Development of the Corpus Callosum and Evolution of Axon Tracts (Anterior Commissure/Axon Guidance/Glial Sling/Substrate Pathways) MICHAEL J

Home , Anterior commissure, Lamina terminalis

Proc. Nati. Acad. Sci. USA Vol. 80, pp. 5936-5940, October 1983 Evolution

Ontophyletics of the nervous system: Development of the corpus callosum and evolution of axon tracts (anterior commissure/axon guidance/glial sling/substrate pathways) MICHAEL J. KATz, RAYMOND J. LASEK, AND JERRY SILVER Department of Developmental Genetics and Anatomy, Case Western Reserve University, Cleveland, OH 44106 Communicated by Walle J. H. Nauta, June 13, 1983 ABSTRACT The evolution of nervous systems has included pus callosum (2, 16, 26-28). Among the placental mammals, the significant changes in the axon tracts of the central nervous sys- corpus callosum generally increases as the neocortex increases tem. These evolutionary changes required changes in axonal growth (16). The corpus callosum is a truly new feature that has ap- in embryos. During development, many axons reach their targets peared in the mammalian phylogeny during evolution (16, 28). by following guidance cues that are organized as pathways in the embryonic substrate, and the overall pattern of the major axon Ontophyletics: An embryological approach to tracts in the adult can be traced back to the fundamental pattern evolutionary questions of such substrate pathways. Embryological and comparative an- atomical studies suggest that most axon tracts, such as the anterior Special constraints operate during the evolution of those struc- commissure, have evolved by the modified use of preexisting sub- tures, such as axon tracts, that are built in complex develop- strate pathways. On the other hand, recent developmental studies mental sequences. These constraints often allow one to infer suggest that a few entirely new substrate pathways have arisen evolutionary history from a comparison of extant chains of de- during evolution; these apparently provided opportunities for the velopmental events in various organisms (29-32). Such formation of completely new axon tracts. The corpus callosum, em- which is found only in placental mammals, may be such a truly bryological analyses of evolution-"ontophyletic analyses"-fo- new axon tract. We propose that the evolution of the corpus cal- cus on the ways that extant developmental sequences limit and losum is founded on the emergence of a new preaxonal substrate channel evolution (32). Here, we present an ontophyletic anal- pathway, the "glial sling," which bridges the two halves of the em- ysis of the evolution of the corpus callosum. bryonic forebrain only in placental mammals. Clearly, the creation of the corpus callosum must have orig- inated in mutations of the genome. However, production of the Increases in the number of central nervous system (CNS) neu- mature corpus callosum involves the cooperation of the prod- rons and in the complexity of the synaptic neuropil characterize ucts of many genes (33), and some of the necessary develop- the vertebrate phylogeny (1, 2). These increases are usually mental interactions appear to be quite distant from the genome manifest as enlargements of specific CNS areas such as the cer- (34-37). Thus, those mutations that originally brought about ebellum, the tectum, and the cerebrum. Local enlargements the evolution of the corpus callosum undoubtedly modified can be attributed to focal increases in the dominant class of ver- certain complex developmental sequences, sequences played tebrate neurons-local circuit neurons-neurons with axons that out some "distance" from the transcription and translation of extend only short distances in the CNS before synapsing (1). individual proteins. Local areas of the CNS are connected by another important Which critical developmental events may have been modi- class of neurons-projection neurons-which send their axons fied during the evolution of the corpus callosum? The corpus along a few long stereotyped routes to target areas far from their callosum is an axon tract of projection neurons. Most projection cell bodies. The white matter of the vertebrate CNS is com- tracts appear to be organized along preexisting substrate path- posed largely of axon tracts of projection neurons, and many of ways-neural "highways" that guide growing axons during de- the same major axon tracts can be identified throughout the velopment. On the basis of recent embryological discoveries, vertebrates. Although the overall pattern of these major tracts we propose that the critical developmental changes underlying appears to have been conserved during evolution (2-8), there the evolution of the corpus callosum included the acquisition are some significant variations. Frequently, the compactness of of a new substrate pathway. To show that this was a likely course particular tracts varies, as in the spinal lemniscal tracts (2, 9- of events, we must first examine both the general role of sub- 11). In some cases, the particular stereotyped route taken by strate pathways in the development of CNS axon tracts and the homologous axons varies-e.g., the stereotyped routes of the specific constraints that substrate pathways impose on evolu- corticospinal tracts differ among most primates, ungulates, ro- tionary changes in nervous systems. dents, and marsupials (12-14). Moreover, a few major vertebrate axon tracts have appeared with no apparent precursors. Substrate pathways The most dramatic example is the (dorsal) corpus callosum, which is found only in