A 3D journey on virtual surfaces and inner structure of ossa genitalia in Primates by means of a non-invasive imaging tool

Novel bio-imaging techniques such as micro-Computed Tomography provide an opportunity to investigate animal anatomy and morphology by overcoming limitations imposed by traditional anatomical drawings. The primate genital bones are complex anatomical structures whose occurrence in both male penis (baculum) and female clitoris (baubellum) may be difficult to assess in individual cadavers. We tested a 3-step methodological protocol, including different techniques ranging from inexpensive/simple to more expensive/sophisticated ones, by applying it to a sample of primate species, and resulting in different levels of data complexity: (1) presence/absence manual palpation method; (2) 2D X-ray plates; 3) 3D micro-CT scans. Manual palpation failed on 2 out of 23 specimens by detecting 1 false negative and 1 false positive; radiography failed once confirming the false positive, however firmly disproved by micro-CT; micro-CT analysis reported the presence of 9 bacula out of 11 male specimens and 1 baubellum out of 12 female specimens. A different baculum position was identified between strepsirrhine and haplorrhine species. We also aim to assess micro-CT as a non-invasive technique providing updated anatomical descriptions of primate ossa genitalia. Micro-CT 3D volumes showed the surface of some bones as rough, with a jagged appearance, whereas in others the surface appeared very smooth and coherent. In addition, four main types of bone internal structure were identified: 1) totally hollow; 2) hollow epiphyses and solid diaphysis with few or several channels inside; 3) totally solid with intricate Haversian channels; 4) totally solid with some channels (structure of single baubellum scanned). Ossa genitalia appeared as a living tissue having its own Haversian-like channels. The high resolution of micro-CT 3D-images of primate genital bones disclosed additional form variability to that available from genital bone 2D images of previous studies, and showed for the first time new internal and external morphological characters. Moreover, micro-CT non-invasive approach proved appropriate to recover much of scientific knowledge still hidden and often neglected in both museum specimens and primate cadavers only destined to necropsy.

Introduction length and prolonged intromission patterns during copulation [43,45,47,48], represents the only available data so far, supporting just one of the few existing hypotheses about baculum adaptive function [22,24,44,49]. Although informative, this analysis is nevertheless firstly limited to bacula only, and secondly restricted to just one morphometric one-dimensional parameter such as length, with all potential information contained in bone shape variability neglected.
In this study we firstly aim to identify the feasibility and potential of a 3-step methodological protocol to reliably detect ossa genitalia in primates by means of non-invasive techniques. We compared data acquisition methods characterized by a progressive increase in both accuracy/ completeness and costs. Our protocol goes from the obtainment of simple occurrence data, to virtual volume and inner structure data by applying: (1) manual palpation, (2) 2D projectional radiography (x-ray plates), and (3) 3D micro-CT. Secondly, we show how 3D high resolution virtual reconstructions of penile and clitoral bones reveal to be a precious source of knowledge in primates as well, disclosing for example (i) exact spatial position of bones inside tissues of external genitals, (ii) external and inner morphological variability, and (iii) the importance of relying on detailed 3D data in order to make meaningful comparative studies among primate species (as well as other mammal orders). 3D models of ossa genitalia have been stored in a permanent archive (morphosource.org), available for further functional analyses.

Samples
Samples of external genitals came from several sources. From the historical (XIX-XX century) primate wet collection of the Natural History Museum of "La Specola" in Florence, Italy (NHMLS) [50] we obtained 13 external genitals (separated in loco from the whole body by cutting at the base of penis and clitoris close to the pubic bone surface) and one baculum (as an already extracted bone), all accounting for 13 different primate species. In addition, 8 fresh samples of external genitals belonging to 5 different species were obtained from corpses at Istituto Zooprofilattico Sperimentale di Lazio e Toscana (IZSLT, Rome, Italy). All Italian Istituti Zooprofilattici Sperimentali (II. ZZ. SS.) usually receive primate corpses (dead for natural causes or trauma/disease) from zoos and/or research institutes, in order to carry out necroscopic investigations aimed at identifying possible risks of zoonoses. Consequently, we established a collaboration with the IZSLT in order to be notified when a primate body underwent necropsy, and therefore the cut off external genitals were available. Generally, after necropsy animal remains are disposed of and most of the anatomical information related to the specimens is completely lost. One additional sample came from the Civic Museum of Zoology (CMZ) in Rome (Italy). In this case, the specimen was a female hybrid orangutan (Pongo pygmaeus x abelii) near to be taxidermied. We were able to obtain external genitals thanks to the taxidermist before the process began. Overall, we collected 23 samples of external genitals belonging to 19 different primate species, 5 of which belonging to the suborder strepsirrhines (N = 5 males) and 18 belonging to the suborder haplorrhines (N = 12 females, and N = 6 males) (S1 Table). As for museum specimens, neither skins nor skeletons but only wet samples could be considered as a source of genital bones, in that skins and skeletons could both be source of falsified data, specifically of false absences. In fact, skins, by definition, no longer include either soft tissues or bones inside them, while skeletons may have lost such small bones after the cleaning procedure (often boiling bones to detach tissues).

