% H2O2) (Stüben & Linsenmair 2008), methyl salicylate, lactic acid (Michels 2007) and Murray's clear (=BABB; 1 part benzyl alcohol + 1 part benzyl benzoate) (Zucker 2006; McGurk et al. 2007). Specimens are kept in the solution for one hour to several days. Soft parts are not drastically affected by this procedure and can be visualized after the process (Deans et al. 2012a).
Depending on the thickness of the specimen, the number of scan channels and the settings of the confocal microscope, the recording of high-quality scans can take several hours. It is important to fix the specimen properly during the scan. Even very slight movements strongly affect the image stack. Scans can be obtained with a minimum of preparation time by fixing the specimen in a drop of glycerine between a slide and a high-precision cover glass supported by small spacers (e.g. wax, self-adhesive rings). Ethanol, buffer and distilled water are also suitable, but evaporation can induce movements of the specimen. For longer scans, embedding media with higher viscosities are recommended.
Agarose (1%), glycerine jelly and mowiol are useful, but traditional media such as Canada balsam or Euparal are also useful (e.g. Schawaroch & Li 2007). They also facilitate the controlled adjustment of the specimen without deformation.
Autofluorescence of insect cuticle is induced by a laser wavelength of 488 nm (e.g. Michels 2007). The emitted light is recorded in two separate spectra: green (about 500–570 nm) and red fluorescence (about 580–690 nm). The overlay of both channels allows for detailed imaging of the grade of sclerotization of the cuticle, with membranes appearing green and sclerites brown (e.g. Deans et al. 2012a). Additionally, resilin-rich structures can be visualized by applying UV radiation (405 nm) and recording the emitted blue light (420–480 nm; e.g. Michels & Gorb 2012). The application of glutaraldehyde as a fixation agent increases the fluorescence at 488 nm. An efficient fluorescence staining of the integument is Congo red, which is excited at a wavelength of 561 nm (emission spectrum 570–670 nm; Michels & Büntzow 2010).
Data obtained with CLSM are mainly useful for volume rendering. Maximum projections of the image stack usually result in lower-quality visualizations. The reconstruction of surface-based 3D models is possible for flat or small isolated objects (see above).
Bleaching is often mentioned in the context of CLSM but is also useful for other investigations, such as analysis of the genital fitting (Kamimura & Mitsumoto 2011a). Kamimura and Mitsumoto (2011a,b) adopted BABB solution to bleach Drosophila spp. couples stabilized by an agarose block (the method was described in Kamimura & Mitsumoto 2011b).
Micro-computer tomography (μ-CT)
This highly efficient anatomical technique was established in insect morphology about ten years ago (Hörnschemeyer et al. 2002). The improved hardware and software of modern desktop scanners yield high-resolution scans and make the technique increasingly attractive for the anatomical study of small and medium sized insects (e.g. Wipfler et al. 2012). The maximum resolution is presently about 0.5 μm. High-end scanners produce data with a resolution below 0.1 μm (nanotomography).
The greatest advantage of μ-CT is the highly accelerated (compared to histology) and non-destructive acquisition of almost artifact-free anatomical image stacks. The data are ideal for fast and precise 3D reconstruction. The specimens can be used for SEM or histological sectioning after the μ-CT scans are obtained. This allows for a very detailed morphological documentation with a minimum of material (see below under “work flow”). As a non-invasive technique, μ-CT can also be applied to very rare species or even type material. High-density resolution based on the specific absorption of different tissue types (e.g. skeleton, muscles) is usually recorded with low beam energy, and facilitates the discrimination of structures (e.g. Friedrich et al. 2008).