placental mammals (15, 16). Axon guidance pathways appear to be important factors in or- The corpus callosum is a large interhemispheric commissure, ganizing the overall layout of axon tracts of projection neurons, and most axons of the corpus callosum interconnect homotopic such as the neurons of the two major cortical commissures-the neocortical areas (17-22). (See refs. 23-25 for exceptions to this corpus callosum and the anterior commissure (38-54). We have general rule.) Monotremes and marsupials have no dorsal cor- called these stereotyped axon pathways "substrate pathways." Descriptions of the normal development of the CNS have shown The publication costs of this article were defrayed in part by page charge that most vertebrate axon tracts are normally organized along payment. This article must therefore be hereby marked "advertise- ment" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviation: CNS, central nervous system. 5936 Downloaded by guest on September 26, 2021 Evolution: Katz et al. Proc. Natl. Acad. Sci. USA 80 (1983) 5937 the processes of nonneuronal substrate cells: the primitive glia, continue growing along the external capsules and thence to their the radial glia, and the developing ependyma of the neural tube terminations in the contralateral cortex. (43-54). (For a similar observation in amphioxus, see ref. 55.) The substrate pathway underlying the anterior commissure These descriptions suggest that the earliest substrate pathways is probably a common feature of the vertebrate ontogeny, be- followed by CNS axon tracts may be formed, at least in part, cause the anterior commissure is a constant feature of the ver- by ordered sets of nonneuronal cell processes. tebrate forebrain (2, 3, 15, 26, 66-75). In mammals lacking a The substrate pathways of the CNS seem to act as common corpus callosum (nonplacental mammals), the anterior com- highways between the cell bodies (NUC in Fig. 1) of projection missure increases in size as the neocortex increases (15, 28). neurons and the axon target areas (TA in Fig. 1) (42). During This suggests that during evolution an increasing number of development, those axons that will form the axon tracts usually axons have used the same underlying anterior commissure sub- grow from groups of immature neurons clustered in repro- strate pathway. We propose that the comparative ontogeny of ducible locations (NUC in Fig. 1) along the intermediate zone the anterior commissure reveals one of the major mechanisms of the developing neural tube (56-59). The first pioneer axons of axon tract evolution. An increased (or a decreased) use of a grow out singly or in small bundles along the early substrate constant preexisting substrate pathway appears to be a common pathways formed by the peripheral processes of particular non- mode of axon tract evolution, as can also be seen in the phy- neuronal substrate cells (47-54, 60), and the early substrate logenies of the corticospinal tracts (76) and the dorsal columns pathways are probably equivalent to the first marginal zones of (10). On the other hand, this mechanism cannot account for the the CNS (54). Subsequent axons appear to fasciculate along evolution of completely new axon tracts. What mechanisms may preexisting axons, and in the developing tracts new axons con- have been responsible for the evolution of those axon tracts, tinue to be added in successive layers to the preexisting bundles such as the corpus callosum, that have emerged without ob- (46-51, 60-63). In this way, the longitudinal pattern of the CNS vious antecedents? axon tracts is largely determined by the pattern of the early substrate pathways, while the radial pattern within an axon tract The corpus callosum develops along a new interhemispheric is largely determined by the sequence of addition of axons. Fi- substrate pathway nally, when reaching an appropriate target area (TA in Fig. 1), an axon probably follows local chemical cues and turns from the Recent embryological studies have shown that the first step in substrate pathway that it has been following (64, 65). the formation of the corpus callosum is a fusion of the two cerebral hemispheres along the midline rostral to the lamina termi- The anterior commissure develops along an ancient nalis. The early callosal axons cross at a discrete point where the vertebrate substrate pathway growing cerebral hemispheres have made this secondary fusion (48). This special crossing point offers another route for inter- The development of the anterior commissure illustrates the de- hemispheric axon connections, apparently independent from velopment of a typical CNS substrate pathway. The anterior nearby primary midline structures such as the lamina termina- commissure, the primal cortical commissure for the verte- lis. Evidence for this independence comes from the following brates, connects the right and the left halves of the brains of observation. The formation of the corpus callosum can be com- all vertebrates (refs. 2, 3, 15, 27, and 66-75; Fig. 3). (A well- pletely blocked by surgically disrupting the secondary point of defined anterior commissure has, however, not been described interhemispheric contact in precallosal (embryonic days 15 and in cartilagenous fishes.) Axons growing in the anterior com- 16) mice. Yet this operation has no effect on other midline commissure originate from and terminate in many cortical areas, missures such as the anterior commissure and the hippocampal especially the olfactory areas, the temporal neocortex (in mam- commissure (48). Moreover, the genetic independence of the mals), and the hippocampal and amygdaloid areas (3, 15, 66- corpus callosum is suggested by the studies of D. Wahlsten of 75). Many axons growing through the anterior commissure are the University of Waterloo (personal communication), who has decussating axons and not true commissural axons. found that these other midline commissures do not vary from In placental mammals, the first anterior commissure axons normal size in acallosal mutant mice. grow out before axons of the corpus callosum (48). The cortical How are callosal axons guided across this special interhemi- anterior commissure axons grow ventrally, following the de- spheric crossing point? Before any callosal axons have reached veloping external capsules, and then grow toward the lamina the midline of the brain (embryonic day 17 in the mouse), a terminalis. In the lamina terminalis, preexisting glial endfeet particular population of nonneuronal substrate cells has accu- form aligned intercellular channels. This suggests that the end- mulated at either side of the secondary interhemispheric con- feet of the local radial glia may constitute a special preaxonal tact (48). These cells are small (each soma is about 5 ,tm in di- substrate pathway that has been constructed before the first ameter) and spider-shaped, with a relatively small cytoplasm- anterior commissure axons arrive at the midline (embryonic day to-nucleus ratio. Their cytoplasm is packed with rough endo- 15 in the mouse) (48). The anterior commissure axons then fol- plasmic reticulum and polysomes, and there are few intracel- low this substrate pathway through the lamina terminalis and lular filaments. These cells never develop neurites and they cross the midline into the contralateral cortices (48). There, they disappear neonatally: thus, they appear to be a transient class of glia. In fact, the cells closely resemble the immature glia pre- NUC viously described in the mammalian corpus callosum (77, 78). TA TA However, these cells do not stain with the Cajal gold stain for ..j astrocytes. On embryonic day 16 in the mouse, these glia mi- fre--j--t- grate across the secondary interhemispheric fusion and form an interwoven cellular bridge between the two hemispheres (Fig. 2 shows a comparable stage in the rat). This bridge [discovered by Silver (48, 79-81)] takes the form of a tightly intermeshed NUC mat of cells and cell processes, all of which are tacked together FIG. 1. CNS substrate pathways act like highways, conducting ax- by many puncta adhaerens junctions. This special transient em- ons between their cell bodies in particular nuclei (NUC) and their syn- bryological structure has been called the "glial sling." The first aptic sites in certain target areas (TA). callosal axons cross the interhemispheric fissure by crawling along Downloaded by guest on September 26, 2021 5938 Evolution: Katz et al. Proc. Natl. Acad. Sci. USA 80 (1983) the dorsal surface of the glial sling. The glial sling, then, ap- OPOSSUM MOUSE pears to be a distinct substrate pathway of nonneuronal cells that guides pioneer axons of the corpus callosum. On about embryonic day 17 in the mouse bundles of callosal axons first reach the edges of the hemispheres and begin to grow along the surface of the glial sling (Fig. 2 shows a comparable stage in the rat). Later axons fasciculate along the earlier axons. In rodents, large numbers of axons are subsequently added ros- trally and still later (embryonic days 18 and 19 in the mouse) they are added caudally. In humans, it appears that yet another midline fusion forms the massa commissuralis just dorsal and caudal to the lamina terminalis, and many of the later human callosal axons grow through this fusion (82). The glial sling is a transient embryonic structure. It disappears soon after birth, and postnatal callosal axons are guided by fasciculating along the preexisting callosal axons (48). The glial sling appears to represent the earliest guide for cal- 4 losal axons between the two hemispheres. In accord with other early CNS substrate pathways, the glial sling is composed of a special class of nonneuronal substrate cells and guides projection axons (in this case, callosal axons) over significant dis- FIG. 3. Diagram of two major cortical commissures in an embry- tances. Experimental evidence indicates that the formation of onic marsupial (opossum; pouch day 18) and an embryonic placental a dorsal corpus callosum is integrally dependent on the creation mammal (mouse; embryonic day 18). All vertebrates have an anterior of this particular substrate pathway (48, 81). How has this im- commissure, routed through the lamina terminalis. Only placental mammals have a corpus callosum, which is founded on a new substrate portant developmental structure figured in the evolution of the pathway. This new substrate pathway, the glial sling, forms a second- corpus callosum? The glial sling appears to be a structure com- ary midline fusion of the cerebral hemispheres rostral to the lamina mon to the ontogenies of many mammals. [Although there has terminalis. (Upper) Oblique horizontal sections at the level of the lam- been some question about its presence in the rat (83), Fig. 2 ina terminalis, which liesjust rostral to the third ventricle. (Rostral is to the top; caudal is to the bottom.) (Lower) Midline sagittal sections through the forebrain. (Rostral is to the left; caudal is to the right; dorsal is to the top.) ac, Developing anterior commissure; cc, developing corpus callosum; mI, third ventricle.