3-step methodological protocol
First step: the manual palpation method. Both penises and clitorises were exposed to a double palpation by two different operators (FS and MC) who touched the genitals with thumb and forefinger of the same hand (a) along medio-lateral axis and (b) along anteroposterior axis. Palpation type (b) may overcome problems related to the palpation type (a) during which a tendon or a ligament may be construed as a small bone instead. Second step: 2D projectional radiography (X-ray plates shot with Arcom Simply system for veterinary radiodiagnostics; 20 mA-40 kV setting, at "Enrico Fermi" veterinary clinic, Rome, Italy). This step was used to either verify data coming from the manual palpation method or select specimens that, possibly confirmed as containing a bone, were suitable for the step 3 tomography protocol. Third step: cone-beam micro-CT, assembled at the Department of Physics and Astronomy of Bologna University (Italy; http://www.fisica-astronomia.unibo.it/it/dipartimento/servizi-e-strutture/ strutture/laboratori-di-ricerca-1/indagini-con-i-raggi-x/radiografia-digitale-e-tomografia-3d). The computed tomography was used to scan samples and obtain 3D virtual volumes of genital bones. For all three steps, no bone extraction from tissues was necessary and integer genitals were also used to investigate both their external and internal morphology.
The total time needed to detect and visualize genital bones depends on the size of each bone. On average, the time required to carry out the 3-step methodological protocol is 3 hours and 50 minutes per bone (see Table 1).

Micro-CT setup and procedure
The set-up of the micro-CT scanner (Fig 1) is based on a microfocus X-ray tube (Thermo Scientific/Kevex PXS10-65W, Scotts Valley, CA, USA; operating voltage range: 45-130 kV; maximum beam current: 0.5 mA; minimum focal spot: 5 μm) and a Photonics Science (Saint Leonards-on-sea, United Kingdom) VHR1:1 Charge Coupled Device (CCD) camera (4008 × 2672 pixels, 9 μm pixel size) with a 36 × 24 mm 2 field of view (FOV). This camera features a high-resolution phosphor screen directly coupled to the CCD by means of a fibre-optic plate that protects the sensor against radiation damage. The sample manipulator consists of a high-precision rotary table (model M-037 by Physik Instrumente, Karlsruhe, Germany), on which the sample is positioned, and a tip-tilt platform (model M-042 by Physik Instrumente, Karlsruhe, Germany) to assure the correct alignment between the vertical rotation axis and the pixel matrix of the detector. The sample manipulator is mounted on a vertical translation axis (model M-415 Physik Instrumente, Karlsruhe, Germany) that makes it possible to analyse, in successive steps, objects with a vertical size larger than the CCD camera's FOV [51]. Table 2 shows the range of setting parameters used to scan the genital samples.

