Palaeontologia Electronica http://palaeo-electronica.org

IMPROVING DEPTH OF FIELD RESOLUTION FORPALYNOLOGICAL PHOTOMICROGRAPHY

Antoine Bercovici, Alan Hadley, and Uxue Villanueva-Amadoz

ABSTRACT

Optical microscopy continues to be the preferred method for imaging in paleopa-lynology. While usefulness of other tools, such as the scanning electron microscope, isnot questioned, the ease of use and timely results of optical microscopy remainsunsurpassed. However, obtaining good quality photomicrographs requires the use ofthe highest magnifying power objectives available, which are inevitably associated withvery limited depth of field. To avoid the need for multiple photomicrographs in order tofully describe each palynomorph, a software solution for reconstructing depth of field isproposed. This solution allows for keeping the main advantages of high magnifyingpower objectives (better resolution and improved contrast) while suppressing theirmain weakness. In addition, photomicrographs published using depth of field recon-struction have a more natural appearance, similar to when directly viewed with the eyeunder the microscope. While this paper deals primarily with the usage of depth of fieldreconstruction for the enhancement of palynological photomicrograph, the techniquecan be applied similarly to many other paleontological and geological objects as well.

Antoine Bercovici. UMR 6118 du CNRS, Géosciences Rennes, Bat. 15 – Université de Rennes 1, Campus de Beaulieu, 35042 Rennes Cedex, France. antoine.bercovici@univ-rennes1.fr Alan Hadley. 5 Ronald Road, Darnall, Sheffield, United Kingdom alan@micropics.org.ukUxue Villanueva-Amadoz. Departamento Ciencias de la Tierra (Paleontología), Universidad de Zaragoza. C/Pedro Cerbuna, 12. 50009 Zaragoza, Spain uxuevillanueva@yahoo.es

KEY WORDS: Image processing; Optical microscopy; Focus stacking; Palynology; Depth of field recon-struction

INTRODUCTION

Optical microscopy remains the preferredmethod for imaging in paleopalynology. The studyof palynomorphs, for both taxonomic descriptionand quantitative data extraction from counts,strongly relies on the observation of palynologicalassemblages mounted on microscope slides usingtraditional techniques of preparation (Doher 1980;Wood et al. 1996; Traverse 2007). With the

improvement of optical microscopy over time,imaging of details on transparent samples (whichuntil then had often produced poor quality results)was made possible by optical means, without theuse of chemical/dying techniques (Pluta 1989;Slayter and Slayter 1992; Davidson and Abramow-itz 2002), which alter the subject in some way.These innovations include various contrastenhancing techniques (Hoffman 1977; Abramow-

PE Article Number: 12.2.5TCopyright: Palaeontological Association August 2009Submission: 18 September 2008. Acceptance: 14 May 2009

Bercovici, Antoine, Hadley, Alan, and Villanueva-Amadoz, Uxue, 2009. Improving Depth of Field Resolution for Palynological Photomicrography. Palaeontologia Electronica Vol. 12, Issue 2; 5T: 12p; http://palaeo-electronica.org/2009_2/170/index.html

Bercovici, Hadley, & Villanueva-Amadoz: Improving Depth of Field

itz 1987; Bradbury and Evennett 1996; Davidsonand Abramowitz 2002) and more especially differ-ential interference contrast (DIC) microscopy,invented in the mid 1950s (Allen et al. 1969;Nomarski 1955). For palynology, DIC allows theobservation of very minute ornamentation of theexine not visible under regular brightfield micros-copy.

However, optical microscopy suffers from res-olution limitations. Horizontal resolution limitation(the resolving power between two points occurringin the plane perpendicular to the optical axis) isoften given and is an easily understood parameteras it directly translates into maximum attainablemagnification of an object. Additionally, optical sys-tems are also characterized by their ability toresolve along the optical axis, which is termed axialresolution. The second resolution, measured in aplane parallel to the optical axis, is known as thedepth of field (Pluta 1989; Slayter and Slayter1992; Davidson and Abramowitz 2002). Practically,depth of field represents the distance that sepa-rates the nearest object plane in focus to the far-thest object plane which is simultaneously in focus(Davidson and Abramowitz 2002). In microscopy,this distance is very short and typically measuredin microns. This very limited depth of field does notpresent an issue for flat objects such as thin sec-tion, but objects preserved in three dimensions(such as palynomorphs), with a greater thicknessthan the depth of field for the objective. In suchcases it is only possible to see a single opticalcross section of the object at a time, and theentirety of the object can only be reconstructed asa mental image by constantly varying the focusadjustment of the microscope (Figure 1).