clearly shows that the embryonic rat has a glial sling.] Perhaps those animals that lack a corpus callosum have ontogenies that also lack the underlying callosal substrate pathway. In particular, monotremes and marsupials do not have a dorsal corpus callosum. Do these animals have a glial sling? The opossum and the evolution of the corpus callosum To answer this question, we (see also ref. 48) examined histological sections of the brain of a developing marsupial, the opossum. By pouch day 18, the anterior commissure was well formed but the cerebral hemispheres still had not contacted in the midline (Fig. 3). The cerebral hemispheres continue to re- main separate throughout the life of the opossum and never fuse along the midline rostral to the lamina terminalis (26). In addition, there is no indication that groups of glial cells collect anywhere along the midline edges of the hemispheres. Thus, two critical steps necessary for the formation of a dorsal corpus callosum do not occur in the opossum: (i) There is no fusion of the cerebral hemispheres along the midline rostral to the lamina terminalis. (ii) There is no formation of a glial sling. These observations lead us to hypothesize that the development of an interhemispheric glial sling was one of the critical steps in the acquisition of a corpus callosum by placental mammals.* We propose that during evolution a small number of critical genomic changes caused a particular group of substrate FIG. 2. Coronal sections through the forebrain ofan embryonic rat (Long-Evans), showingthematofglial cells (the "glialsling") that guides * Histological observations of the acallosal mutant mouse are consistent thefirstcallosal axons as they grow across thedorsal surface ofthe sep- with our hypothesis that, besides interhemispheric fusion, a glial sling tum. (Upper) Embryonic day 18, before callosal axons have grown across is also necessary for the formation of a corpus callosum (48). Although themidline. (Lower) Embryonicday 19, whenmany callosal axons have there is fusion of the midline in these mutants, the glial sling is missing grown alongthe upper surface ofthe sling. CC, callosal axons; GS, glial and the normal callosal axons cross through other commissures, follow sling, LF, longitudinal fissure between the two hemispheres ofthe fore- preexisting ipsilateral pathways, or form a neuroma when they reach brain; S, septum; V, lateral ventricle. (Ten-micrometer paraffin sec- the midline. Identical results were found after surgical disruption of tions, hematoxylin stain, x70.) the glial sling in otherwise normal mouse embryos. Downloaded by guest on September 26, 2021 Evolution: Katz et al. Proc. Natl. Acad. Sci. USA 80 (1983) 5939 cells to form the interhemispheric glial sling. The glial sling then erate during development to produce concordance among the offered a new substrate pathway to both new and existing axons: many disparate developmental events (30-32, 89). those axons that took advantage of the newly available glial sling One example of the ontogenetic buffer mechanisms nor- became the corpus callosum. Once the glial sling became an mally at work during the development of the corpus callosum established feature of the ontogenies of placental mammals, is "selective axon elimination." During development, callosal further evolution of the corpus callosum occurred in a typical axons grow to most areas of the neocortex. As an animal ma- fashion for axon tracts. As the size of the neocortex increased, tures, many of these axons are selectively eliminated from cer- the size of the corpus callosum increased-with the additional tain