3-step methodological protocol
Step n˚Time (mins) The step-by-step scanning procedure consists of: (a) preliminary check of correct alignment of the system; (b) X-ray acquisition in order to verify presence of bone as identified for its high radiopacity, and to find the optimal scanning parameters; (c) trials of rotation to set sample position within the capture field; (d) setting correct parameters for each sample and launch the scanning procedure. A single scan lasts about 75 minutes at the end of which, samples are returned to their original ethanol containers.
Final phases are the post-processing and the reconstruction phase that were performed with the proprietary software PARREC (it uses parallel processing on several cores and CPUs). Radiographic projections are normalized with respect to the white field and dark field  (respectively the image acquired with no object and the image with X-rays off) producing the so called atenrads whose values are natural logarithm of the intensity transmission values ranging from 0 to 1 and corresponding to the different attenuation of the object along the X-ray path. After the adjustment of the optimal rotation centre on the images the actual reconstruction process starts and atenrads are first properly filtered in the frequency domain and then back projected with the cone beam algorithm. The result is a set of sections (in this case vertical sections) called slices whose values are proportional to density and atomic number of the materials (in this case bone and soft tissue). The stack of slices is loaded with VGStudioMax 2.1, a rendering and data processing commercial software capable to show a 3D data representation. A "histogram"-based segmentation (Fig 2) of different materials (in our case bone and soft tissue) visualizes a virtual three-dimensional scene and allows the grey-level adjustment. In particular, by using the classification tool it is possible to enable or disable different segments of the grey-level histogram corresponding to tissues of different density in the scanned sample, thus bringing out the genitals and/or the bone inside.
In addition, the rendering software allows the manipulation (translation, rotation and virtual slicing) of the object volume. It was therefore possible to virtually dissect scanned bones, in order to investigate their internal structures. In fact, in the "clipbox mode" the object may be clipped along the desired clipping direction, thus obtaining 3D orthogonal cross sections: axial ( Despite linear length data do not require the application of 3D techniques to be sampled, 3D reconstructions allow to collect linear data for museum specimens that cannot undergo a destructive sampling. By using the distance measurement tool, start and end points were set on the distal and proximal tips, respectively. Lengths were blindly sampled for three times by the same operator (FS) and means were calculated.
Limited to two fresh samples (baculum of Mandrillus sphinx, Linnaeus, 1758, and baubellum of Sapajus apella, Linnaeus, 1758; provider IZSLT, non-museum samples), the bone was extracted in order to directly investigate both orientation and position inside tissues. We followed a five-step extraction protocol: 1) a transversal cut of penis at the proximal end of bone; 2) a longitudinal cut, from the urethra external opening (distal end of genitals) until the proximal end of bone, therefore opening the whole penis in two halves; 3) several cuts in order to separate the bone from the penis; 4) boiling the extracted bone in water for 5 minutes; 5) detaching all tissues from bone to obtain a clean and smooth surface. Table 3. If we consider micro-CT (step 3) as the most sophisticated and sensitive technique, manual palpation failed on 2 out of 23 specimens, with a false negative and a false positive, while radiography (step 2) failed once. The latter sample, interestingly, though corresponding to the false positive by manual palpation, was firmly proved wrong by the true absence detected by the micro-CT (see discussion).