Both axial and horizontal resolutions aredriven by the numerical aperture of the objective,but in different ways. As the axial resolutiondecreases, the horizontal resolution increases withthe numerical aperture. Therefore, objectives withhigher numerical apertures give more contrast andhigher magnification but a lower depth of field, thusit is necessary to select an appropriate trade-offbetween these quantities. Traditionally this trade-off is solved in photographic descriptions of palyno-morphs by including a general view of the speci-men taken at a lower magnification to avoid depthof field problems, and a series of pictures at highermagnification to highlight details, or to expose dif-ferent views of the same specimen as separateoptical sections. While the inclusion of multiple pic-tures for a single palynomorph is necessary andcommonly used for taxonomically oriented publica-tions (such as for description of new species), it israrely done in publications involving description ofpalynological assemblages for biostratigraphy.Theamount of space required for depicting detailedpalynological plates can be quite high for largeassemblages, and multiple views of a singlepalynomorph for identification is not practical. As aresult, palynomorphs are depicted as a singlelower resolution photograph that may not show allthe necessary details for easy identification andverification.

To circumvent this problem, dedicated soft-ware solutions may be employed to reconstruct asingle image from multiple optical sections, eachcontaining only parts of the object in focus. Thistechnique allows obtaining the advantage of highermagnification and contrast levels given by the high-

21

3456

Light source

Mounting mediumCover slide

Immersion oil

Objective

Microscope slide

Level of focus

FIGURE 1. Residue resulting from the standard processing of rock samples is mounted on a microscope slide usinga mounting medium of appropriate refractive index. The slide is observed using transmitted light microscopy (withDIC if available), using the more powerful objective with immersion oil. Varying the focus adjustment of the micro-scope allows one to cross successive optical sections (planes 1 to 6 are examples) across the palynomorph.

2

PALAEO-ELECTRONICA.ORG

est magnification objectives, whilst eliminating thelimited depth of field issue.

RECONSTRUCTING DEPTH OF FIELD BY SOFTWARE PROCESSING

The solution relies on a software algorithmcapable of reconstructing a full image based on astack of multiple individual pictures. There are twopopular image processing techniques based onimage stacks: the first is called image stacking andconsists of using multiple image of an object takenin the same conditions. Image stacking is effec-tively working on limitations due to the imaging sys-tem (and especially digital camera sensors) bycombining all of the separate images into a singleimage with reduced noise, defects, and increasedresolution. This reconstruction technique, verypopular for astronomical photography, should notbe confused with the second image processingtechnique called focus stacking. Focus stacking

uses a stack of images taken in different focus con-ditions in order to reconstruct a complete image ofthe object.

Both commercial and free open sourced soft-ware incorporate or were specifically developed toapply image stacking and/or focus stacking on aset of images. A summary of commonly availablesoftware is listed in Table 1 along with the operat-ing system on which it can be installed, as well asthe type of license for its usage (free or commer-cial). For the purpose of this paper, focus stackingwas tested on each of the listed software to com-pare the quality of the end result image. However,keep in mind that the majority of this software isdesigned for use by photographers using the“macro” mode of their camera. While this is com-patible for use with opaque paleontological andgeological objects, translucent palynomorphs maybe only partially reconstructed as two overlappingareas (e.g., the proximal pole and the distal pole of

1 2

FIGURE 2. Digital image of a checkerboard pattern 1: in focus, 2: out of focus.

3

Software Author Licence type Usage Linux/Unix MacOS Windows Image Stacking Focus Stacking

ALE

AutoPano Pro

CombineZM

DeepSkyStacker

DpxView Pro AZ

Extended Depth of Field (ImageJ plugin)

Helicon Focus

Image Pro Express

Image Pro Plus

ImageJ

PhotoAcute

Photoshop CS3 Extended

Photoshop CS4 Extended

Picolay

RegiStax

SAR Image Processor

Stack Focuser (ImageJ plugin)

David Hilvert Open source, GPL Free Yes Yes Yes Yes

Kolor Proprietary Commercial Yes Yes Yes Yes

Alan Hadley Open source, GPL Free Yes Yes

Luc Coiffier Freeware Free Yes Yes

DeltaPix Proprietary Commercial Yes Yes

Alex Prudencio Proprietary Free for research use Yes (Java) Yes (Java) Yes (Java) Yes

Danylo Kozub Proprietary Commercial Yes Yes Yes

Media Cybernetics Proprietary Commercial Yes Yes Yes

Media Cybernetics Proprietary Commercial Yes Yes Yes

Wayne Rasband Open source, GPL Free Yes (Java) Yes (Java) Yes (Java) Yes

Almalence Inc. Proprietary Commercial Yes Yes Yes Yes

Adobe Inc. Proprietary Commercial Yes Yes Yes

Adobe Inc. Proprietary Commercial Yes Yes Yes Yes

Heribert Cypionka Freeware Free Yes Yes Yes

Cor Berrevoets Freeware Free Yes Yes

Saruzinsky Shareware Commercial Yes Yes Yes

Michael Umorin Open source, GPL Free Yes (Java) Yes (Java) Yes (Java) Yes

TABLE 1. List of available software with image stacking and/or focus stacking capabilities.

