Paul De Ley

Department of Nematology, University of California, Riverside CA 92521, USA. email: paul.deley@ucr.edu

Biology Department, Ghent University, Ledeganckstraat 35, 9000 Ghent, Belgium. email: wim.bert@rug.ac.be

November 28^th, 2001

Development of the protocols outlined below was supported by a research equipment grant of the Fund for Scientific Research, Belgium (Krediet aan Navorsers G.0292.00). We are grateful to J.G. Baldwin, A. Ploeg and S. Edwards (Department of Nematology, UC-Riverside) for critical reading and helpful comments.

Summary: Morphological identification of nematodes usually requires permanent slides for detailed observation. This significantly limits the range of research methods that can be applied to such specimens. Permanent slides are never truly permanent, as taxonomically important reference material may be lost by accident, or degrades slowly over time. In order to efficiently record the morphology of nematodes, in a format that allows easy archiving, editing and distribution, we have assembled and tested two micrographic Video Capture and Editing (VCE) configurations. These comparatively inexpensive, easily customized systems allow for the production of short video clips that mimic multifocal observation of nematode specimens through a light microscope. These clips can be used for many purposes, including teaching and training, management and online access of taxonomic collections, routine screening of fixed or unfixed specimens, recording of ephemeral staining patterns, or recording of freshly dissected internal organs prior to their decomposition. We provide details on the components and operation of both systems, evaluate their efficiency in the aforementioned applications, and provide illustrations of the obtained image quality. We expect this approach to become widely used in nematology research and teaching.

Most nematodes are transparent micro-invertebrates, and the light microscope (LM) is therefore the first-line tool for nematode identification. This will undoubtedly remain true for a long time, despite the proliferation in recent decades of various ultrastructural and molecular techniques. Unfortunately, many situations occur in which a specimen's micro-morphological characters are irreversibly lost. For example, type specimens deposited in taxonomic reference collections deteriorate inevitably, and are sometimes accidentally lost or damaged, while teaching material often gets destroyed by inexperienced hands long before preservation fails. Specimens prepared for histochemical studies or Scanning Electron Microscopy (SEM) lose transparency during staining or coating protocols, and often deteriorate within hours, days or weeks. Certain kinds of observations on internal structures, such as the cellular architecture of the female gonad, can only be made on freshly dissected specimens, which usually plasmolyse and decompose quickly. Even more drastically, techniques such as Transmission Electron Microscopy (TEM) and molecular analysis require the destruction of the nematode being studied, as a prerequisite to the extraction of ultrastructural or macromolecular data.

Loss of optical morphology of a specimen invariably interferes with subsequent verification of previous identifications and observations. It also complicates procedures for applying multiple methodologies to the same individual nematode. For example, although it is possible to obtain DNA sequence data from single nematodes fixed years earlier in formalin (Dorris & Blaxter, pers. comm.), chances of success are highest when DNA extraction is performed soon after fixation (Thomas et al., 1997) in order to maximize succesful PCR amplification. And regardless of the time elapsed since fixation, at least part of the specimen must inevitably be destroyed. Some of the consequences are, (i) that morphology-based identification of an individual nematode must essentially be complete before its molecular analysis can begin, (ii) that curators of reference collections can rarely permit attempts at molecular analysis of older type specimens, and (iii) that it constrains practical feasibility of large-scale surveys of nematode diversity combining morphological and molecular data of each individual.

For many years, line drawings and still photographs have acted as the sole means of illustrating, distributing and preserving LM-visible morphology. Both types of images have important drawbacks. First, they depend on the skills of the person producing them (especially so for line drawings), and second, they simplify or omit three-dimensional topology of structures (especially so for still photos). Eisenback (1988) presented a simple method for superimposing several focal planes in one photomicrograph, but this method is only applicable to a maximum of five focal planes, and the act of superimposition reduces contrast per focal plane, as well as discarding all topological information along the focal axis.

We have therefore set out to find alternative approaches for recording the morphological characters of nematodes as seen with LM. Taking cues from four-dimensional microscopy, as utilized in comparative embryology of nematodes for multifocal image recording through time (Thomas et al., 1996; Schnabel et al., 1997), we have assembled two simplified, non-automated configurations for multifocal recording of nematode morphology. We assembled both general-purpose Video Capture and Editing (VCE) systems from inexpensive and widely available components, tested them and developed simple protocols for the recording, optimization and distribution of multifocal images of nematode specimens. The resulting video files can easily be exchanged on disk or through networks, and can be viewed with various widely distributed video player programs.