Validation of our 3-step protocol is shown in
Micro-CT validated X-rays showing the presence of a baculum in 9 out of 11 male specimens, i.e., 5 strepsirrhines and 4 haplorrhines (as shown in Fig 4):1) Galagoides demidoff; 2) Galago gallarum; 3) Loris lydekkerianus; 4) Loris tardigradus; 5) Otolemur crassicaudatus; 6) Chlorocebus aethiops; 7) Mandrillus sphinx; 8) Papio cynocephalus; 9) Saguinus niger. Only 1 out of 12 female specimens revealed the presence of a baubellum, namely Sapajus apella (haplorrhine). Therefore, total number of 3D models of ossa genitalia obtained (available at https://www. morphosource.org/Detail/ProjectDetail/Show/project_id/771) was 10. Scanning was firstly devoted to genital bone reconstruction in order to investigate both surface and inner structure characteristics. Concerning the external surface, two different types were found, either rough with a jagged surface or very smooth and coherent; whereas for the bone internal structure, four main types were found (see Table 4 and Figs 5 and 6). A first type of bone had a totally hollow inner structure with neither trabecular-like systems nor labyrinthic system of channels inside (Figs 5 and 6A). By hollow we did not mean that bone is empty, rather that it does not contain anything radiopaque. However, only a destructive sampling could verify the presence (or not) of fatty bone marrow inside the bone. A second type would include two alternative versions: one having hollow epiphyses and solid diaphysis, with a few or several channels inside (Figs 5 and 6B), and another (baculum of S. niger) showing a hollow proximal epiphysis, but solid diaphysis and distal epiphysis. A third type was an entirely solid inner structure with intricate networks of trabecular-like systems and channels in both epiphyses and diaphysis (Figs 5 and 6C). Last type was a totally solid internal structure with only some channels (Figs 5 and 6D) which belong to a baubellum sample. Linear length data of each genital bone were reported in Table 4.
Secondly, scanned samples were also used to investigate the position of the bone inside the genitals. To do this, the area corresponding to peak of soft tissues which was disabled to let the bone appearing, was reduced and soft tissues appeared again in transparency (Table 4 and Fig  7). It was glaring the difference of baculum position between strepsirrhine and haplorrhine primates, the former showing baculum filling the entire length of penis, whereas in the latter genital bones were placed in the distal half of penis. Additional details about both baculum position and orientation inside the penis come from the manual bone extraction in one male  Presence (1) and absence (0) of genital bones in each specimen resulting from double palpation method (along latero-lateral and anteroposterior axis; Palpation), X-ray plates (X-rays), and micro-CT scans (micro-CT). In bold, selected specimens for micro-CT scans. Species are listed following a phylogenetic order (Suborder): firstly strepsirrhines (S) and after haplorrhines (H) and for every species sex (F = female; M = male) is also reported.
M. sphinx in which a dorsal position relative to the urethra was revealed (longitudinal section of penis, Fig 8A). However, while the baculum proximal epiphysis (considering the baculum as a long bone, see below) was visible and dorsal to the urethra, the distal epiphysis could not be seen from that perspective. In fact, an apparent slight torsion of baculum along proximo-distal axis made the distal epiphysis, inside the glans (i.e., as Dixson suggested [44]), becoming lateral to the urethra, resulting in proximal and distal epiphyses on different sagittal planes (Fig 8B  and 8C).  b. Comparing the first (Fig 9A) and the last (Fig 9B) radiographic projection of the genitals of Loris lydekkerianus, a mismatch appeared as if the sample was subjected to a (grisly!) repositioning. In fact, the overlaying of the red outline (radiographic projection of sample before the scan) and the yellow outline (radiographic projection of sample after the scan) revealed a sample volume reduction (Fig 9C, see discussion). The bidimensional blurry slices consequently obtained did not show suitable for 3D reconstruction.