Bercovici, Hadley, & Villanueva-Amadoz: Improving Depth of Field

a spore) can appear in focus at a different level.For this reason, among all of the available softwarelisted in Table 1, only CombineZM (Hadley 2006)will be discussed herein. This particular softwarewas adapted to work on translucent objects andgive significantly better results for palynologicaluse.

The algorithm behind the “magic”

Several methods can be adopted for theimage processing required to complete depth offield reconstruction. The method used by Combi-neZM consists of identifying zones within eachframe that appear to be in focus. These zones aresubsequently merged while out of focus areas areignored. Identification of zones in focus is deter-mined by using the amount of details visible in aframe, such as edges and dots: when parts of theobject are out of focus, details will appear blurredor will not be seen at all. Details on a picture areusually recognizable as zones with well-definededges and strong contrast, a characteristic thattends to disappear if out of focus. In a digitizedimage, this translates into the amount of variationthat can be observed among neighboring pixels. Toillustrate this, Figure 2 shows a checkerboard pat-tern imaged in the focus plane (1) and out of focus(2). Indication that the checkerboard pattern A isactually in focus is identified by the well-definededges of each black and white square (represent-ing a maximum variation of 100% (white) to 0%(black) on the edge of each squares). The checker-board pattern in B depicts a much more gradualvariation in the value of each adjacent pixel’sbrightness. The algorithm used in CombineZMuses this principle and quantifies the variations inthe values of adjacent pixels, then sums thesequantities thus giving each pixel a score. The areasaround the pixels with the highest scores are thenmerged to form the composite image.

Method of reconstruction

CombineZM provides a set of two predefinedroutines (macros), each using a different methodfor reconstructing depth of field. These methodscan be customized by editing the macros, (e.g. thesharpness and/or contrast of the resultant imagescan be adjusted to attain better results dependingof the equipment used). Correct operation of theprogram requires the original pictures to be nor-malized; the reconstructed image consists of anoverlay of pieces of the individual original frames,all of the individual frames must be normalized tothe same value of contrast, brightness and colour.

Also, the object of interest must be scaled at thesame size ratio and the same position on eachframe in order to be rendered correctly. Combi-neZM can perform all of these adjustments auto-matically using the top frame of the stack as thereference for adjusting all other pictures. Thus it isnecessary to ensure that the brightness/contrastratio as well as the colour balance is adequate onthis first frame. Once normalized, two processingmacros are proposed:

- Do stack

This macro selects areas that are identified tobe in focus and uses them to reconstruct the objectby stacking them on top of each other. Adjacentpatches from different frames are merged into acomposite image. However, this macro does notmerge details that reside on different frames at thesame location. Due to this limitation, details on thefront, within, and on the back of transparent objectsmay not appear in the final reconstructed image.

- Do weighted average

This macro uses the weighted average of cor-responding pixels in the stack. The weighting factorused is the score produced by summing the differ-ence between each pixel and its neighbours on thesame frame. As translucent objects such aspalynomorphs create a challenge for the “DoStack” macro above (having parts of the object thatcan overlap), this algorithm may be more suitablein dome cases. This macro overcomes the problemoutlined earlier, however, its drawback is thatbecause several frames may contribute to theimage at each location there may be some loss ofsharpness, some fogging, and the merging ofobjects that in reality are far apart along the opticalaxis.

Guidelines for photography

Regarding the operations conducted duringsoftware processing, a number of guidelines needto be followed during acquisition of the individualframes of the stack. Particular attention should bemade to use the same settings for image acquisi-tion on each frame. Even if the CombineZM soft-ware is able to correct disparities between frames,it is always preferable to keep them at a minimum.The camera should be set to a manual mode inwhich both white balance and exposure time canbe kept identical for all pictures. Additionally, atten-tion should be paid to avoiding possible motion ofthe subject or of the optical setup (especially if thecamera is mounted on the top of a long optical tube

4

PALAEO-ELECTRONICA.ORG

subject to vibration) between frames. Remote shut-ter release (electronic or cable) and mirror lockup(for single-lens reflex cameras [SLRs]) can help toprevent these vibrations.

