AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 95:355-362 (1994) Notes and Comments Endocasts and Meningeal Vascular Patterns comparison of data derived directly from the skull with what was preserved on the endocast. In most cases, the endocasts resulted in a loss of information, or, worse, preserved Michael K. Diamond an incomplete and easily misinterpreted pattern. As for the endocasts I examined Department of Anatomy, Palmer College of Chiropractic, Davenport, Iowa 52803 that were prepared by others, information In a recent issue of the AJPA, Falk (1993) was frequently absent, ambiguous, or mispresented frequency distributions of leading when compared with data derived meningeal groove patterns in great apes. directly from skulls of the same species or From her data she concluded that the genera. Each primate genus presents its own pemeningeal arteries of the middle cranial culiar problems in the interpretation of crafossa and the endocranial surface superior nial vascular traces (Diamond, 1988:1&20, to it (“middle braincase”) are mainly sup268-287, 335-353). Most of these problems plied from the orbit (via the ophthalmic artery) in Pongo, while Pan relies mainly on a are exacerbated through the use of ensupply from the infratemporal fossa (via the docasts. In the great apes there tends to be a middle meningeal artery). The relative ex- lot of tunneling, especially around such crittent to which Gorilla relies on these two al- ical areas as the orbit and floor of the middle ternative sources falls between the other cranial fossa (Fig. 1).The situation is partictwo genera. Falk believes that none of the ularly bad in Gorilla, where there is somegreat apes relies on the middle meningeal times extreme overgrowth of the internal table of bone that buries meningeal grooves. In artery to the extent that humans do. Pongo, meningeal grooves running along the Falk collected her data from latex enpetrosquamous and petrosphenoid sutures docasts. This is a highly questionable source are frequently converted into canals that of information. An endocast is two steps will, perforce, not show up on endocasts (see, away from the circulatory pattern that one 1992, Fig. 7). Pan for example, Diamond, hopes to reconstruct from it. It records a poshas the fewest problems, but even in this itive impression of the bony traces left by genus the situation around the orbit is meningeal arteries and veins as well as poorly resolved by endocasts. other blood vessels such as dural sinuses Falk‘s data would be more tenable had she and emissary veins. Endocasts are very good physically verified the intramural arterial at reproducing vascular grooves, but these pathways that she has chosen to reconstruct are not the only kind of traces that blood of endocasts. She could solely on the basis vessels leave behind. Vessels also leave behave used sectioned skulls and either shined hind canals and foramina. Latex will preor injected a colored a light, passed a bristle, serve the very proximal part of an intramusolution through the meningeal canals and ral vessel trace but will seldon reproduce its foramina before initiating the casting procefull course or consistently reveal the existdure. The course and connections of intraence or position of an exit point. mural vessel traces can only be reliably In the course of prior research (Diamond 1988,1991a) I prepared a number of homi- established by physically tracing them (Dianoid endocasts (17 human and one each of mond, 1988; 1991a). Without such a procePan, Gorilla, Pongo, and Hylobates). All of these endocasts were taken from skulls that were hemisected or that had their roof reReceived December 31,1993; accepted February 27,1994. moved. I was thus able to make a qualitative 0 1994 WILEY-LISS, INC. M.K.DIAMOND 356 mbn. an m a x . a. Fig. 1. Depiction of the left lateral side of an endocast of Pun with the meningeal grooves (heavy black lines) drawn in a somewhat simplified manner. Broken lines indicate those meningeal grooves that are most subject to being converted into canals in great apes. The two sources of supply to the meningeal arteries of the middle cranial fossa have been drawn in schematic fashion. Numbers 1, 2, and 3 indicate the three main branches (or territories) that Falk (1993) considered in her analysis: 1, lower temporal branch 2, sylvian branch; 3, coronal branch. Abbreviations: ext. car. a., external carotid artery; int. car. a., internal carotid artery; max. a,, maxillary artery; men.