target areas (34, 90-92). At least some cortical neurons which axons following the preexisting callosal substrate pathway (16). selectively eliminate a callosal (a contralateral) projection during development maintain another projection to the ipsilateral Glial sling mutant cortex (92). It appears that cortical neurons initially send axons to several different cortical target sites and then selectively A mutant mouse in which the corpus callosum fails to develop eliminate those axons that are not appropriately (perhaps, func- provides further support for our proposed evolutionary scenar- tionally) matched (62, 93-95).§ io. Wahlsten has shown that the BALB/c strain of mice con- Ontogenetic buffer mechanisms allow the organism to either tains a recessive incompletely penetrant mutation (apparently integrate or eliminate developmental variants (89), variants that multifactorial) affecting the corpus callosum (33). Complete inevitably arise during complex developmental sequences. In expression is agenesis of the corpus callosum, and this callosal this way, buffer mechanisms can help to ensure the accurate defect appears to be the primary CNS lesion (48, 84). One of development of the complex neural circuitry necessary for the the earliest defects appearing in the acallosal mutant mouse is normal functioning of many organisms. Some evolutionary in- the failure of glia to collect along the medial walls of the form- novations such as the glial sling probably represented devel- ing hemispheres (48). Subsequently, no glial sling forms. Agen- opmental variants when they first appeared in the ancestors of esis of the corpus callosum is also a well-documented condition the placental mammals. The use of the glial sling as a key de- in humans (85-87). In such cases, callosal axons form thick velopmental step in building a corpus callosum illustrates how aberrant bundles running ipsilaterally and longitudinally along ontogenetic buffer mechanisms can sometimes integrate such the roofs of the lateral ventricles. These aberrant axons are known developmental variants into viable and ultimately quite suc- as the bundles of Probst. In experimental animals, bundles of cessful ontogenies. Probst can be produced artificially by disrupting the glial sling early in its development (48). Bundles of Probst are also found Conclusions in the acallosal mutant mouse (48, 81, 84). The acallosal mutation appears to selectively delete the glial Embryological studies indicate that axons and substrate path- sling while leaving the callosal axons intact. Thus, one impor- ways are the natural developmental units underlying axon tracts. tant inference is that the glial sling is in many ways an inde- Evolution of axon tracts can occur through changes in either of pendent ontogenetic unit, developmentally separate from the these units. The two major forebrain commissures of mam- axons that use it as a substrate pathway. The successful iden- mals-the anterior commissure and the corpus callosum-pro- tification of mutations that specifically affect critical develop- vide good demonstrations of both routes of axon tract evolu- mental units is the first step toward unpuzzling those genomic tion. The anterior commissure appears to have evolved through changes that originally gave rise to the corpus callosum during the use of the same ancient substrate pathway by new axons. the evolution of placental mammals. On the other hand, the corpus callosum probably first evolved via the emergence of an entirely new substrate pathway, the The ontogenetic integration of genetic mutations glial sling. During subsequent evolution of the