It was not that long ago that those of us doing paleoneurology and actively making and studying endocasts could probably have fit into a London phone booth, albeit with protest. Of course, study of the evolution of the brain, in general, has always been a popular subset of the Zoological sciences (think of the Edingers, Tilly, and father Ludwig, C. Ariens Kappers, F. Weidenreich, the Jesuit scholar CJ Connolly, H. Jerison, I. Kotchekova, to mention but a few), including Anthropology. It seems to have finally dawned on a growing number of paleoanthropologists that the most direct evidence for hominin (and hominid) brain evolution is to be found by studying the only really direct evidence, those sometimes ugly, seldom complete, and data-impoverished objects we call endocasts, moot with regard to almost all cerebral morphology. Ecological opportunities are conspicuously available in paleoneurology, particularly given the age profile of its current combatants….

It is no secret that endocasts provide few cortical convolutional details and this understanding has a long and controversial history (Symington 1916; Clark et al. 1936; Hirschler 1942; Bailey and Bonin 1951; Keith 1931; Ariens-Kappers 1934; Balzeau and Gillissen 2010; Black 1932; Edinger 1949; Grimaud-Herve 1997; Holloway et al. 2001, 2010; Holloway 2012; Radinsky 1968; Shellshear and Smith 1934; Weidenreich 1936, 1941, 1943; Wu et al 2006), and it is embarrassing to remember that such a highly regarded neuroanatomist as Smith (1926) regarded the Piltdown endocast as having an extremely primitive pattern, more so than in Homo erectus from Indonesia. Connolly’s (1950) book is more or less our “bible” on these issues, which also provided ontogenetic, and comparative “racial” observations, based on the collections at the Smithsonian Institute. Nevertheless, endocasts are the closest we can come to what was once an actual living brain, and is what I describe as the only true “direct” evidence regarding hominin brain evolution (Holloway 1964, 1996, 2009; Holloway et al. 2004). The goal of course is to synthesize the “direct” paleoneurological evidence with the rest of the fossil record, archaeological materials, particularly stone tools, living sites, faunal associates, current neuroscience, and human behavioral/social adaptations gleaned from comparative neuroscience and animal studies.

Today, it is becoming difficult to stay on top of the paleoneurological game, particularly if we include all of its speculation, and simply impossible to stay on top of the neurosciences. Nevertheless, the field of paleoneurology is in an expansive phase, even if ignored by most physical anthropologists, whether postcranial or dental specialists, and archaeologists (see Holloway 1997, 2008, 2009, for a brief history). I would like to offer a speculation as to why, aside from the obvious interest in human brain evolution: paleoneurology is becoming an aesthetic and a digital turn-on for younger scholars born with mice in hand. Not that art did not exist earlier as any study of the papers by Retzius, Ariens Kappers, Weidenreich, and many others would show, in their illustrations and lists of linear and arc measurements, but today’s software programs such as Amira, Analyze, Osirix, ITK-SNAP, Endex, to mention a few, offer both aesthetic and morphometric delights that surely must satisfy paleoneurological nerds, even dinosaurs like myself. One need only look at the beautiful images produced by Dean Falk and colleagues on LB1, Bruner, Weber, Neubauer, Gunz, Schoenemann, Balzeau, Gilissen, Grimuad-Hervé, Subsol, Thibaut, and Wu. Add to this virtual virtuosity sophisticated statistical packages and the advancement of morphometric techniques e.g., mirror-imaging, spline analysis, and algorithms for correcting some forms of distortion, add missing data points, and one can see much advancement over the days when endocast reconstruction reliability was scored between 1–4 (Holloway 1970). In short, these advances have made it possible to evolve a Paleoneurology that is more empirical, quantitative, advancing actual hypotheses for testing, and perhaps most importantly, the sharing of endocast data, where colleagues can challenge each other’s reconstructions and interpretations, given, of course, there exists a wholesome intellectual environment, and not a sheer competitive landscape brought on by a dearth of academic jobs available for an over-abundance of paleoneurologists.