In this paper we describe the various components and protocols used for our VCE installations, as well as their relevant functions, problems and observed disadvantages. We conclude that VCE has a wide range of possible uses in nematological teaching and research, and that it is complementary to still photography rather than a direct methodological competitor or replacement.

In accordance with the financial limitations imposed on the research budget of most nematologists (including our own) we assembled a first basic VCE configuration that was comparatively low-priced, not restricted for use with any particular brand of microscope, and easily customized for different laboratory environments and for a wide range of applications. This setup is further referred to as Configuration 1 (Fig. 1, Table 1), and was used to establish the basic protocol outlined below and summarized in Fig. 2.

To compare the quality of the captured video images with still photography on 35 mm film, multifocal series of pictures of some of the specimens captured with Configuration 1 were also taken with an Olympus SC35 single-lens reflex camera, mounted on the same microscope.

Next, on the basis of our experience with Configuration 1, we assembled a second version of the above system, including various updated components allowing greater flexibility in magnification and more versatility in various other aspects. Changes in this more expensive second setup, Configuration 2, included a separate video monitor, an optical magnification changer, exchangeable 1x and 0.35x camera adapters, and a video capture card equipped for both analog and digital video input (Table 2). Also, in this second system we replaced the semi-apochromat 40x and oil 100x objectives of Configuration 1 with a single apochromat oil 60x objective, and compensated for lower on-screen magnification by utilising a video camera with 1/3" CCD chip instead of a 1/2" CCD chip. The higher numerical aperture of this 60x objective, in combination with the magnification changer and exchangeable camera adapters, allowed capture of images that were comparable in quality to those obtained with the first setup at highest magnifications, but significantly better than those obtained with Configuration 1 at mid-range magnifications. We also added a water immersion 100x objective of lower numerical aperture, to allow us to capture unfixed specimens on slides without coverslip. Finally, the computer was equipped with two hard drives, one fixed disk for program files and one removable disk for captured files, to allow swapping with other computers for e.g. file backup.

Except for the microscopes and optics of either system, we did not opt for any video or computer components designed specifically for microscopy. Instead, we chose general-purpose components that are aimed at more advanced home users, readily available from many different electronics suppliers, manufactured by companies that are among the market leaders in their respective areas, and compatible with many models and brands of other components.

We have not tested alternative brands for the optics, electronics and software we haven chosen, and do not wish to claim or imply any preference for the tested brands. Also, although our configurations were both based on IBM-compatible computer components, various Macintosh-compatible hardware and software counterparts exist and operate along identical principles. As to our own configurations, the respective manufacturers of the different components are listed in Table 6.

The following terms are used throughout this paper:

Background redundancy: Occurs when the subject does not occupy the entire field of view, so that individual images (frames) of a clip contain large areas of background. This is especially frequent in images of vermiform animals like nematodes.

Clip: a sequence of images recorded chronologically in one operation.

Codec: A compression/decompression algorithm. This allows reduction of the file size of the source clip through a range of algorithms, some of which will preserve image quality while others will reduce it. Some codecs decompress at the same speed as compression, while others take longer to compress than to decompress. Also, different movie player programs are equipped with different selections of codecs.

Duration redundancy: Occurs when some of the frames at the start and/or end of a source clip contain no focused parts of the nematode. In order to obtain a complete series of focusing levels, it is best to start recording a fraction of a second earlier than strictly required, and to stop recording a fraction of a second later. In that case, the clip will definitely include the outermost surfaces of the specimen, but the first and last frames of the source clip will be redundant.

Finished clip: A clip edited, modified and/or compressed in various ways to optimize information content and minimize file size.

Frame: One individual image in the chronologically contiguous series of images that constitutes a clip.

Frame redundancy: Occurs when some frames in a source clip are effectively indistinguishable. While it is an essential advantage of video capture that many levels of focus are recorded in one operation, this does not mean that all levels of focus can and should be captured. Usually, only about 25-30 levels are significantly different ÷ as is also the case with 4-D microscopy - and in that case the remainder of the captured focusing levels are basically redundant.

Pixel redundancy: Occurs when the data rate of video capture exceeds the value required for maximal discernible quality of the final images. In other words, redundant information is stored because more bytes are assigned to each pixel than can be depicted on the end user's computer display.