The number of optical section picturesrequired for a particular stack is not critical, as longas at least each frame contains the focus intervalto be reconstructed in the stacked image. How-ever, there is no need to take pictures at closeintervals, or to take duplicated pictures, as they donot add extra data and may alter the final image.Ideally, pictures should be ordered starting at theuppermost frame (processed first by CombineZM),to the lowermost frame. With the advances of digi-tal photography, digital cameras, or even high-endSLRs, are commonly mounted directly on micro-scopes, and provide sensor resolutions in the tensof megapixels range. While creating large, ten-megapixel images may be useful for certain situa-tions, processing a series of such images requiresa tremendous amount of memory and processingpower on the host computer running CombineZM.Not only will the processing time exponentiallyincrease, but the quality of the output decreases ascamera resolution increases (close examination ofthese images usually reveals that edges anddetails are less sharp at the pixel level). It is pre-ferred to downsample such large images to a moreacceptable value of five, two, or even one mega-pixel. Lowering the resolution not only speeds upthe processing, but also artificially improves theidentification of details, as the reduced number ofpixel artificially increase the difference betweenneighboring pixels located at the edge of eachdetail. Remember that at a standard printing reso-lution of 300 dpi, a five megapixel picture of apalynomorph occupying the entire field of view ofthe camera represents an actual size of 22 x16.5 cm, which surpasses the size requirementsfor publication, so downsampling images shouldnot be a concern.

Batch processing

In addition to the CombineZM software, anextra plug-in called CZBatch was developed as away to automatically apply a predefined set ofoperations to a group of stacks. Each group offrames representing a stack needs to be collectedin an individual folder prior to running CZBatch.While the batch processing plug-in can dramati-cally speed up the processing of large numbers ofstacks of pictures, it is recommended to carefullycontrol frames that are chosen as input for eachstack.

Post-processing

Resulting stacked pictures can be post-pro-cessed prior to assembling a palynological plate.Because of the stacking technique, all areas of thepalynomorph now appear in focus, and the mar-gins are sharp and clear, which allows for easyremoval of unwanted background around eachpalynomorph. Several palynomorphs can then beassembled on a single plate with minimum lostspace and on a more desirable, totally white back-ground. On occasions the stacking process mayintroduce undesirable artifacts such as halos nextto some edges, smearing of reflected and refractedhighlights, trails of dots left when a single “badpixel” or “dust spot” on the camera sensor appearsin a different place on each frame after resizing andalignment, and erroneous patches often caused bynoise in the original frames. Such artifacts can beeasily edited manually in an image editing pro-gram.

RESULTS AND EXAMPLES

As illustrations of the results obtained by theuse of depth of field reconstruction on actualpalynomorphs, pictures depicting selected speci-mens from two different palynofloras are pre-sented. The first palynoflora is the latestCretaceous (Maastrichtian) Hell Creek flora fromsouthwestern North Dakota, USA. Abundant litera-ture on the description of the palynological assem-blage and species exists (Funkhouser and Evitt1959; Stanley 1965; Leffingwell 1971; Tschudy1971; Jarzen 1977; Sweet 1986; Jerzykiewicz andSweet 1986; Hotton 1988; Srivastava 1994; Sweetand Braman 2001) and are summarized in thework of Nichols (2002) and Nichols and Johnson(2002). This particular flora is located within thenorthern Aquilapollenites biogeographical domain(Herngreen and Chlonova 1981), with the occur-rence of easily recognizable pollen grains of thetriprojectate complex. The second palynoflora usedis the mid-Cretaceous (Albian - Cenomanian)palynoflora of the region of Teruel, northeasternSpain. Also, references on these palynologicalassemblages can be consulted for the originaldescriptions (Menéndez Amor and Esteras Martín1964; Médus 1970; Cabanés and Solé de Porta1986; Querol and Solé de Porta 1989; Solé dePorta et al. 1994; Solé de Porta and Salas 1994;Peyrot et al. 2005, 2007a, 2007b). This palynofloracorresponds to the Cerebropollenites biogeograph-ical domain. However, there is a significant influ-

5

Bercovici, Hadley, & Villanueva-Amadoz: Improving Depth of Field

ence of the Northern Gondwana Province(Elaterate biogeographical domain) as well.

These palynofloras display a great diversity ofmorphological types with excellent preservationquality, which makes them ideal candidates toevaluate the performance of depth of field recon-struction for each case. Each palynomorph isdepicted on a series of pictures representing theindividual optical sections used for reconstructionfrom bottom to top. The reconstructions are shownat the end of the sequence as well and commentedbelow.

• Example of Figures 3 and 4 shows twotriprojectate pollen grains (Aquilapollenitesquadricretaeus and Aquilapollenites attenua-tus) from the Hell Creek flora. The very com-

plex three-dimensional structure makesthem spectacular and convincing candidatesfor depth of field correction. Using the “DoStack” macro efficiently combines all of theframes into a single comprehensive imagepreserving all the details and ornamentationof the exine.