corpus callosum, additional axons have then used this same underlying A special developmental structure-the glial sling-appears to substrate pathway. be one of the critical evolutionary innovations underlying the In general, the extant pattern of available substrate pathways creation of the corpus callosum. In addition, however, there will always impose a fundamental constraint on the evolution must be many other concordant integrated changes throughout of axon tracts. Unless new substrate pathways arise, new axons development. To create a functional corpus callosum it is not must follow preexisting substrate pathways. These substrate sufficient merely to create a new substrate pathway bridging pathways interconnect only certain predetermined areas, and the two cerebral hemispheres. Axons must also be guided to the thus not all neural populations can be directly connected, even new substrate pathway, and later the callosal axons must form by new axons. For example, it appears that the eye and the cer- appropriate synapses. What generates all of the additional co- ebellum cannot be directly interconnected by using extant sub- ordinated developmental changes that are necessary to make a strate pathways. Nonetheless, new substrate pathways, such as functional corpus callosum? the glial sling, may occasionally evolve, permitting direct and On the one hand, the particular genomic changes respon- specific communication between completely new areas of the sible for the secondary midline fusion and for the glial sling may nervous system. In this way, the glial sling is an example of a be pleiotropic (88), perhaps bringing about certain additional truly fundamental evolutionary change in the topology of the coordinated changes during development. t On the other hand, CNS axon tracts. the critical evolutionary mutations may have produced effects We thank V. S. Caviness, Jr., F. F. Ebner, G. M. Innocenti, and G. that must be secondarily integrated into the developmental se- 0. Ivy for critically reading this manuscript and for suggesting useful quence. f This secondary integration could be brought about by changes. F. F. Ebner also provided information about the comparative various ontogenetic buffer mechanisms, which normally op- anatomy of the forebrain commissures. R. G. Northcutt advised us on the neuroanatomy of the anterior commissure. D. Wahlsten instructed t Concordant pleiotropic effects of a mutation produce what we have called a "type I evolutionary change" (29). § The existence of this ontogenetic buffer mechanism-widespread tar- t Preexisting ontogenetic buffer mechanisms can sometimes integrate get sampling coupled with selective axon elimination-may have been potentially discordant mutations into functional evolutionary changes, a prerequisite for the initial formation of the corpus callosum along the producing what we have called "type II evolutionary changes" (29). phylogenetically new substrate pathway, the glial sling. Downloaded by guest on September 26, 2021 5940 Evolution: Katz et al. Proc. Natl. Acad. Sci. USA 80 (1983)

us on the genetics-of acallosal mutant mice. G. 0. Ivy pointed out the 47. Silver, J. & Sapiro, J. (1981)J. Comp. Neurol. 202, 521-538. possible evolutionary importance of target sampling by developing ax- 48. Silver, J., Lorenz, S. E., Wahlsten; D. & Coughlin, J. (1982) J. ons. This work was supported by grants from the National Institutes of Comp. Neurol. 210, 10-29. Health and from the Whitehall Foundation. 49. Nordlander, R. H., Singer, J. F., Beck, R. & Singer, M. (1981)J. Comp. Neurol. 199, 535-552. 