Redundancy: The recording and storage of more information than effectively needed. Five kinds are distinguished here: background, duration, frame, pixel and resolution redundancy.

Resolution redundancy: Occurs when the resolution of video capture exceeds the maximal value imposed by limits of clip file size. In other words, redundant information is generated because more pixels are stored per frame than can be accommodated within file size limits set by data storage devices (e.g. 1.4 Mb maximal capacity for floppy disks) or transmission methods (e.g. 1 Mb maximal size for email attachments sent through certain service providers).

Scrubbing: Playback of a clip (forwards or in reverse) by clicking and dragging a slider at the bottom of a video display window.

Sequence: A series of clips edited and arranged in sequential order so as to obtain a finished "video program".

Source clip: An unedited clip as obtained immediately after capture, prior to any editing or modification.

Experimentation with the two hardware and software configurations led us to a series of general steps for obtaining edited video clips (Fig. 2, Table 3,4). More details on various aspects of these protocols are available at http://faculty.ucr.edu/~pdeley/vce.html.

- A first major stage consist of the capturing of multifocal series of images, which actually involves two separate steps. First, the equipment and software must be initialized in such a way that microscope images will be transmitted correctly to the capture card, converted by the capture software into acceptable real-time video images displayed on the computer monitor, and stored in the proper folders of the computer's hard disk(s) during actual capture.

The next step is the capturing itself, of properly positioned parts of a single specimen at a time. Capturing and specimen positioning can be repeated as many times as necessary to record an entire body part at highest magnification (e.g. pharynx, genital system, male tail region, ·), or even to record a complete specimen.

- The second stage consists of editing and compression of the captured video clips. Although each source clip contains all the morphological information provided by the combination of specimen, microscope and camera, it often uses much more disk space and memory than strictly necessary, because usually at least one of five kinds of redundancy (see glossary) occurs. Also, the clips can still be compressed by various algorithms, without visible loss of image quality or extension of playback speed.

Pixel redundancy is actually minimised before capture by selecting the appropriate software settings in the initialization step. The four other kinds of data redundancy must be minimised or removed after capture, using video editing software included with the video capture card. The clip is therefore edited more or less extensively to optimize its final size. As an optional feature it is also possible to add text, scale bars and arrows to some or all of its frames, so as to highlight e.g. structures that are particularly important or hard to see.

- In the last stage, the edited clip is tested and verified on various video player programs, to ensure that the end users will be able to play it correctly, regardless of their particular hardware and software facilities.

Optionally, if complete specimens were captured as a large series of clips, then it is also possible to index the clips by means of an overview of their position along the body of the specimen. We have done this by constructing "image maps" in World Wide Web pages written in HTML and accessible online or on disk through any WWW browser. On such an image map, a specimen is depicted as a low-magnification image with clickable boxes or bars delineating the different body parts captured at different magnifications, and providing embedded HTML links to the corresponding video clips. When a given box or bar is clicked, the appropriate clip will open automatically in the video player software installed as plug-in of the browser used. We refer to the aforementioned website for more details and examples.

After familiarizing ourselves with the hardware and software, both our VCE systems proved to be versatile and easy to use. In terms of time needed for capture and editing, the entire procedure took between five minutes and a whole day, depending mainly on the size of the specimen, the concomitant number of files captured, and the addition of text and labels of varying complexity (see examples in Table 5).

Nevertheless, some limitations were evident, both in the optical and electronic components. Variability in the smoothness and speed of manual focusing meant that some parts of a specimen were sometimes captured with less multifocal resolution than others, or that different operators would consistently capture with greater or lesser duration redundancy. These problems could probably be overcome with a motorized focusing system, although this would represent a significant extra cost.

With the combination of objectives and camera adapter assembly of Configuration 1, small nematodes (body length below 2 mm) were easily recorded, but for larger nematodes or dissected organs it was more difficult to obtain a good fit and positioning in the video display area. The entire nematode could not always be covered by one frame at 4x magnification, while video capture of elongate structures (such as a long tail, a cylindrical bulb or an extended dissected gonad) was less convenient, because the 40x objective did not always provide enough resolution, and the display area obtained with 100x required too many separate files to be recorded.

If at least the width of the studied object was completely contained within the display area, then it was fairly easy to capture it as a series of separate files, and distill a reconstruction of the entire object - much as if one were to directly observe it in several steps with the LM. However, when neither the width nor the length fit within the display area, e.g. when recording larger dissected gonads from Meloidogyne spp., then it became too difficult to compile a set of files and to use these subsequently for re-assembling an overview. In that case, we were forced to compromise on resolution by choosing a lower magnification.