The preparation of specimens is simple. Usually they are dehydrated in an alcohol series and dried at the critical point. Thus a high contrast between the tissues and the surrounding medium (air) is generated. The specimen can be mounted directly on a holder with superglue or fixed within a small container (e.g. Eppendorf tube, pipette tip). To ensure maximum spatial resolution, the rotation axis should correspond to the longitudinal axis of the specimen. If drying is not possible, scanning in alcohol or distilled water is an alternative. Even living organisms can be used, if they do not move during scans (e.g. Lowe et al. 2013). However, the similar electron density of water or alcohol and the insect tissues greatly reduces the contrast. To enhance the contrast, reversible staining can be applied. The best results are obtained with alcoholic iodine solution (I2E; 1% iodine in pure ethanol) for one to five days followed by washing in alcohol (e.g. Metscher 2009). The stained specimens have to be firmly fixed to avoid movement caused by circulation of the liquid medium. Cotton and foamed plastic are suitable materials for this purpose. The iodine can be removed by washing the specimen in alcohol.
Specimens prepared for TEM investigations can be documented using μ-CT. The osmium tetroxide–stained and resin-embedded samples show good contrast between tissues and embedding media. This is usually not the case if samples embedded for microtome sectioning are used without specific staining. However, such specimens can be scanned with equipment optimized for phase contrast. The technique is also an excellent tool for investigating amber fossils (e.g. Tafforeau et al. 2006; Pohl et al. 2010). Prior to scanning a dispensable piece of amber should be tested to estimate possible darkening of the material. However, this effect can be reversed by placing the specimen under a UV lamp for some time. A disadvantage of phase contrasted μ-CT scans is that different tissue types can barely be distinguished.
Recently two teams successfully established this technique in Japan (Fig. 7). The first team is composed of Shun-ichi Kinoshita, Osamu Sasaki (Tohoku University), Yoshiaki Hashimoto (University of Hyogo/Museum of Nature and Human Activities, Hyogo), Katsuyuki Eguchi (Tokyo Metropolitan University) and Takuji Tachi and Shingo Hosoishi (Kyushu University). Sample images are shown in Figure 7b–d and their activities are described on http://webdb2.museum.tohoku.ac.jp/e-foram/indexj.html and http://www.museum.tohoku.ac.jp/press_info/news_letter/index.htm. The second team comprises Yukihiro Nishikawa (Kyoto Institute of Technology) and Kyohei Watanabe (Kanagawa Prefectural Museum of Natural History). They used μ-CT not only for anatomical research, but also for improving the usage of museum specimens, in ecological studies and in education. A study describing their activities will be published in the near future (Watanabe et al., unpubl. data, 2013).
Figure 7. Development of μ-CT in Japan. (a) μ-CT at Museum of Natural History of Tohoku University and S. Kinoshita. (b–d) Sample images taken by S. Kinoshita; (b) Lema coronata, (c,d) Pristomyrmex punctatus, color variation means density variation of material composition; red: high, blue: low. (e,f) μ-CTs in Kyoto Institute of Technology with K Watanabe (Kanagawa Prefectural Museum of Natural History) (e) and Y Nishikawa (Kyoto Institute of Technology) (f).
Download figure to PowerPoint
Computer-based 3-dimensional reconstruction
Three-dimensional reconstructions can be based on different types of data sets. Most common sources are serial sections, μCT image stacks and CLSM data. Three-dimensional reconstructions greatly facilitate analyses of the spatial arrangement of morphological structures by using virtual section planes computed by the software. Smaller data sets can be visualized on standard desktop computers, but powerful graphic workstations are needed for high-resolution data and extensive imaging procedures (Fig. 8).
Figure 8. Computer-based three-dimensional reconstruction. Tracing materials of pictures of histological sections by RG Beutel (Friedrich-Schiller-Universität, Jena).
Download figure to PowerPoint
Perfectly aligned data as for μ-CT and SBFSEM image stacks can be immediately used for 3D reconstruction and visualization. Data based on other sources (e.g. images of histological or TEM sections) have to be aligned, either using specific functions implemented in commercial software packages such as Amira (FEI Visualization Sciences Group, Mérignac Cedex, France) or AutoAligner (Bitplane Imaris, Zurich, Switzerland) or using open-source tools (e.g. ImageJ).