1. Rakic, P. (1976) Local Circuit Neurons (MIT Press, Cambridge, 50. Nordlander, R. H. & Singer, M. (1982) Exp. Neurol. 75, 221-228. MA). 51. Nordlander, R. H. & Singer, M. (1982) Dev. Brain Res. 4, 181- 2. Sarnat, H. B. & Netsky, M. (1974) Evolution of the Nervous Sys- 193. tem (Oxford Univ. Press, New York). 52. Holley, J. A., Nornes, J. 0. & Morita, M. (1982)J. Comp. Neurol. 3. Herrick, C. J. (1948) The Brain of the Tiger Salamander (Univ. 205, 360-370. of Chicago Press, Chicago). 53. Holley, J. A. (1982)J.,Comp. Neurol. 205, 371-382. 4. Nieuwenhuvs, R. (1967) Prog. Brain Res. 23, 1-64. 54. Kevetter, G. A. & Lasek, R. J. (1982) Dev. Brain Res. 4, 195-208. 5. Crosby, E. C. (1969) in Neurobiology of Cerebellar Evolution and 55. Bone, Q. (1960)J. Comp. Neurol. 115, 27-64. Development, ed. Llinas, R. (Am. Med. Assoc., Chicago), pp. 19- 56. Wenger, E. K. (1950)J. Exp. Zool. 114, 51-85. 41. 57. Bergquist, H. & Kallen, B. (1954)J. Comp. Neurol. 100, 627-659. 6. Nauta, W. J. H. & Karten, H. J. (1970) in The Neurosciences: Sec- 58. Kallen, B. (1977) in Handbook of Clinical Neurology, eds. Vin- ond Study Program, ed. Schmitt, F. 0. (Rockefeller Univ. Press, ken, P. J. & Bruyn, B. W. (North-Holland, Amsterdam), Vol. 30, New York), pp. 7-26. pp. 41-83. 7. Nieuwenhuys, R. (1972) J. Comp. Neurol. 145, 165-178. 59. Nornes, H. 0. & Cary, M. (1978) Brain Res. 159, 1-16. 8. Ebbesson, S. 0. E. (1980) Cell Tissue Res. 213, 179-212. 60. Sapiro, J., Silver, J. & Singer, M. (1980) Invest. Opthalmol. Vis. 9. Herrick, C. J. & Bishop, G. H. (1958) in Reticular Formation of Sci. 19, 247. the Brain, eds. Jasper, H. H. & Proctor, L. D. (Little, Brown, 61. Scholes, J. H. (1979) Nature (London) 278, 620-624. Boston), pp. 353-360. 62. Nornes, H. O., Hart, H. & Cary, M. (1980)J. Comp. Neurol. 192, 10. Norton, A. C. (1969-1970) The Dorsal Column System of the Spinal 119-132. Cord (UCLA Brain Inf. Serv., Los Angeles), pp. 82-107. 63. Easter, S. S., Jr., Rusoff, A. C. & Kish, P. E. (1981)J. Neurosci. 11. Jane, J. A. & Schroeder, D. M. (1971) Exp. Neurol. 10, 1-17. 1, 793-811. 12. Schoen, J. H. R. (1964) Prog. Brain Res. 11, 203-222. 64. Attardi, D. G. & Sperry, R. W. (1963) Exp. Neurol. 7, 46-64. 13. Phillips, C. G. & Porter, R. (1977) Corticospinal Neurones (Aca- 65. Fujisawa, H., Tani, N., Watanabe, K. & Ibata, Y. (1982) Dev. Biol. demic, New York). 90, 43-57. 14. Watson, C. R. R. & Freeman, B. W. (1977) Brain Behav. Evol. 14, 66. Kuhlenbeck, H. (1978) The Central Nervous System of Verte- 341-351. brates (Karger, Basel, Switzerland), Vol. 5, Part 2, pp. 384-478. 15. Ariens Kappers, C. U., Huber, C. G. & Crosby, E. C. (1936) The 67. Johnston, J. B. (1902) J. Comp. Neurol. 12, 1-86. Comparative Anatomy of the Nervous System of Vertebrates, In- 68. Allison, A. C. (1953) Biol. Rev. 28, 195-244. cluding Man (Macmillan, New York). 69. Faucett, J. R. (1969)J.Comp. Neurol. 137, 377-406. 16. Abbie, A. A. (1939) J. Comp. Neurol. 70, 9-44. 70. Fox, C. A. M., Fisher, R. P. & Desalva, S. J. (1948)J. Comp. Neu- 17. Jacobson, S. (1965)J. Comp. Neurol. 124, 131-146. rol. 89, 245-277: 18. Karol, E. A. & Pandya, D. N. (1971) Brain 94, 471-486. 71. Ebner, F. F. & Myers, R. E. (1965) J. Comp. Neurol. 124, 353- 19. Innocenti, G. M., Manzoni, T. & Spidalieri, G. (1974) Exp. Brain 366. Res. 19, 447-466. 72. Crosby, E. C., DeJonge, B. R. & Schneider, R. C. (1966) in Evo- 20. Yorker, C. H., Jr., & Caviness, V. S., Jr. (1975)J. Comp. Neurol. lution of the Forebrain, eds. Hassler, R. & Stephan, H. (Thieme, 164, 223-246. Stuttgart, FRG), pp. 117-135. 21. Goldman, P. S. & Nauta, W. J. H. (1977) Brain Res. 122, 393-413. 