-lac. anast., meningeal arteryfiacrimal artery anstomosis; mid. men. a,, middle meningeal artery; oph. a,, ophthalmic artery; trans. sin., transverse sinus sulcus; sig. sin., sigmoid sinus sulcus. dure it is often difficult to determine whether a vessel re-enters the cranial cavity, exits the skull, or terminates in the diploe. For example, grooves on an endocast that appear to be emerging from the orbit may actually be connected to the “anterior branch” middle meningeal groove (and the infratemporal fossa) via a canal running through the sylvian crest of the parietal bone. In other words, an apparent supply from the ophthalmic artery might actually be a supply from the maxillary artery (via the middle meningeal). Falk’s endocasts were independently coded by a graduate assistant, and the two sets of results were in close agreement. However, this only proves that Falk’s method of interpretation is transferrable t o others. If a feature is hidden on an endo- cast, it will be consistently invisible to all observers. Falk commits a t least one clear error of interpretation on the Pongo endocast illustrated in her Figure 5. She believes it shows evidence of an artery that emerged into the middle cranial fossa through the postglenoid foramen. She also mentions other Pongo endocasts with a similar pattern. I expect she was led to this hypothesis by a failure to find any evidence of a more typical orbital or infratemporal arterial entry. This failure must be due t o the roofing over or tunneling of arteries and not due to genuine absence because her reconstruction is impossible. A postglenoid artery has never been verified in any placental mammal, either through dissection or through the examination of sectioned embryos (Wible, 1990). I have certainly never seen one in any of the placental mammals I have dissected (Diamond, 1991a, 1992).The only mammals for which a postglenoid artery is known are some marsupials, and the vein it accompanies is not homologous with the postglenoid vein of placentals (Wible, 1990). Assuming that the illustrated specimen does, in fact, preserve a cast of the postglenoid foramen, a more likely interpretation is that the endocast reflects partial drainage of the petrosquamous sinus via the postglenoid vein. The petrosquamous sinus and its accompanying meningeal artery run along the petrosquamous suture in a common sulcus. An endocast cannot discriminate between the two vessels. The artery presumably gave off an ascending branch to the cranial side wall opposite the postglenoid foramen-a coincidental relationship. Any confidence one might place in endocasts is further eroded by the very poor correspondence between Falk’s published data and data that I previously collected from the sectioned skulls of extant hominoids (see Diamond, 1988). While we each asked different questions and coded our data in different ways, there were some areas of overlap. From my original data, I was able to extract the frequency with which there was no communication between the meningeal vessels of the middle cranial fossa and the orbit. This would correspond to 357 NOTES AND COMMENTS TABLE 1. Comparison of frequencies of meningeal groove patterns in hominoids' Falk (1993) Pan troglodytes Gorilla Diamond (previously unpublished) Pongo Pan troglodytes ~~ Sample size Pattern' A1 E (all ~ a t e g o r i e s ) ~ Unclassified A1 1%) E (76) Unclassified (%) 'R Gorilla _ Pongo_ Homo ~ _ R R R R B R B R B R 20 37 17 31 6 13 44 37 45 100 2 12 2 - _ 4 23 10 1 21 78 10 60 8 54 41 41 0 35 9 0 80 20 20 20 11 44 4 _ _ 3 - 7 1 2 7 1 0 - 1 0 _ 39 32 - _ _ 17 0 - _ 0 15 0 62 27 2 23 51 1 21 78 samples consisting of only right hemicrania; B = samples consisting of both left and right hemicrania. with most individual specimens therefore counted twice. 'A1 = complete supply to the middle cranial fossa from ophthalmic artery; E = complete supply to the middle cranial fossa from middle meningeal artery. 'Specimens allocated to pattern E by Diamond include only those specimens in which there is no communication between the meningeal grooves of the middle cranial fossa and the orbit. = an extreme example of Falks pattern E (exclusive supply from the middle meningeal artery). I also recorded the frequency with which there was no connection between the meningeal vessels in the middle cranial fossa and the infratemporal fossa. This would correspond to Falk's pattern A1 (total supply from the ophthalmic artery). The frequencies are shown in Table 1.For some species, the number of individuals Falk studied was relatively small. As a result, her samples drawn from only one side are subject to sampling error. Therefore, I have reproduced her data from the right side and both sides combined. Because I examined many more specimens than Falk, my comparison sample consists only of right hemicrania, except in a few cases where only the left side of a specimen was accessible. I had no data on Pan paniscus, so only samples of Pan troglodytes are compared. In my sample, 9 specimens of Pan (20%) n = 44)) 8 specimens of Gorilla (21%, n = 37), and 23 specimens of Pongo (51%, n = 45) were unclassified due to connections of the meningeal grooves to both the orbit and the infratemporal fossa. Two additional specimens of Gorilla (5%)were unclassified due to there being no detectable connection to either