Volume rendering is an easy way to visualize image data and yields results similar to low-magnification SEM images. An advantage compared to standard SEM imaging is the easy assessment of the thickness of the body wall (and other structures). Volume models can be used to create 3D images (e.g. for red–cyan goggles) or videos (e.g. with ImageJ). In order to show internal features, cutting planes can be applied to the volume rendering. This procedure is fast, but limited in its potential to visualizing complex internal configurations. To obtain more specific 3D reconstructions (e.g. of the skeleto–muscular or nervous systems), segmentation of the data set is necessary. Structures of interest can be manually outlined or semi-automatically labeled throughout the stack using commercial software packages (FEI VSG Amira, Bitplane Imaris, VGStudio MAX) or open source tools (e.g. Reconstruct). The prepared data can be used to produce segmented volume renderings of selected structural complexes, which allows for easy coloration of volume data.
The segmented image stacks can also be used for the automatic creation of surface objects of discrete structures. Unlike volume renderings, surface objects are hollow and represent only the outline of the structures, usually resulting in a more or less simplified 3D model. The main advantages of this visualization technique are the potential for multiple modifications (e.g. simplification, minimizing artifacts), coloration, illumination and animation of anatomical structures using software (e.g. Autodesk, Maya, Luxology Modo and open-source Blender). Surface models can be used to calculate the volume of structures, to carry out finite element analyses (FEA) and print magnified, solid models of the structures by rapid prototyping. Furthermore, surface-based 3D models can be included in scientific presentations and publications using the common pdf file format, allowing for easy exchange of 3D contents.
This technique is not widely used in Japan, but some morphological studies were published by Japanese authors (e.g. Nagashima et al. 2009; Kaji et al. 2011; Adachi & Kuratani 2012; Tokita et al. 2012; Matsumura et al. 2013), even though only a few of them study insects, and some of them in collaboration with European researchers. Nagashima et al. (2009) used free software (imageJ) for 3D reconstructions to illustrate the development of turtles. This technique is frequently used in vertebrate developmental studies (e.g. Nagashima et al. 2009; Adachi & Kuratani 2012; Tokita et al. 2012) and will also be established in the entomological laboratory of R. Machida (University of Tsukuba) in the near future (Y Nakagaki & R Machida, pers. comm., 2013).
Focus stacking of digital images
Focus stacking is a technique for extending the depth of focus (Fig. 9). Due to the special optical conditions the area of the object rendered in optimal focus is extremely narrow when taking photographs of very small objects. Thus focus stacking procedures are very helpful in macro or micro photography. However, the techniques can be applied whenever images similar to photographs are available, for example, SEM micrographs.
Figure 9. Sepsis fulgens (Diptera: Sepsidae). (a) Partly focused image taken with a digital SLR equipped with a macro lens. (b) Focus stacked image of 142 partly focused images.
Download figure to PowerPoint
The focus stacking procedure follows a simple principle: as many differently focused photographs should be taken as are necessary to get well-focused images of all areas of the object that should be visualized in the final picture. Then only the focused areas from each photograph is selected and combined in the final image. This shows much more of the object in perfect focus than any combination of photographic parameters would produce (e.g. lens or aperture). It is possible to carry out all steps of this procedure manually using any software for cutting certain areas from an existing digital picture and pasting them into another file. As this is an arduous and error-prone process, it is recommendable to use specific software for combining the images.
Most of the larger microscope manufacturers offer special systems, for example microscopes where the focus is operated through software, automatically generating microscopic images with an extreme depth of focus. These systems are expensive, but reasonably priced or even free alternatives producing final images of comparable quality are available (see below).
M. Maruyama (Kyushu University) is one of few experts producing high quality images using this technique in Japan (e.g. Ballerio & Maruyama 2010; Maruyama 2010, 2012). Maruyama uses the software Combine ZM, which is freely available from Alan Hadley (http://hadleyweb.pwp.blueyonder.co.uk/), to produce stacks of photos. The procedure is described in detail at: https://sites.google.com/site/myrmekophilos/czm. A commercial but reasonably priced alternative is Helicon Focus (http://www.heliconsoft.com/) or Zerene Stacker (http://zerenesystems.com/cms/stacker), which allows for a very comfortable work flow. With Helicon Focus it is even possible to remotely operate a SLR-camera and stepper motors to adjust the camera's position relative to the object, making semi-automatic image capture and processing possible.