Such limitations were largely overcome, albeit at additional expense, by the various components of Configuration 2 providing greater flexibility in magnification and viewing area. In practice, the highest useful setting of the magnification changer was 1.6x, as the 2x setting only represented "empty magnification" without any gain in resolution on the video monitor or computer screen.

The resolution provided by both the tested video cameras was sufficient for many purposes, but nevertheless clearly limited in comparison to the actual resolution of the microscope optics. In particular, VCE images of fixed specimens did not always resolve very small or thin features such as cuticular punctations, nerve endings or minute openings, even when these were relatively distinctly visible with the LM itself. Similarly, in dissections of gonads the precise appearance of cell membranes was often resolved much more poorly in the video clips than through the eyepieces of the LM. Improvements in this respect must await the advent of digital video cameras with both high resolution and appropriate adapters for fitting to LM.

This being said, the obtained resolution was definitely good enough for most illustrative purposes. When assembled into a figure and printed next to photographic plates, for instance, VCE frames were somewhat poorer than still photographs in contrast, but slightly better in resolution, less grainy, and the contrast was easily enhanced with image editing software (Fig. 3, 4). Again, the magnification-changing add-ins of Configuration 2 were less affected by resolution issues, although at reduced magnifications using the 60x objective and the 0.35x camera adapter, there was an even greater discrepancy between the actual quality of the microscope optics and the obtained final on-screen resolution.

Configuration 1 was installed on an ordinary benchtop in a room below three floor centrifuges, while configuration 2 was set up in a laboratory with various kinds of large and noisy equipment (fume hood, flow bench, incubators, etc.). The obtained images were in both cases affected by microvibrations, which were reduced to a minimum by mounting the microscopes on a heavy board resting on various kinds of elastic material (bubble wrap, packing foam, or a rectangular lifesaver). This largely took care of the problem, to the extent that video clips were no longer significantly affected, even when there were still sufficient vibrations to prevent traditional photography.

On the computer side, the DC30plus video capture card is known to be prone to overheating, in which case it produces clips containing frames with green or red bands and desynchronised interlacing, resulting in playback problems. This problem occurred with Configuration 1, and was minimised by equipping the computer housing with better two ventilation fans instead of the usual single one. Also, these and other occasional errors in single frames could be edited out of the captured clip, by removing the affected frame in the Monitor window of Adobe Premiere 5.1. The DV500 card in Configuration 2 did not present such problems.

We encountered an intriguing bug with Configuration 1 in Adobe Premiere 5.1, in the "Special Processing" window of the "Export Movie" settings. At 720 x 540 frame size, the cropping rectangle control positions were not properly converted into coordinates, the latter values only being half of the correct ones (e.g. if the rectangle controls are set at 70 pixels from left, 100 pixels from right, 50 pixels from top and 124 pixels from bottom, then the coordinate boxes display the respective values 35, 50, 25 and 62). Apparently the program incorrectly assumed that frame size was 360 x 270, regardless of the actual value in "video settings".

Without compensation, this resulted in less extensive cropping of the frames, and therefore suboptimal reduction of background redundancy. Fortunately, in most cases the error could easily be compensated for by always selecting the proper rectangle first, and then typing the double of each value in each coordinate box (but never the other way round). However, if the corrected coordinate values were such that they would add up to more than 360 pixels horizontally or 270 pixels vertically, then a conflict arose and it was not possible to crop the desired area completely.

With Configuration 2, Adobe Premiere 5.1 invariably caused a complete freeze of the operating system. However, this problem disappeared after upgrading to Premiere 6.0. The latter version has a number of minor bugs affecting e.g. windows refreshing or available menu options, the most notable one being that the function for capturing still images instead of movie clips did not work, despite being listed in the appropriate menu and in the manual. This meant that still images had to be obtained instead by exporting them out of captured clips. Otherwise, none of the observed bugs significantly affected the procedure summarized in Table 4. Finally, we experienced one major crash of the removable hard drive when it was moved from the computer of Configuration 2 to another system. We therefore adopted the practice of copying sets of newly captured or edited files to CD-R immediately after capture or editing.