73. Pandya, D. N., Karol, E. A. & Lele, P. P. (1973) Brain Res. 49, 22. Gould, H. J. & Ebner, F. F. (1978) J. Comp. Neurol. 177, 473- 177-180. 502. 74. Kuhlenbeck, H., Malewitz, T. D. & Beasley, A. B. (1966) in Evo- 23. Jones, E. G. & Powell, T. P. S. (1968)J. Anat. 103, 433-455. lution of the Forebrain, eds. Hassler, R. & Stephan, H. (Thieme, 24. Keller, G. & Innocenti, G. M. (1981) Neuroscience 6, 703-712. Stuttgart, FRG), pp. 9-19. 25. Hedreen, J. C. & Yin, T. C. T. (1981) J. Comp. Neurol. 197, 605- 75. Crosby, E. C. & Schnitzlein, H. N. (1982) Comparative Correl- 621. ative Neuroanatomy of the Vertebrate Telencephalon (Macmil- 26. Tilney, F. (1938) Bull. Neurol. Inst. N.Y. 7, 1-77. lan, New York). 27. Ebner, F. F. (1969) Ann. N.Y. Acad. Sci. 167, 241-257. 76. Towe, A. L. (1973) Brain Behav. Evol. 7, 1-17. 28. Heath, C. J. & Jones, E. G. (1971)J. Anat. 109, 253-270. 77. Paterson, J. A., Privat, A., Ling, E. A. & Leblond, C. P. (1973)J. 29. Katz, M. J. & Lasek, R. J. (1978) Proc. Natl. Acad. Sci. USA 75, Comp. Neurol. 149, 83-102. 1349-1352. 78. Sturrock, R. R. (1976)J. Anat. 122, 521-537. 30. Katz, M. J., Lasek, R. J. & Kaiserman-Abramof, I. R. (1981) Proc. 79. Silver, J. (1980) Soc. Neurosci. Abstr. 6, 487. Natl. Acad. Sci. USA 78, 397-401. 80. Silver, J. (1981) Soc. Neurosci. Abstr. 7, 548. 31. Katz, M. J. (1982) in Evolution and Development, ed. Bonner, J. 81. Silver, J. & Ogawa, M. Y. (1983) Science 220, 1067-1069. T. (Springer, Berlin), pp. 207-212. 82. Rakic, P. & Yakovlev, P. I. (1968) J. Comp. Neurol. 132, 45-72. 32. Katz, M. J. (1983) Perspect. Biol. Med. 26, 323-333. 83. Valentino, K. L., & Jones, E. G. (1982)J. Neurocytol. 4, 583-609. 33. Wahlsten, D. (1982)J. Hered. 73, 281-285. 84. Wahlsten, D. (1982) Brain Res. 239, 329-347. 34. Innocenti, G. M; (1981) Science 212, 824-827. 85. Slager, U. T., Kelly, A. B. & Wagner, J. A. (1957) N. Engl.J. Med. 35. Caviness, V. S., Jr., & Yorke, C. H., Jr. (1976) J. Comp. Neurol. 256, 1171-1176. 170, 449-460. 86. Loesser, J. D. & Alvord, E. C., Jr. (1968) Brain 91, 553-570. 36. Caviness, V. S., Jr., & Rakic, P. (1978) Annu. Rev. Neurosci. 1, 297- 87. Friede, R. L. (1975) Developmental Neuropathology (Springer, 326. New York), pp. 287-292. 37. Goldman-Rakic, P. S. (1981) in The Organization of the Cerebral 88. Stern, C. (1955) in Analysis of Development, eds. Willier, B. H., Cortex, ed. Schmitt, F. 0. (MIT Press, Cambridge, MA), pp. 69- Weiss, P. A. & Hamburger, V. (Saunders, Philadelphia), pp. 151- 97. 169; reprinted (1971) by Hafner, New York. 38. Kallen, B. (1954) J. Embryol. Exp. Morphol. 2, 87-100. 89. Katz, M. J. & Grenander, U. (1982)J. Theor. Biol. 98, 501-517. 39. Katz, M. J. & Lasek, R. J. (1979)J. Comp. Neurol. 183, 817-832. 90. Ivy, G. 0. & Killackey, H. P. (1981) J. Comp. Neurol. 195, 367- 40. Katz, M. J. & Lasek, R. J. (1980) Cell Motility 1, 141-157. 389. 41. Katz, M. J., Lasek, R. J. & Nauta, W. J. H. (1980) Neuroscience 91. O'Leary, D. D. M., Stanfield, B. B. & Cowan, W. M. (1981) Dec. 5, 821-833. Brain Res. 1, 607^-617. 42. Katz, M. J. & Lasek, R. J. (1981) J. Comp. Neurol. 195, 627-641. 92. Ivy, G. 0. & Killackey, H. P. (1982)J. Neurosci. 2, 735-743. 43. Simpson, S. B., Jr. (1968)J. Comp. Neurol. 134, 193-210. 93. Innocenti, G. M. & Frost, D. 0. (1979) Nature (London) 280, 231- 44. Singer, M., Nordlander, R. H. & Egar, M. (1979)J. Comp. Neu- 234. rol. 185, 1-22. 94. Innocenti, G. M. & Frost, D. 0. (1980) Exp. Brain Res. 39,365-375. 45. Silver, J. & Robb, R. M. (1979) Dev. Biol. 68, 175-190. 95. Rhoades, R. W. & Dellacroce, D. D. (1980) Brain Res. 202, 189- 46. Silver, J. & Sidman, R. L. (1980)J. Comp. Neurol. 189, 101-111. 195. Downloaded by guest on September 26, 2021