The images produced by focus stacking are ideal, for example, for documenting type specimens, showing stunning pictures of specimens in publications or printing sharp pictures of small or large specimens for exhibitions.
Geometric morphometrics is a collection of approaches for the multivariate statistical analysis of Cartesian coordinate data, usually limited to landmark point locations. Shape is the geometrical information that remains when location, scale and rotational effects are filtered out from an object (Kendall 1977). The analysis of shape is a fundamental element of biological research (Bai & Yang 2007).
Morphometrics in a traditional sense is the application of multivariate statistical analyses to sets of quantitative variables such as width, length, depth, volume and area (Bookstein 1998). This approach was limited in different ways, for instance by the difficulty to assess the homology of linear distances. Another inherent problem is that the geometric relationships among the variables are not preserved. Consequently, alternative methods of quantifying and analyzing morphological shape were explored. In the 1980s, the nature of data gathered and analyzed changed fundamentally, with a focus on the coordinates of landmarks and the geometric information about their relative positions. The developing novel approach was referred to as geometric morphometrics (Adams et al. 2004).
The multivariate part of geometric morphometrics is usually carried out in a linear tangent space to the non-Euclidean shape space in the vicinity of the mean shape. Traditionally, morphometrics was mainly focused on size, which plays an important role. However, geometric morphometrics is better suited to assess information about shape. More generally, it is the class of morphometric methods that preserves complete information about the relative spatial arrangements of the data throughout an analysis. As such, these methods allow for the visualization of group and individual differences, sample variation, and other results in the space of the original specimens.
The crucial first step is the selection of specimens suitable for the addressed scientific question. The second step is to collect photos (using a camera, or based on drawings, SEM micrographs, etc.) based on the same direction rule, which ensures the comparability of shape. Although in principle there is no limitation for the photo size, it should be as small as possible (as long as it is still clear enough). The third step is to build a data file (such as a .tps file) which links all photos involved. The fourth step is to gather landmark or semi-landmark data when the photos are loaded in software, such as TPS-Dig (Rohlf 2006). The procedure for 3D data collection is similar, but different in the 3D photo reconstruction and landmark gathering. Criteria for selecting landmarks and curves are homology, which means that a specific point is considered the “same” in all specimens, adequacy of coverage, consistency of relative position, coplanarity and repeatability.
For most practical applications, the parameters describing the shapes for a sample of homologous landmark configurations are estimated by a Procrustes superimposition. This procedure is a least-squares oriented approach involving three steps. The position, scale and orientation of the specimens are removed, retaining only the shape for which the same position, size and orientation is assumed. After this the shapes are placed on top of one another. Two alternative methods are used: Procrustes superimposition and Bookstein shape coordinates.
The greatest strength of geometric morphometric methods is that graphical representations of results are possible as configurations of landmark points rather than as customary statistical scatterplots. The analysis of landmark data can be estimated by a superimposition procedure, followed by the projection of the aligned coordinates on a linear tangent Kendall's shape space for multivariate analyses, and the graphical visualization of results in terms of the configurations of landmarks (Kendall 1984; Rohlf 1999; Slice 2001). In Kendall's shape space, distances between pairs of points (specimens) approximate the Procrustes distances between the corresponding pairs of landmark configurations. Partial warps, which are the eigenvectors of the bending energy matrix ordered, from the thin-plate spline plus the uniform shape components (Rohlf & Bookstein 2003), are a convenient set of shape variables that can be interpreted as axes for this space. Scores on these axes can then be treated as multivariate data representing shape, and can be used in conventional multivariate analyses (Caldecutt & Adams 1998; Bookstein et al. 1999; Adams & Rohlf 2000; Bai et al. 2010, 2011, 2012; Gharaibeh et al. 2000; Klingenberg & Leamy 2001; Rüber & Adams 2001). Differences in shape among objects can be described not only as a plot, but also in tree form via cluster analysis. Different models, including single linkage (the distance between two units is the distance between the two closest members of those clusters), UPGMA (unweighted pair group method with arithmetic mean; the distance between two clusters is the average of the distances between units in one cluster and units in the other cluster) and Ward's (minimizes the variance of intra-cluster distances), are used to calculate the similarity among clusters.