• Depending on the mounting medium usedfor preparing the palynological slides,palynomorphs can more or less frequentlybe situated in an unfavorable position, notlying flat in the plane perpendicular to theoptical axis. For example, Figure 5 (Nyssa-pollenites spp.) shows that the use of depthof field reconstruction is very efficient in com-pensating for this problem.

FIGURE 3. Example of reconstruction for Aquilapollenites quadricretaeus Chlonova 1961 (48 µm, dyed, Hell Creek),a pollen grain with complex three dimensional structure. The final reconstructed image is the result of “Do Stack” onall seven optical sections.

FIGURE 4. Example of reconstruction for Aquilapollenites attenuatus Funkhouser 1961 (75 µm, dyed, Hell Creek), apollen grain with complex three dimensional structure. The final reconstructed image is the result of “Do Stack” on allseven optical sections.

6

PALAEO-ELECTRONICA.ORG

• Palynomorphs with very developed externalsculptures also benefit from depth of fieldreconstruction. Four examples are given, 1)Erdtmanipollis cretaceus (Figure 6) with awellpronounced crotonoid sculpture, 2)Retitriletes spp. (Figure 7) with a reticulatescultpure, 3) Wodehouseia spinata (Figure8) with echinate sculpture, and 4) Aequitrira-dites spp. (Figure 9) with reticulate-spinulosesculpture. More care should be taken in theselection of input frames for processing.Variable results can be obtained when pat-terns tend to overlap. The use of only a lim-ited number of optical sections may help.

• Figures 10 to 13 depict a series of palyno-morphs with a different structure betweenthe proximal and distal poles. Here, the “DoWeighted Average” macro was tested on alloptical sections. As can be seen, it is notnow possible to distinguish ornamentationthat occurs on the proximal pole from thatoccurring on the distal pole. Separating theinput images into two sets may be necessaryto describe some palynomorphs fully without

presenting a false image of everythingstacked in a single plane.

OTHER RELATED TECHNIQUES

While this paper represents the first mentionof the use of such software for paleopalynology,other superior imaging techniques do exist. Themost familiar is of course the scanning electronmicroscope (SEM), which uses a focused beam ofelectrons for analyzing the specimen. The possibil-ity of magnification with larger depth of field makesit a very desirable and useful tool for palynology(Ferguson et al. 2007). While it is evident that elec-tron microscopy is an important tool for palynologi-cal morphological analysis, its use is limited by anumber of factors. The preparation techniques forthe samples differ completely from those routinelyused for optical microscopy. Pollen grains must bemounted on a metallic stub and metallized by ionsputtering prior to observation. This implies thatsamples previously mounted in a permanentmedium on a microscope slide cannot be analyzedunder SEM.

FIGURE 5. Example of reconstruction for Nyssapollenites spp. (25 µm, dyed, Hell Creek), this particular specimen isnot mounted flat on the cover slide. The final reconstructed image is the result of “Do Stack” on all five optical sec-tions.

FIGURE 6. Example of reconstruction for Erdtmanipollis cretaceus (Stanley 1965) Norton in Norton and Hall 1969(42 µm, dyed, Hell Creek), a pollen grain with crotonoid ornamentation. The final reconstructed image is the result of“Do Stack” on all seven optical sections.

7

Bercovici, Hadley, & Villanueva-Amadoz: Improving Depth of Field

8

FIGURE 7. Example of reconstruction for Retritriletes sp. (45 µm, dyed, Hell Creek), a spore with complex ornamen-tation. The final reconstructed image is the result of “Do Stack” on all six optical sections.

FIGURE 8. Example of reconstruction for Wodehouseia spinata Stanley 1961 (80 µm, undyed, Hell Creek), a pollengrain with complex ornamentation. The final reconstructed image is the result of “Do Stack” on all six optical sections.

FIGURE 9. Example of reconstruction for Aequitriradites spp. (93 µm, undyed, Utrillas Formation), a spore with com-plex ornamentation. The final reconstructed image is the result of “Do Stack” on all five optical sections.

FIGURE 10. Example of reconstruction for an Unidentified liverwort spore (60 µm, undyed, Hell Creek), with differentmorphological features on proximal and distal view. The final reconstructed images are the result of “Do WeightedAverage” for the complete reconstruction, and of “Do Stack” for the reconstructions using only a subset of the originaloptical sections.