Despite the limited resolution and some technical flaws, the tested configurations were easy to use and yielded finished clips that were of acceptable image quality and file size. Compared to film-based still photography, VCE was in all respects faster, more versatile and more forgiving. Firstly, no film development was required, so the captured clip could be checked instantaneously for flaws, and retaken if necessary. Secondly, while it was very easy to obtain still images from a clip, it was rather difficult and time-consuming to photograph a continuous series of focusing levels ÷ all the more so because it was not possible to check the result on the spot. Thirdly, the video cameras had a much broader tolerance for low light conditions and/or slight movements of the specimen. Although both situations did affect image quality, fuzziness or graininess of the image became a problem much less quickly than with still photography on film. Thus, it was quite difficult to obtain a good series of still photographs of a freshly dissected gonad, because the temporary mount still allowed for some movement of the specimen due to muscular contractions and plasmolysis. By comparison, the VCE system immediately yielded a good series of images of the same specimen (Fig. 4).

Digital still photography has many advantages over its film-based counterpart, in that it matches or even exceeds the level of resolution, while bypassing the development step. Digital still cameras are also rapidly becoming more affordable and more compatible with microcomputers, and more advanced models are increasingly combining still photography with video functions (albeit at reduced resolution and frame rate).

Compared to digital stills, analog video capture is limited in terms of resolution attainable with existing cameras, but on the other hand it provides a distinct advantage with respect to multifocal recording. Not only does this allow capture of a series of focal planes in one quick operation, but it also results in files that retain this focal series as a single object, without necessarily using more disk space than a single high-resolution digital still image. VCE and digital still photography are therefore complementary rather than competing approaches: the latter is to be preferred when resolution is critical, while the former is more appropriate for reproducing the focusing action of a microscope, thereby translating three-dimensional topologies into a series of two-dimensional optical sections.

Undoubtedly, the distinction between digital stills and analog video clips will disappear quickly if technological developments continue along current lines. Digital video cameras are now becoming available, as well as video capture cards that accept digital input (such as the tested DV500 capture card). Within the next few years, we can therefore expect prices to drop and equipment to develop sufficiently rapidly to allow the use of completely digital VCE configurations. In terms of computer hardware, the process will become more and more efficient with the continued development of larger computer memory chips, larger hard disks, faster processors, faster internet connections, more efficient codecs and more powerful video editors.

Another relevant technological development, is image capture and digital enhancement resulting in still images with Extended Depth of Field (EDF, see Tucker et al., 1999), i.e. a much greater range of focus than allowed for by standard optics. This approach holds great promise for the production of single still images containing contrast-rich and focused structures occurring at different levels, but as far as known to us, inexpensive systems are not yet available on the market, and we are not aware of any examples where EDF has already been applied to nematodes or other transparent invertebrates. One problem could be that heavily sclerotized structures like buccal capsules and stylets will blot out over- or underlying structures. Also, a VCE clip preserves structural clues to angle of view and position of a specimen, whereas an EDF image combines information from multiple standard focal planes into a single omnifocal plane. In the case of nematodes, this could mean that a clear single EDF image could be obtained of a nematode lip region in sublateral view, with both amphids in focus, but without any remaining clue allowing distinction between a dorso-sublateral and a ventro-lateral angle of view. Practical application to nematode specimens will show whether these concerns are important or not, and to what extent EDF can be combined with VCE and other types of microscopy imaging.

A last useful comparison that needs to be be made, concerns the performance and uses of VCE as compared to Confocal Laser Scanning Microscopy (CLSM, see e.g. Czymmek et al., 1994; Donaldson & Lausberg, 1998). The latter approach has many advantages, in that it e.g. significantly enhances resolution, as well as relying on software capable of measuring distances and constructing three-dimensional models. We have not yet tested our configurations in the context of automated biometric analyses, as the emphasis of the approach outlined here is on the recording and representation of morphology itself, rather than on subsequent biomathematical applications. Nevertheless, it is presumably possible to expand our hardware and software to incorporate three-dimensional imaging functions (see e.g. Omasa & Kouda, 1998).

As a trade-off for the benefits, it should also be noted that CLSM systems are significantly more expensive, and alter the properties of the microscope itself in ways that make its use rather different from routine practice in nematology, so that for most nematologists it is not a realistic option to convert their research microscope into a CLSM system. By comparison, adding the various components of a VCE system can be done at a much lower cost and with greater ease, without modifying the basic properties of the microscope itself.