Geometric morphometrics data can be applied to many research fields, such as phylogeny, development and ecology. This often requires a sophisticated quantitative representation of the phenotype that captures the functional, genetic or developmental attributes that are biologically important for the hypothesis to be tested, by linking the data. As existing coordinate-based geometric morphometric methods can not be easily extended to 3D data, geometric methods for the analysis of 3D data is still a developing area. Further improvements are needed to address an even broader field of problems with greater sophistication than is possible today.
Practical reviews and specific studies by Japanese entomologists are also available (Fukudome & Sakamaki 2011; Konuma 2011; Novkovic 2011; Takahashi 2011; Tatsuta & Sakamaki 2011).
Storage of morphological data in data banks
Impressive amounts of high quality morphological data have been produced in recent years. This increases the need for suitable data storage facilities (e.g. μ-CT raw data). Moreover, to facilitate the maximum use of the information, easy accessibility for members of the scientific community should be guaranteed. A presently available data bank is the Morph⋅D⋅Base. It is an online data repository for morphological metadata and media files and a freely accessible general communication platform for scientists (https://www.morphdbase.de/).
Optimized storage and accessibility is invaluable for scientific progress. Nevertheless, standards and rules for reporting and documenting data are not fully established yet (e.g. Deans et al. 2012b; Vogt et al. 2013).
Depending on the aim of an investigation and the available material and facilities, different combinations of techniques can be used to produce optimal results with maximum efficiency. Even single specimens can be processed to produce very detailed results. Microscopic drawings from different perspectives would usually be the first step. This is useful for assessing the major external features including properties of the cuticle, for example membranous vs sclerotized areas, and also internal structures (e.g. muscles) visible through the transparent or semitransparent integument. CLSM imaging in glycerine provides additional information on external and more or less superficial internal structures. After critical point drying, SEM can be applied for a detailed documentation of surface structures. Micro-computed tomography using the same specimen provides information on the entire external and internal configuration. Finally, 3D reconstruction based on obtained image stacks is an ideal tool for attractive and informative visualization, including animated 3D pdf files, which are increasingly used in morphological publications. Using this or a slightly different protocol, a minimum of material can yield an immense wealth of information without destroying the exemplars. If additional specimens are available, a very detailed documentation of internal details including histological properties of tissues can be obtained using semi-thin sectioning. Ultrastructural studies require specifically fixed additional specimens and the use of TEM or SBFSEM.
The data obtained with the approach outlined here are an ideal basis for morphology-based phylogenetic evaluations and studies in evolutionary morphology, also using molecular phylogenies as a background for developing complex evolutionary scenarios. For investigations mainly focused on comparisons of shapes, geometric morphometric analyses are highly recommended, also in a phylogenetic and evolutionary context.
- Top of page
- Traditional and modern techniques in insect morphology
It was demonstrated in a recent project on the phylogeny of Holometabola that an optimized combined application of more traditional and innovative morphological techniques can greatly facilitate and accelerate the acquisition and documentation of high-quality morphological data. In approximately two years a data matrix containing 356 characters of different body parts and developmental stages was compiled. The results (Beutel et al. 2011) yielded a single highly resolved cladogram (parsimony analysis) and were almost completely congruent with analyses of an extensive molecular data set (Wiegmann et al. 2009).
An often-raised question is whether modern techniques also lead to the discovery of novel characters. The introduction of TEM and SEM doubtlessly revolutionized investigations on the cellular level and of surface structures, respectively, and both techniques have contributed to the exploration of new character systems (e.g. spermatozoan ultrastructure, Dallai & Afzelius 1984; Jamieson et al. 1999). To a much lesser degree new characters can be expected using μ-CT. The advantage of this technique is greatly accelerated data acquisition. Likewise computer-based 3D reconstruction will rarely reveal new phylogenetically relevant features. However, this approach has tremendously facilitated the understanding of complex structural constellations and greatly improved the visualization of structures and characters.