PALAEO-ELECTRONICA.ORG

Another lesser known tool is the confocallaser scanning microscope (CLSM). This micro-scopical technique, derived from optical micros-copy, allows for the imaging of very thin opticalsections (at a micrometer scale) (Cogswell andSheppard 1992; Sheppard and Shotton 1997;Claxton et al. 2005). These optical sections are notonly very thin, but also do not present any out offocus zone corresponding to the parts of the objectabove and below the focus plane. Computer recon-struction from these thin sections allows for veryprecise three- dimensional reconstruction that canbe viewed from different angles. This approach isvaluable as it represents a non-invasive way of re-imaging and re-investigating type collections. Useof CLSM for paleopalynology on the Duxbury(1983) Albian dinoflagellate cysts collection waspart of a project undertaken by the Natural HistoryMuseum of London (Feist-Burkhardt et al 1998;Feist-Burkhardt and Pross 1998).

However, both electron microscopy and con-focal laser microscopy require the use of expen-sive heavy items of equipment, which may not beroutinely available in any palynological laboratory.In contrast, optical microscopes are part of thestandard equipment, and optical photomicrographsare still the standard for comparison. CombineZM,and similar software, can be routinely used as aconvenient way to depict palynomorphs in publica-tions, but they do not pretend to replace moreadvanced imaging methods.

CONCLUSIONS

This method was verified to be a greatenhancement to the imaging of palynomorphsunder optical microscopy. As taxonomic nomencla-ture can change over time, and possible differ-ences in the identification among differentpalynologists can occur, illustration may be the only

FIGURE 11. Example of reconstruction for Libopollis jarzenii Farabee et al. 1984 (33 µm, dyed, Hell Creek), with dif-ferent morphological features on proximal and distal view. The final reconstructed images are the result of “Do Stack”on selected optical sections.

FIGURE 12. Examples of reconstruction for Cibotiumspora juncta (30 µm, undyed, Escucha Formation), with differ-ent morphological features on proximal and distal view. The final reconstructed images are the result of “Do Stack”on selected optical sections.

9

Bercovici, Hadley, & Villanueva-Amadoz: Improving Depth of Field

way to validate or verify published data years afterthe publication date. Thus, it cannot be overem-phasized how important it is to use the best possi-ble figures for palynomorphs, in association withthe publication of palynological taxonomic lists. Notonly does the software allow for reducing the num-ber of photomicrographs necessary to describepalynomorphs, it also provides a more naturalappearance to the depicted palynomorphs, morelike that seen under the optical microscope. Thisimprovement is obvious and helps with the recog-nition of specimens on a palynological slide bycomparison with published taxa as printed in ajournal. However, it should be noted that as a mainlimitation to this method, three-dimensional struc-tures are collapsed into a two-dimensional image,which can result in one feature obscuring another,or two or more features being merged.

Additionally, while this type of image process-ing has proved to be very helpful for paleopalynol-ogy, the method can be extrapolated in any othercase where depth of field reconstruction is needed.This technique could scale to others geological orpaleontological objects such as microfossils (fora-minifers, microvertebrates, and many others)observed under the binocular microscope, up tothe largest dinosaur bones photographed with acamera stand or macro lenses.

ACKNOWLEDGMENTS

We thank J. Broutin for inviting us to write thisreview and both J. Dejax and D. De Franceschi,

Muséum National d’Histoire Naturelle, for providingaccess and support using the microscope equip-ment. Additionally, we gratefully acknowledge thevaluable comments of D. Nichols and anotheranonymous reviewer who helped to improve thismanuscript.

REFERENCES

Abramowitz, M., 1987. Contrast Methods in Microscopy:Transmitted light, Olympus America Inc., Melville,New York.

Allen, R.D., David, G.B., and Nomarski, G. 1969. TheZeiss-Nomarski differential interference equipmentfor transmitted-light microscopy, Z. Wiss. Mikrosk,69:193-221.

Bradbury, S., and Evennett, P.J. 1996. Constrast Tech-niques in Light Microscopy, BIOS Scientific Publish-ers Ltd., Oxford, UK.

Cabanés, R., and Solé de Porta, N. 1986. Nuevas aport-aciones sobre la edad de la Formación Arcillas y Lig-nitos de la Traiguera. Cordillera Ibérica Oriental.Maestrazgo, Abstracts XI Congreso Español de Sed-imentología, Barcelona, pp. 36.

Chlonova, A.F. 1961. Spores and pollen of the upper halfof the Upper Cretaceous sediments of the easternpart of the West-Siberian Depression. Trudy InstitutaGeologii i Geofisiki Sibirskoe otdelenie, Akademyianauk SSSR 7:137 pp.

Claxton, N.S., Fellers, T.J., and Davidson, M.W. 2005.Laser scanning confocal microscopy, Department ofOptical Microscopy and Digital Imaging, NationalHigh Magnetic Field Laboratory, Florida State Univer-sity, 37 p., Unpublished (http://www.olympusflu-oview.com/theory/LSCMIntro.pdf).