Based on the developed procedures, we have tested the quality (Fig. 3-5), resulting file size and time load (Table 5) of producing VCE files for a number of different purposes. Examples can be seen at http://faculty.ucr.edu/~pdeley/vce.html

a) Recording for websites: VCE files can easily be incorporated into web pages, allowing internet users around the world to view the captured part of a specimen as if it were positioned under a microscope. However, current limits to download speeds make it impractical to have to access a file that is several Mb large, especially if one is connected by means of a modem rather than an Ethernet or Local Area Network connection. For basic website purposes, it is therefore better to produce smaller files by compromising somewhat on quality and/or using more efficient codecs such as "Sorenson". Web-connected users who do not have recent versions of video playing software, will in most cases be able to update their software online, and obtain newer versions that include the newer codecs.

First, using only half of the available resolution (360 x 270) but the high data transfer rate of 4000 kbps, we produced very modestly sized files after compression with the "Cinepak" codec. (see Radopholus bridgei in Table 5). Although clearly inferior to direct observation with LM, these low-resolution files nevertheless contained sufficient detail to capture and label many important characters, illustrating that a VCE file can offset lower frame resolution to an important degree by its multifocal information contents, compared to a single still image captured at high resolution. Next, we tried a limited data rate during capture, but this resulted in negligible differences in file size of the final clip (see Prionchulus punctatum, with and without limited data rate during capture,Table 5). In order to produce acceptable clips for websites, it is therefore sufficient to limit resolution during clip exporting.

b) Recording demonstration material for disk access: If the purpose is to provide students and inexperienced users with a detailed overview of important mophological structures, then it is sufficient to capture only those parts of a specimen that are relevant, such as the lip region, all or part of the pharynx, all or part of the repoductive system, and the tail. The files can be opened and read from CD-ROM, or preferably copied to and opened from hard disk for maximal playback speed, without the constraints imposed by download speed. Each multifocal series can therefore be captured at maximal resolution (in our case, 720 x 540 pixels) and with high data transfer rate.

Capturing the various body regions takes very little time in itself, but for maximum transfer of knowledge it is best to use the titling facilities of the video editor to their fullest, by naming all informative or interesting structures, and adding arrows to point them out in those frames where they occur only (Fig. 5). This titling step is much more time-consuming (Table 5), especially during the first few attempts to utilise the various functions. However, in our experience the result is an extremely effective demonstration tool, especially when compared to traditional line drawings, or having to point out structures through each student's LM in a full class. Not only does it allow the viewer to learn to locate and distinguish each structure, but it also greatly improves the three-dimensional perception of inexperienced microscopists, by mimicking the focusing action of the microscope, and encouraging the use of focusing as a tool for assessing the relative positions of different structures at different transverse levels in a transparent specimen.

c) Archiving large numbers of specimens and slides: Many databasing and surveying applications benefit from an ability to link one or more visual records to a text record of each collected specimen. VCE clips can easily be linked to various popular database programs and used for such purposes.

In the case of nematodes, it is often useful to verify the appearance and identification of specimens by taking a quick look at its lip region, cardia, vulval or spicular region, and tail. VCE protocols can be used to routinely capture and optimize these five body parts for large numbers of nematode specimens, and the resulting VCE clips can then be included in, or linked to, a database. Overall body shape and size can be captured on a single still image at appropriate magnification (4x or 10x objective).

Depending on space constraints, it may be advisable to restrict file size of archive clips in the same ways as for website files, and "Sorenson" compression would then be preferable over "Cinepak" if the database will be used strictly by local access. However, we assume that this problem will largely disappear with the ever-increasing capacity and speed of hard disks, CD-writers and DVD-writers, as the download speeds allowed on a Local Area Network are usually sufficient to handle quite large VCE files.

To test feasibility of routine screening and capturing of specimens, we have therefore processed a series of 10 slides by capturing at 720 x 540 pixels and 4000 kbps, and timed the average duration per specimen at 22 minutes (Table 5). No titling or scale bars were added, as we assume that routine application would not allow for the added time required per specimen, and in that case it is essential to document each clip by including all relevant information in its file name, or in a separate text file or database record.

Furthermore, the 100x water immersion objective of Configuration 2 allows screening and capture of nematodes immediately after extraction, without any fixation or the need to place a coverslip over the specimens. We have performed trial runs of sample screening, by heat-killing extracted nematodes on a hot plate at 65° C, and then applying VCE to specimens in an open drop of water on glass slides. Although the 100x objective is inferior in resolution to its oil immersion counterparts, it nevertheless provides images of sufficient detail to allow rapid screening and archiving of nematodes, without requiring the time investment and health risks of fixation and transfer to permanent slides.

d) Recording type material as "image vouchers": In this case the goal is to store an entire specimen digitally at maximal resolution and at different magnifications, in order to produce a set of reference files that can replace the type specimen for most purposes of morphological and taxonomic studies. We therefore set out to determine the amounts of time and disk space required for creating a "virtual type specimen".