Nevertheless, I remember fondly the good old days of pouring liquid latex into skulls, vulcanizing them, and extracting the endocast out from the foramen magnum with a satisfying expanding pop, hoping that the sella turcica was not included. Or, using Dentsply on sectioned crania and getting beautiful blue (or green) endocranial portions that hopefully would last forever with exquisite detail, as the molding was meant for dental crowns. When the endocasts were incomplete, as was the case for almost all of the African, Indonesian, and European ones I reconstructed since 1969, (see Holloway et al. 2004 for examples), the sculptor became alive and happy as I tried to add plasticine to the missing regions based on other endocasts of the same taxon that were more complete. Of course, that is a route for reducing endocranial volumetric variance, but one where decades of experience should count for something! Nor was there any lack of pleasure in dunking the endocast into various sized beakers to see if Archimedes’ Principle really worked across the taxa! Currently, one just hits the “volume” key in the software package, and voila! a volume with three (or more) decimal places appears! Sad to say, however, that the roughly 200 ape endocasts I made during the ‘70’s have deteriorated, their latex surfaces growing caramelized as I write this article. Fortunately, almost all of these have been scanned and are available at ORSA, University of Penn, under the directorship of Drs. Janet Monge and Tom Schoenemann.

Next was the task of sharing one’s work with one’s colleagues, and making sure that the home of the discovered crania was given an endocast, which meant then molding the endocast reconstruction and mailing them off to various parts of the world, which was not an easy, or enjoyable task. These were usually done by exchanging endocasts between respective collections, which could be difficult for those with nothing to exchange! Today, not only can one send the reconstruction or original electronically, but one can even make a 3-D model from the scan data. Just consider the beautiful reconstruction of the Malapa (MH1) endocast done by Berger, Carlson et al. (2011) from South Africa. Contrast that with the inability of myself to get a copy of the Daka endocast after doing the description and made to return the original cast without making a copy. The same for the Konso specimen. The lesson being, of course, that politics is still a major barrier to full academic sharing between fossil discoveries and major research centers which control the CT scan data of the specimens and are loath to share until every last pixel and voxel have been described. Even then politics tends to overcome our better instincts.

Of course, advances in technologies and electronic manipulations of data do not necessarily guarantee more reliable error-free results. Considerable neuroanatomical knowledge is still required, collections need expansion, morphology still needs identification and interpretation, and individual biases still play their nefarious roles in selective perception, segmenting, measuring, etc. The classic “garbage-in garbage-out” meme always remains a possibility, even with sharing of data. Having now segmented some 500+ modern human endocasts from museum crania using Analyze 11, ITK-SNAP, I know full well how it is possible to err in interpreting objects slice-by-slice, and selecting points with the mouse. Working segmentation with fossils still containing adhering debris is extremely challenging, particularly when the fossils are fragmented and/or distorted, as in several of the Neandertal (e.g., Forbes Quarry, Skhul 9, Amud), Homo erectus, and australopithecine specimens, e.g., Stw505, Hadar AL 444. Below are some examples of where the “new” and “virtual” raises some issues and doubts. Weber et al. (2012) have a nice paper on how some of the defects can be corrected.

The exchanges between Dean Falk and me over the decades can serve as prime examples of the above. Consider our recent exchanges in the American Journal of Physical Anthropology (Falk and Clarke 2007, 2012a, b; Holloway and Broadfield 2011, 2012a, b). Here, the issue was using a modern technique of mirror-imaging (truly modern?) the right side of the Taung endocast and finding a new volume of 382 ml, quite different than my 402 ml volume published in 1970, which was quite different than the 525 cc volume published by Dart. Inherent in the above arguments was the notion that somehow using laser scans and mirror-imaging was a real improvement over my (Holloway 1970) defining a midline, scribing it, and sanding it down to the midline, and multiplying the resulting hemi-endocast volume by 2, assuming hemispheric symmetry. When their results were published (Falk and Clarke 2007), it was apparent that they had not only not defined a midline explicitly by defining actual anatomical landmarks, but their mirror image showed a visible asymmetry between left and right sides, hence our critical response. The quest by Falk and colleagues to assert that my earlier works on australopithecine endocasts (Falk et al. 2000) provided inflated endocasts does not appear to have much support, judging by the recent Neubauer et al. (2012) paper. Nevertheless, despite our disagreement, the work by Falk and Clarke on that specimen is valuable and suggests that such methods and challenges to prior research is useful and welcome.