FIGURE 13. Example of reconstruction for Taurocusporites segmentatus (46 µm, undyed, Utrillas Formation), withdifferent morphological features on proximal and distal view. The final reconstructed images are the result of “DoWeighted Average” for the complete reconstruction, and of “Do Stack” for the reconstructions using only a subsetof the original optical sections.

10

PALAEO-ELECTRONICA.ORG

Cogswell, C.J., and Sheppard, C.R.J. 1992. Confocal dif-ferential interference contrast (DIC) microscopy:including a theoretical analysis of conventional andconfocal DIC imaging, J. Microscopy, Blackwell Sci-ence, 165-1:81-101.

Davidson, M.W., and Abramowitz, M. 2002. OpticalMicroscopy, In Hornak, J. (ed.), Encyclopedia ofImaging Science and Technology, Wiley-Inter-science, New York, 1106-1141.

Doher, L.I. 1980. Palynomorph preparation procedurescurrently used in the paleontology and stratigraphylaboratories, USGS Circular 830.

Duxbury, S. 1983. A study of dinoflagellate cysts andacritarchs from the Lower Greensand (Aptian toLower Albian) in the Isle of Wight, Southern England.Palaeontographica Abt. B., 186,1-3:18-80.

Farabee, M.J. Daghlian, C.P. Canright, J.E. Oftedahl, O.1984. Libopollis, a new pollen genus from the UpperCretaceous (Maestichtian) of North America. Palyn-ology, 8: 145-164.

Feist-Burkhardt, S., and Pross, J. 1998. Morphologicalanalysis and description of Middle Jurassic dinofla-gellate cyst marker species using confocal laserscanning microscopy, digital optical microscopy andconventional light microscopy, Bulletin du Centre derecherches Elf Exploration Production, Pau, 22-1:103-145.

Feist-Burkhardt, S., Henderson, A.S., McLachlan, I., andDuxbury S. 1998. 1983. Database of Cretaceousdinoflagellate cysts, (http://www.nhm.ac.uk/research-curation/projects/duxbury/)

Ferguson, D.K., Zetter, R., and Paudayal, K.N. 2007.The need for the SEM in palaeopalynology. C. R.Palévol, 6:423-430.

Funkhouser, J.W. 1961. Pollen of the genus Aquilapolle-nites. Micropaleontology, 7-2: 193-198.

Funkhouser, J.W., and Evitt, W.R. 1959. Preparationtechniques for acid-insoluble microfossils, Micropale-ontology, 5-3:369-375.

Hadley, A. 2006. CombineZM and CZBatch - Free imagestacking software for depth of field correction , http://www.hadleyweb.pwp.blueyonder.co.uk/

Herngreen, G.F.W., and Chlonova, A.F. 1981. Creta-ceous microfloral provinces. Pollen et Spores,23:441-555.

Hoffman, R. 1977. The modulation contrast microscope:principles and performance, J. Microscopy, Oxford,110:205-222.

Hotton, C.L. 1988. Palynology of the Cretaceous-Tertiaryboundary in central Montana, U.S.A., and its implica-tion for extraterrestrial impact. Ph.D. dissertation,University of California, Davis, California.

Jarzen, D.M. 1977. Angiosperm pollen as indicators ofCretaceous-Tertiary environments, Syllogeus, 12:39-49.

Jerzykiewicz, T., and Sweet, A.R. 1986. The Cretaceous-Tertiary boundary in the central Alberta Foothills:stratigraphy. Canadian Journal of Earth Sciences,23:1356-1374.

Leffingwell, H.A. 1971. Palynology of the Lance (LateCretaceous) and Fort Union (Paleocene) formationsof the type Lance area, Wyoming. In Kosanke, R.M.,and Cross, A.T. (ed.), Symposium on palynology ofthe Late Cretaceous and early Tertiary, GeologicalSociety of America Special Paper, 127:1-64.

Medús, J. 1970. A palynological method for stratigraphi-cal correlations. A study of the Barremian, Aptian andAlbian Complex of North-Eastern Spain and Roussil-lon in France. Grana, 10-2:49-158.

Menéndez Amor, J., and Esteras Martín, M. 1964.Observaciones palinológicas sobre la microflora dela cuenca lignitífera de Utrillas (Teruel). EstudiosGeológicos, 20,1-2:171-174.

Norton, N.J., Hall, J.W. 1969. Palynology of the UpperCretaceous and Lower Tertiary in the type locality ofthe Hell Creek Formation, Montana, U.S.A. Palaeon-tographica Abteilung B 125: 1-64

Nichols, D.J. 2002. Palynology and palynostratigraphy ofthe Hell Creek Formation in North Dakota: A micro-fossil record of plants at the end of Cretaceous time.In Hartman, J.H., Johnson, K.R., and Nichols, D.J.(ed.), The Hell Creek Formation and the Cretaceous-Tertiary boundary in the northern Great Plains: anintegrated continental record of the end of the Creta-ceous, Geological Society of America Special Paper,361:393-456.