It is apparent that the study of the morphology of insects or other organisms is challenged by the tremendous progress of molecular analyses in the “age of phylogenomics” (e.g. Beutel & Kristensen 2012; Trautwein et al. 2012). It was shown in a recent review of systematic studies (Bybee et al. 2008) that the use of morphological data in insect phylogenetics loses ground when compared to molecular systematics. However, it was demonstrated by Wheeler (2008) that even when morphological parts of combined matrices are seemingly dwarfed by the sheer number of informative sites in extensive molecular data sets (e.g. 775 genes in Meusemann et al. (2010)), they still can have a substantial impact on the resolution of the deeper nodes of the trees, which are often insufficiently resolved with molecular data alone (e.g. von Reumont et al. 2009). Extensive molecular data sets including transcriptomes and complete genomes are already available for a considerable number of taxa (e.g. Meusemann et al. 2010; von Reumont et al. 2012; see also 1KITE http://www.1KITE.org). These data sets will possibly yield very robust phylogenies without using any morphological information (e.g. Niehuis et al. 2012). However, mere branching patterns, robust as they may be, provide only limited insights into the evolution of insects or other groups of organisms. As pointed out in Beutel et al. (2011), the knowledge of the morphological transformations is essential for the reconstruction of complex evolutionary scenarios. It is the phenotype with its morphological features which interacts with the environment and which is primarily exposed to natural selection (Beutel et al. 2009, 2011). The concept of evolutionary morphology outlined by Wirkner and Richter (2010) emphasizes the importance of in-depth morphological investigations in a context that goes beyond mere phylogenetic branching patterns. This also includes investigations in a functional context (e.g. Beutel & Gorb 2001, 2006, 2008) or the detailed study and documentation of the morphology of different life stages of model organisms such as Drosophila or Tribolium.
For several reasons it appears likely that morphology will continue to play a vital role in insect systematics and evolutionary biology. The careful investigation of structures is an essential precondition for understanding the functions of diverse structures, which may have played an important role in the evolution of the groups in question, and are often also helpful in reconstructing relationships (e.g. attachment structures, Beutel & Gorb 2001, 2006; wing base sclerites, Yoshizawa 2011). Morphology provides an independent data set for critically re-evaluating the results of molecular studies (and vice versa). Analyses exclusively based on molecular data may provide robust phylogenies but do not provide a complex picture of a group. Without knowing the transformations on the phenotypic level it is not possible to develop a complex and meaningful evolutionary scenario for insects or any other group with a complex morphology.
A field in which morphology will obviously play an exclusive role is paleontology. Fossils will only provide useful DNA sequences in extremely rare cases, but at least amber fossils can be studied with innovative morphological approaches and yield detailed information (e.g. Pohl et al. 2010). Beutel et al. (2008) demonstrated that the neglect of fossils can lead to serious misinterpretations in phylogenetic reconstruction, and it is trivial that the knowledge and interpretation of extinct taxa is essential for understanding the evolutionary history of any group of organisms.
- Top of page
- Traditional and modern techniques in insect morphology
This research was supported by the National Basic Research Program of China (973 Program) (No. 2011CB302102), the National Natural Science Foundation of China (Nos. 31010103913, 31172143), and by a Humboldt Fellowship (to M.B.) from Alexander von Humboldt Foundation, by the VolkswagenStiftung (to R.G.B.), by the DFG Heisenberg grant (HO2306/7-1) of the Deutsche Forschungsgemeinschaft (to T.H.), by JSPS Postdoctoral Fellowships for Research Abroad (to Y.M.). We also thank K Yoshizawa, S Kinoshita, K Watanabe, K Sekiya, R Machida, Y Nakagaki, M Ôhara, I Ohshima, T Tsumumi, M Shimomura, K Ohta, T Hariyama, M Maruyama and S Niitsu for valuable information on the technical advances in Japan. Thanks to S Kinoshita, K Watanabe, K Sekiya, M Ôhara for offering us images as well.