the orbit or the infratemporal fossa (an anatomical impossibility). I did not lump my unclassified specimens into categories A2-A4 (dual supply to the middle cranial fossa from both the oph- thalmic and middle meningeal arteries) for two reasons. First, the coding procedure I employed does not adequately resolve these patterns. Second, presence of an orbital communication does not necessarily mean a contribution t o the middle cranial fossa from the orbit (a requirement of pattern A). Blood could just as easily be going toward the orbit, depending on the details of the pattern. Unlike Falk, I often found it difficult to make a determination of blood flow direction in such specimens. By the same reasoning, the percentages listed under pattern E in my column represent mimimum values and consist only of those specimens without any communication between the middle cranial fossa and the orbit. The actual percentages of pattern E are likely to be considerably higher. In humans, for example, complete separation of the middle cranial fossa from the orbit is much less common than in the African apes (humans are in fact much closer to orangutans in this regard; see below). Yet it is clear that humans rely almost exclusively on the middle meningeal artery to supply the middle cranial fossa (Diamond, 1988, 1991b). There are some marked discrepancies between Falk's data and mine. According to my data, in Pan troglodytes pattern E is much more frequent (80%vs. 54%).This figure is also much greater than Falk reports for Pan paniscus (63%, n = 19, both sides _ 358 M.K. DIAMOND combined). In Gorilla, pattern A1 is much less frequent (11%vs. 39%)and pattern E is much more frequent (62% vs. 32%) than Falk reports. In Pongo, pattern A1 is much more frequent (44%vs. 15%).The only point on which our data seem to agree closely is that Pan troglodytes has a low frequency of pattern A1 (0%vs. 8%). While my frequency percentages are quite different, the general morphocline that Falk constructed is supported. Pongo does rely on the ophthalmic artery to a greater extent than the other comparison species. While I have no idea how frequent pattern E really is in Pongo, it has t o be less than in any of the other species given the high frequency of pattern A l . Among the African apes, Gorilla does indeed rely more heavily on the ophthalmic artery than Pan, albeit only slightly. Given that Pan has a minimum, value of 80% for Pattern E, I doubt that it is any less reliant than humans on the middle meningeal artery to supply the middle cranial fossa. However,if encroachment into the orbit is also included, humans may yet occupy the far end of the pattern E morphocline. To assess this possibility, I re-analyzed data from a subsample of 100 right hemicrania drawn from two skeletal populations that were reported on previously by Diamond (1988). Fifty skulls were from the University of Chicago’s Department of Anatomy and presumably trace their origin to the Indian subcontinent. The other 50 were from white Americans of both sexes drawn from the Terry Collection of the U.S. National Museum. Of these 100 hemicrania, 79% showed an apparent communication between the orbit and the meningeal grooves of the middle cranial fossa. Comparison percentages are 20% in Pan (n = 44), 33% in Gorilla (n = 37), and 96%in Pongo (n = 45). In all but one of the human hemicrania, the meningeal grooves also communicated with the infratemporal fossa through the foramen spinosum. Therefore, 78% could not be classified as conclusively possessing either pattern E or pattern A2-A4 (Table 1). The 79%value for humans is quite close to published data drawn from dissection studies of the human orbit and its collateral circulation. Ducasse et al. (1985) report that 83% of the orbits they dissected (n = 70) showed anastomosis between the arteries of the orbit and the middle cranial fossa. Hayreh (1962) reported a frequency percentage of 85% (n = 59). Despite the high frequency of arterial communication between the orbit and middle cranial fossa, neither of these papers reported a single case of a supply to the middle cranial fossa from the orbit, only the reverse. A supply to the middle cranial fossa from the ophthalmic artery is unusual enough that it is the subject of a substantial number of case reports in the medical literature (e.g., McLennan et al., 1974; Dilenge and Ascherl, 1979). In my own examination of an expanded sample of 638 human hemicrania, only 17 (3%) showed unmistakable evidence of a supply to the middle cranial fossa from the ophthalmic artery (Diamond, 1991b). The high frequency of arterial communication between the orbit and the middle cranial fossa may indicate that humans have been extending the territory of the middle meningeal artery into the orbit and thus expanding the boundaries of pattern E. Of the dissection studies cited above, Ducasse et al. (1985) reported that 27% of orbits (n = 70) definitely showed some takeover by the middle meningeal artery (usually the lacrimal branch of the ophthalmic artery), while Hayreh (1962) implied that around 22% of the orbits he examined (n = 59) showed some degree of annexation by the middle meningeal artery. Australopithecines and pre-erectus Homo appear to have a basically chimpanzee-like pattern of meningeal grooves (Diamond, 1988). It would not be surprising if the similarity extended to a low frequency of communication between arteries of the orbit and middle cranial fossa. One can speculate that the re-establishment in later Homo of communication a t frequencies nearly as high as Pongo would be a case of phenotypic similarity but functional disparity. LITERATURE CITED Diamond MK (1988) Cephalic Vascular Evolution and Development in Primates: The Stapedial Artery and NOTES AND COMMENTS its Companion Venous Sinuses. Ph.D. dissertation, University of Chicago. Ann Arbor: University Microfilms. Diamond MK(1991a) Homologies of the stapedial artery in humans, with a reconstruction of the primitive stapedial artery configuration of euprimates. Am. J. Phys. Anthropol. 84:433-462. Diamond MK (1991b) Homologies of the meningeal-orbital arteries of humans: A reappraisal. J . Anat. 178: 223-241. Diamond MK (1992) Homology and evolution of the orbitotemporal venous sinuses of humans. Am. J. Phys. Anthropol. 88:211-244. Dilenge D, and Ascherl GF (1979) Variations of the ophthalmic and middle meningeal arteries. AJNR 1: 45-53. Ducasse A, Segal A, Delattre JF, Burette A, and Flament J B (1985) La participation de l’artere carotide externe a la vascularisation orbitaire. J . Fr. Ophtalmol. 8r333-339. Falk D (1993)Meningeal arterial patterns in great apes: Implications for hominoid vascular evolution. Am. J. Phys. Anthropol. 92:81-97. Hayreh SS (1962) The ophthalmic artery. 111. Branches. Br. J. Ophthalmal. 46:212-247. McLennan JE, Rosenbaum AE, and Haughton VM (1974) Internal carotid origins of the middle meningeal artery. Neuroradiology 7:265-275, Wible J R (1990) Petrosals of Late Cretaceous marsupials from North America and a cladistic analysis of the petrosal in therian mammals. J Vert Paleont 10:183205. Reply to Dr. Diamond Dean Falk Department of Anthropology, SUNY at Albany, Albany New York 12222 I was pleased to read that in Dr. Diamonds Comments (Diamond, 1994), “the general morphocline that Falk constructed is supported”; yet I was surprised to read that “data from latex endocasts. . . [are] a highly questionable source of information.” Until now, the only thing that has not been controversial about primate endocranial casts (endocasts) is the observation that they clearly reproduce the meningeal vessels that course on the external surface of the dura mater and are “grooved into the internal surface of the braincase. In fact, it is often easier to discern meningeal arteries 359 from veins on endocasts than in the cranium alone because the smaller arteries stand out on endocasts as “zigzag beading marks” (Jones, 1912) that course on or near their wider venous companions. Nevertheless, Diamond asserts that there are “peculiar problems” associated with interpreting meningeal patterns on ape endocasts because of tunneling of meningeal vessels and their tendency to be covered by (‘overgrowth of the bony table, etc., and he therefore questions the accuracy of my observations of meningeal arterial patterns on 100 hemispheres from pongid endocasts. A possible explanation of these “peculiar problems” is that Diamond has prepared only one endocast (or hemicast; it is not clear which) each from Pan, Gorilla, and Pongo and compared them with their respective skulls. Thus he is basing his generalization of a lack of relationship between ape endocasts and their respective crania on sample sizes of one, a practice that is considered statistically invalid. Diamond has also observed endocasts “prepared by others,” but he does not specify their number nor species nor say if he compared these endocasts with their respective skulls. On page 351 of his dissertation (Diamond, 19881, Diamond states that he did a statistical study of 86 endocasts (hemicasts?) of Pan paniscus but does not list any endocasts for the other great apes. This is confusing because in the present note he states that “I had no data on Pan paniscus, so only samples of Pan troglodytes are compared.” This is discussed below. With regard to the endocasts that I studied, I used a casting procedure that employed a thin liquid latex that cast foramina including openings of bony canals (Fig. 1). I had no need to “have used sectioned skulls and either shined a light, passed a bristle, or injected a colored solution through the meningeal canals and foramina before initiating the casting procedure” (Diamond, 1994). With the endocasts that I used, there were no “peculiar problems.” The eight pat- Received March 28,1994; accepted April 20,1994.