Discussion
While traditional approaches and methodologies would entail genital bone extraction from tissues and histology (i.e., a completely disruptive approach) in order to study their position, orientation, external surface and internal structure, innovative extant non-invasive techniques would not. In fact, depending on museum requirements (e.g., in-situ availability of micro-CT, museum willingness to loan samples), the micro-Computed Tomography turns out to be the most suitable and exhaustive modern technique to achieve the above mentioned results. In particular, micro-CT gives high resolution information about external surface and internal structure nevertheless making an unaltered sample available for future researchers. Thanks to micro-CT, we recovered a high 3D-form variability: the "Y"-shaped baculum found in G. demidoff; the "stick"-baculum of O. crassicaudatus; the hook-baculum of M. sphinx and P. cynocephalus (both cercopithecids). Even more significantly, we disclosed for the first time the external and internal 3D morphology of 9 bacula for 9 species, and of a single baubellum as a tenth species. Differences detected in the internal structure, in particular, allowed to exclude the interpretation of ossa genitalia as a simple calcareous deposit and supported, instead, that of a living tissue having its own Haversian-like channels [13,52]. Four different internal structures clearly appeared, depending on presence/absence and concentration of trabecular-and Haversian-like systems: 1) a hollow baculum; 2) several bacula with solid epiphyses and hollow diaphysis; 3) two solid bacula; 4) a solid baubellum.
Although based on a small sample size, by comparing specimens to each other we detected two alternative positions of the baculum inside the penis: filling its entire length in Strepsirrhine species and placed in its distal half in Haplorrhine species. Despite no baubella were found in female strepsirrhines available, the single specimen of os clitoris scanned in the haplorrhine S. apella was placed inside the ending tip of clitoris, resembling haplorrhine baculum position. In previous anatomical studies, the positioning of both baculum and baubellum in the investigated species was either not reported, or explicitly stated in the text, or else simply inferable from drawings [24,31,35,41,52]. Position data, however, have never been emphasized as a discriminant trait between suborders and the only difference clearly reported is about length. In fact, it is possible that the longer bacula found in strepsirrhines [44] may contribute to the different position we observed relative to penis length. Our linear length measurements of scanned bacula supported that the longest one belongs to a strepsirrhine species (Otolemur crassicaudatus). In conclusion, in order to draw up a general pattern out of these individual observations and making it valid for the entire order of primates, differences between the positioning of the genital bones in Strepsirrhines and Haplorrhines must be confirmed by increasing sample size, for both baculum and baubellum. Asymmetry found in M. sphinx baculum should also find an appropriate context to be discussed and deepened. The proposed 3-step protocol of investigation might be used firstly as a tool to obtain either simple (presence/absence) or complex (3D) data-depending on the study conditions and limitations-and secondly as a sample sorting procedure to select qualified specimens for micro-CT scanning which, as reported above, represent a time consuming procedure. Step 1: although proved useful in anesthetized living individuals (Carosi pers. obs.), manual palpation in museum specimens could either be hampered by the stiffness of dead tissues long preserved in either alcohol or formalin (e.g., female S. niger dating 1843 in our sample) or else fail to c., C.ae., M.s., P.c., S.n.) and 1 baubellum (S.a.). a = strepsirrhine species; b = haplorrhine species. Letters below each 3D image stay for specimens' acronym (see Table 3).
Step 2: X-rays might represent a valuable tool if, exploring museum wet primate collections (i) a destructive sampling application is rejected, (ii) a micro-CT scanner is not available, and/or (iii) specimens available are only entire bodies too big to fit into the micro-CT scanner. Nevertheless, X-rays failed (one case) to discriminate between bone and radiopaque artefacts.
Step 3: micro-CT allowed to (1) recognise a radiopaque artefact, (2) virtually repair any kind of bone fracture (recovering a precious otherwise useless sample), (3) investigate both external and internal structure of genital bones.
Nonetheless, in the case of fresh samples from primate remains (fated to be cremated), a traditional disruptive extraction protocol turned out to be useful to investigate more in detail position and orientation of bone, in particular as related to the urethra. This finding is particularly useful if considering that many old museum specimens, even if available for disruptive sampling, would unlikely be unsuitable for such investigation due to the formalin perfusion, i. e., the old traditional museum preservation technique. Primate cadavers should then be considered precious material as well as museum primate collections and should not be wasted. Ancient treatises on primate comparative anatomy [23,35,36,53] came from the sacrifice of many wild individuals which nevertheless allowed to carry out several studies about (but not limited to) morphology and anatomy. Nowadays all primate species are currently listed in the main international conventions for endangered fauna preservation, and collection of wild individuals for research purposes is not allowed anymore based on both ethical and conservation issues. However, national and international natural history museums still host many of those wild-caught individuals which still dramatically support investigations on anatomy, morphology, systematics and taxonomy. In this scenario, in order to preserve these precious collections, an increased application of non-invasive vs. invasive techniques, should be encouraged.

Technical recommendations
During tomography scanning, it is mandatory that samples do not move in order to avoid resulting blurry images of slices not suitable for 3D reconstruction. In fact, for samples fixed in alcohol, evaporation may produce undesired specimen movements while scanning and, in addition, it may induce a reduction of tissues' volume. In order to prevent all this to happen and based on empirical trials, we suggest two alternative preventive measures: 1) specimens may be removed from alcohol and sit overnight to let the alcohol evaporating before the micro CT-scanning, and 2) before positioning the specimen on rotating turntable, it could be inserted in a small plastic bag which, briefly saturating with evaporated alcohol, will avoid the volume reduction.
We believe that advanced anatomical data about form, position and inner structure of genital bones in primates as those provided by micro-CT scanning technique might improve the study of the neglected topic of ossa genitalia. An investigation about relations between genital c., C.ae., S.n.). a = strepsirrhine species; b = haplorrhine species. Letters below each 3D image stay for specimens' acronym (see Table 3).
https://doi.org/10.1371/journal.pone.0228131.g007 bone form variation and species reproductive and ecological correlates, in a phylogenetic perspective, might help unravelling further functional and evolutionary hypotheses. In addition, virtual data about genital bone surface could be analyzed by using either geometric morphometric techniques (for example following the same landmark and semi-landmark method applied to investigate molluscan shell shape variation by [54][55][56]) or other methods, such as the finite element analysis [57][58][59][60][61] and "alpha-shapes" [62]. In conclusion, micro-CT technique proved appropriate to discover knowledge still hidden in museum specimens or fresh 3D journey on primate ossa genitalia cadavers, preventing them not only to be stored in a closet or else get trashed after necropsy, rather allowing them to still contribute to science and live a second life.