Nichols, D.J., and Johnson, K.R. 2002. Palynostratigra-phy and microstratigraphy of Cretaceous-Tertiaryboundary sections in southwestern North Dakota. InHartman, J.H., Johnson, K.R., and Nichols, D.J.(ed.), The Hell Creek Formation and the Cretaceous-Tertiary boundary in the northern Great Plains: anintegrated continental record of the end of the Creta-ceous, Geological Society of America Special Paper,361:95-143.

Nomarski, G. 1955. Microinterféromètre différentiel àondes polarisées, J. Phys. Radium, Paris, 16:S9-S13.

Peyrot, D., Jolly, D., and Barrón, E. 2005. Nuevas aport-aciones a la Palinología del Cretácico Inferior de laSubcuenca de Oliete (Fm Escucha, Teruel). Bienalde la Historia Natural, pp.195-168.

Peyrot, D., Rodríguez-López, J.P., Barrón, E., and Melé-ndez, N. 2007a. Palynology and biostratigraphy ofthe Escucha Formation in the early Cretaceous Oli-ete Sub-basin, Teruel, Spain. Revista Española deMicropaleontología, 39,1-2:135-154.

Peyrot, D., Rodríguez-López, J.P., Lassaletta L., Melén-dez, N., and Barrón, E. 2007b. Contributions to thepalaeoenvironmental knowledge of the Escucha For-mation in the Lower Cretaceous Oliete Sub-basin,Teruel, Spain. C. R. Palevol, 6:469-481.

Pluta, M. 1989. Advanced Light Microscopy, (3 vols.),Elsevier, New York.

11

Bercovici, Hadley, & Villanueva-Amadoz: Improving Depth of Field

Querol, X., and Solé de Porta, N. 1989. Precisiones cro-noestratigráficas sobre la Fm. Escucha en el sectornoroeste de la cuenca del Maestrazgo. CordilleraIbérica oriental. Acta Geológica Hispánica, 24-2:73-82.

Sheppard, C.R.J., and Shotton, D.M. 1997. ConfocalLaser Scanning Microscopy, BIOS Scientific Publish-ers Ltd., Oxford, UK.

Slayter, E.M., and Slayter, H.S. 1992. Light and ElectronMicroscopy, Cambridge University Press, Cam-bridge, UK.

Solé de Porta, N., and Salas, R. 1994. Conjuntos micro-florísticos del Cretácico Inferior de la Cuenca delMaestrazgo, Cordillera Ibérica Oriental (NE de Espa-ña). Cuadernos de Geología Ibérica, 18:355-368.

Solé de Porta, N., Querol, X., Cabanés, R., and Salas,R. 1994. Nuevas aportaciones a la palinología ypaleoclimatologia de la Formacion Escucha (Albi-ense inferior-medio) en las cubetas de Utrillas y Oli-ete; Cordillera Iberica Oriental. Cuadernos deGeología Ibérica, 18:203-215.

Srivastava, S.K. 1994. Palynology of the Cretaceous-Tertiary boundary in the Scollard Formation ofAlberta, Canada, and global KTB events, Review ofPalaeobotany and Palynology, 83:137-158.

Stanley, E.A. 1965. Upper Cretaceous and Paleoceneplant microfossils and Paleocene Dinoflagellates andHystrichosphaerids from northwestern South Dakota,of American Paleontology Bulletin, 49:179-384.

Sweet, A.R. 1986. The Cretaceous-Tertiary boundary inthe central Alberta Foothills, II. Miospores and pollentaxonomy, Canadian Journal of Earth Sciences,23:1375-1388.

Sweet, A.R., and Braman, D.R. 2001. Cretaceous-Ter-tiary palynoflora perturbations and extinctions withinthe Aquilapollenites phytogeographic province.Canadian Journal of Earth Sciences, 38:249-269.

Traverse, A. 2007. Paleopalynology, Second Edition(Topics in Geobiology 28). Springer, Dordrecht, Neth-erlands, 813.

Tschudy, R.H. 1971. Palynology of the Cretaceous-Ter-tiary boundary in the northern Rocky Mountain andMississippi Embayment regions. In Kosanke, R.M.,and and Cross, A.T., (ed.), Symposium on palynologyof the Late Cretaceous and early Tertiary. GeologicalSociety of America Special Paper 127:65-111.

Wood, D.G., Gabriel, A.M., and Lawson, J.C. 1996.Palynological techniques - processing and micros-copy. In Jansonius, J., and McGregor, D.C. (eds.),Palynology: Principes and applications, AASP Foun-dation, 1:29-50.

12