For smaller nematodes like Hemicriconemoides variodus, Panagrobelus stammeri, or Plectonchus n. sp., the required number of captured clips was fairly limited (up to about 30 files, cf. Table 5), and it therefore took little time to give all edited files a name including basic information such as taxon name, body part, slide number and magnification. Also, in order to allow subsequent on-screen measurements, we first recorded the scale on a calibration slide at each magnification, as clips or TIFF files. For H. variodus and Plectonchus n. sp., we then used the titling functions to construct a scale bar for each magnification, and paste it into the appropriate clips with captured parts of the specimen. We also titled each clip with taxon name and slide number.

As appears from Table 5, total processing time for H. variodus and Plectonchus n.sp. was moderately to significantly longer than that of Panagrobelus stammeri, where we did not apply titling. In practice, adding titles may require too much time when large numbers of type material need to be processed, and in this case it may suffice to give each clip an informative name. Nevertheless, to eliminate any confusion we would nevertheless recommend that titling always be applied to at least the holotype.

In larger nematodes like Paraxiphidorus michelluci, the number of files required for complete capture becomes much too large to be feasible without automation of the focusing and stage controls. It even becomes too time-consuming to use detailed filenames, and in this case we therefore just named the files as a numbered series for each magnification. In such cases, an acceptable compromise is to capture an incomplete series of multifocal files, by omitting only those parts of the body that are irrelevant to diagnosis. Thus, we did not record those parts of the P. michelluci holotype where the body only contained intestine, and this yielded file numbers, time and disk requirements that proved more manageable.

In order to keep track of all recorded parts of a completely or incompletely captured specimen, we constructed clickable image maps to show the magnification and display area of each VCE file. Examples are shown at http://faculty.ucr.edu/~pdeley/vce.html, for the captured holotypes of Paraxiphidorus michelluci and Hemicriconemoides variodus. However, the construction of this type of map requires several hours more work, and we would therefore not recommend it for routine VCE processing of type material.

e) Recording and documenting parts of dissected organs: It is impossible to keep permanent mounts of dissected organs without drastic reduction of quality. Furthermore, when one is interested in the cellular architecture of these organs, then a clear three-dimensional overview is indispensable. By preserving the multifocal nature of observations through the microscope, our approach provides a relatively cheap and fast tool for fulfilling both these demands, even in the absence of dedicated 3-D software.

The most informative components were recorded of several female nematode gonads , i.e. the region around the spermatheca: end of the ovary, oviduct, spermatheca and beginning of uterus (see Fig. 4 for an example). In most cases, the region of interest could be captured within the display area, and the production of a single videoclip was sufficient to record the dissected structures. In those cases, capturing and editing times were minimal: per gonad, we timed an average duration of only 4 minutes without titles and 45 minutes with titles (Table 5).

f) Recording the staining of specific cells: In certain types of histological or molecular studies, results are visualized by means of staining reactions in target cells or structures. In many cases, this stain is not permanent, fading with prolonged exposure to light, due to diffusion, and/or because of decomposition of the tissues. Capturing the staining pattern with a VCE system allows the observer to permanently record which cells or structures are stained. The basic procedure is the same as for slide archiving (item b above). Obviously, a confocal microscopy system will provide more detailed recording and resolution for this type of application. However, the lower cost of adding VCE components to an existing microscope, or even of buying a complete VCE system, will undoubtedly make it an affordable alternative for many laboratories unable to budget a confocal system.

We tested a slightly modified regressive staining technique (essentially carmine and propionic acid) based on Khrustalev & Hoberg (1996). Stained Panagrolaimus rigidus specimens were mounted directly in the destaining solution (50 % acetic acid and 70% ethanol), and we then captured a section of the intestine at several time intervals during the destaining process. Afterwards, we could easily select the clip with the best contrast between nuclei and background, thus recording the cellular structure of the intestine on a whole mount individual. The effective time spent on recording and editing of these clips was minimal (see Table 5).

Video Capture and Editing allows for the production of video clips that mimic the act of observing nematode specimens through a light microscope. These clips can be modified and distributed in a number of ways, allowing for a wide range of applications at the interface between preservation, illustration and distribution of actual nematode specimens. The equipment tested here, apart from the compound microscope itself, is comparatively cheap and easily interchangeable with different microscope brands and models, without requiring computer and microscopy expertise beyond those skills needed routinely in a laboratory environment.

Although the optical quality of the obtained final images is inferior to that of the microscope optics alone, it is nevertheless sufficient to supplement or even replace direct observations through the microscope, in a wide range of situations. Also, even though the obtained file sizes and types are too demanding for older microcomputer systems, they are well within the operational limits of speed and capacity of the current generations of computers, and future developments will undoubtedly allow for dramatic further increases in efficiency.

We therefore predict that this methodology will quickly become widespread in nematology laboratories, and that many further uses will be developed in the process. Other applications that we wish to investigate ourselves include capture of images from immobilized live nematodes, as a means of combining high-magnification microscopy with molecular analyses of individual specimens. In the longer term, we hope that it may even become possible to altogether omit the need for production of permanent slides, with its concomitant chemical hazards of handling and storing carcinogenic and toxic fixatives. Keeping these possibilities in mind, we are continuing to experiment with other accessories that may further expand the versatility of VCE.

Czymmek, K.J., Whallon, J.H., and Klomparens, K.L. 1994. Confocal microscopy in mycological research. Experimental Mycology 18: 275-293.

Donaldson, L.A., and Lausberg, M.J.F. 1998. Comparison of conventional transmitted light and confocal microscopy for measuring wood cell dimensions by image analysis. IAWA Journal 19: 321-336.

Eisenback, J.D. 1988. Multiple focus and exposure photomicroscopy of nematodes for increased depth of field. Journal of Nematology 20: 333-334.

Khrustalev, V. A. & E. P. Hoberg. 1996. Carmine-propionic acid stain for elucidating of fine cellular structure in nematodes. Journal of Parasitology 82: 176-178.

Omasa, K., and Kouda, M. 1998. 3-D color video microscopy of intact plants: a new method for measuring shape and growth. Environment Control in Biology 36: 217-226.

Schnabel, R., H. Hutter, D. Moerman, and H. Schnabel. 1997. Assessing normal embryogenesis in Caenorhabditis elegans using a 4D microscope: Variability of development and regional specification. Developmental Biology 184: 234-265.

Thomas, C., P. De Vries, J. Hardin, and J. White. 1996. Four-dimensional imaging: Computer visualization of 3D movements in living specimens. Science 273: 603-607.

Thomas, W.K., J.T. Vida, L.M. Frisse, M. Mundo, and J.G. Baldwin. 1997. DNA sequences from formalin-fixed nematodes: integrating molecular and morphological approaches to taxonomy. Journal of Nematology 29: 250-254.

Tucker, S.C., W.T. Cathey, and E.R. Dowski, Jr. 1999. Extended depth of field and aberration control for inexpensive digital microscope systems. Optics Express 4: 467-474.
 

 
Legends for figures

Fig. 1. Photograph of the main components of VCE Configuration 1. The video camera is connected to the computer by means of an S-VHS cable plugging into the external connector box or "breakout box" of the video capture card.

Fig. 2. Overview of the main steps in the VCE procedure (on left), and where these steps are implemented in the relevant software of configurations 1 and 2 (on right).

Fig. 3. Comparison of image quality obtained by VCE Configuration 1 (left column) versus 35 mm still photography through the same optics (right column) of different focusing levels of the lip region of a formalin-fixed female Panagrobelus stammeri (cf. Stock et al., submitted) mounted in glycerin. Note the occurrence of setiform processes on the lips (arrows), which are resolved slightly better by VCE than 35 mm film.

Fig. 4. Comparison of image quality obtained by VCE Configuration 1 (left column) versus 35 mm still photography through the same optics (right column) of the freshly dissected spermatheca of Coslenchus sp. Note cell membranes (arrows) resolved better by VCE than 35 mm film.

Fig. 5. Several frames from an annotated VCE clip of a female paratype of Radopholus bridgei, produced with VCE Configuration 1. Text and arrows (in color in the original clip) are embedded in one frame or in a series of subsequent frames, as appropriate for the various structures visible at the corresponding levels of focus. Note the small projections just anterior to the stylet knobs, which were not described before for this species, and which are nevertheless resolved by the video camera.