In the developing brain, the phenomenon of neurogenesis is manifested heterotopically, that is, much the same neurogenetic steps occur at different places with a different timetable. This is due apparently to early molecular regionalization of the neural tube wall in the anteroposterior and dorsoventral dimensions, in a checkerboard pattern of more or less deformed quadrangular histogenetic areas. Their respective fate is apparently specified by a locally specific combination of active/repressed genes known as “molecular profile.” This leads to position-dependent differential control of proliferation, neurogenesis, differentiation, and other aspects, eventually in a heterochronic manner across adjacent areal units with sufficiently different molecular profiles. It is not known how fixed these heterochronic patterns are. We reexamined here comparatively early patterns of forebrain and hindbrain neurogenesis in a lizard (Lacerta gallotia galloti), a bird (the chick), and a mammal (the rat), as demonstrated by activation of acetylcholinesterase (AChE). This is an early marker of postmitotic neurons, which leaves unlabeled the neuroepithelial ventricular cells, so that we can examine cleared wholemounts of the reacted brains to have a birds-eye view of the emergent neuronal pattern at each stage. There is overall heterochrony between the basal and alar plates of the brain, a known fact, but, remarkably, heterochrony occurs even within the precocious basal plate among its final anteroposterior neuromeric subdivisions and their internal microzonal subdivisions. Some neuromeric units or microzones are precocious, while others follow suit without any specific spatial order or gradient; other similar neuromeric units remain retarded in the midst of quite advanced neighbors, though they do produce similar neurogenetic patterns at later stages. It was found that some details of such neuromeric heterochrony are species-specific, possibly related to differential morphogenetic properties. Given the molecular causal underpinning of the updated prosomeric model used here for interpretation, we comment on the close correlation between some genetic patterns and the observed AChE differentiation patterns.

© 2022 S. Karger AG, Basel

Introduction

Developmental heterochrony refers to comparable mechanisms or patterns of development that occur displaced in relative time. It applies to phylogeny, when comparing different species in evolutionary context, but also to ontogeny, when the same process or pattern occurs at different ontogenetic times at different positions (e.g., in the brain). The latter use accordingly examines ontogenetic heterotopic heterochrony (position-related temporal displacement of a particular developmental process). The embryonic brain shows many instances of heterotopic heterochrony due to its complex anteroposterior and dorsoventral regionalization into areal neuroepithelial progenitor domains (called fundamental morphogenetic units) [Nieuwenhuys and Puelles, 2016; Nieuwenhuys, 2017]. It is common knowledge now that each of these has an unique developmental molecular profile that leads them to regulate independently their proliferative and differentiative processes according to the local combination of active/inactive genes. Sharing of significant genes among adjacent serially transverse morphogenetic units apparently underlies metameric repeating of an histogenetic pattern along a series of units (e.g., forming plurisegmental patterns that generate multimodular sensory columns) [Marin et al., 2008; Puelles, 2013; Tomás-Roca et al., 2016]. Examined across species, the map of molecularly defined brain fundamental morphogenetic units is known to show considerable evolutionary conservatism, defining the so-called ontogenetic brain Bauplan. In this, fixed component units may vary in relative size and observable histogenetic properties over time but keep their neighbors or topologic boundaries. It is less clear whether the heterochronic neuronal differentiation patterns are also stereotyped across the shared Bauplan, irrespective of the fact that the speed and length of development may change in different species.

The present research examines heterochronic spatial patterns in the topology of emerging acetylcholinesterase (AChE)-positive young postmitotic neurons in the sauropsidian (chick and lizard) and mammalian (rat) forebrain and hindbrain. We will reexamine and partly reinterpret previously published chick and rat material [Puelles et al., 1987a, 2015a], adding unpublished chick forebrain and rat hindbrain results from Amat’s [1986] doctoral thesis, and including as well some hitherto unpublished results from our AChE studies on lizard embryos, done in collaboration with C.M. Trujillo. Emphasis will be placed on the chicken pattern.

The present work reflects a talk on “neurogenetic heterochrony” given at the Karger Symposium in 2020. Our comparative embryonic AChE material was interpreted within the updated prosomeric model [Puelles et al., 2012a; Puelles, 2013; Puelles and Rubenstein, 2015; Puelles, 2018], thus correcting some errors found in our previous publications.

The Puelles et al. [1987a] study of chicken early whole­mount patterns of AChE-labeled newborn neurons employed a then completely heterodox neuromeric approach which was largely based on the previous neuromeric embryologic work of von Kupffer [1906], Palmgren [1921], Rendahl [1924], Bergquist and Kallen [1953, 1954], Coggeshall [1964], Vaage [1969, 1973], Keyser [1972], and Gribnau and Geijberts [1985].

This original model was largely developed by the so-called “Nordic school,” though there were earlier cogent notions on neuromeres since Orr [1887] (work on lizard embryo brains). Except for the German von Kupffer, the authors cited above worked in Sweden, Norway, Denmark, and Holland (thus the “Nordic school”); they apparently were directly or indirectly inspired by the Swedish comparative brain histologist and embryologist Niels Holmgren, who supervised the theses of Palmgren, Rendahl, and Bergquist, among others. However, they all tended to work independently [Källén, personal communication to L.P. in 2002]. This model was tentatively applied to diencephalic development by L.P. as of 1977 due to previous failure to explain various sorts of developmental Golgi and AChE data using the standard columnar model of Herrick [1910, 1933, 1948] and Kuhlenbeck [1973].

The slightly modified version of the Nordic neuromeric model offered in Puelles et al. [1987a] was the immediate antecedent of the now well-known prosomeric model, produced subsequent to a complementary gene mapping approach in mouse embryos done in collaboration by L.P. and J.L.R. Rubenstein since 1992 [Bulfone et al., 1993; Puelles and Rubenstein, 1993; Rubenstein et al., 1994; Bulfone et al., 1995; Shimamura et al., 1995; Puelles, 1995]. This model was soon tested in amniote and anamniote model vertebrates including agnatha [Pombal et al., 2009], as well as the cephalochordate Amphioxus [Albuixech-Crespo et al., 2017]. In the latter species a number of brain Bauplan components normally present in vertebrates are underdeveloped or absent.

Updates of the prosomeric model later appeared in Puelles and Rubenstein [2003], Puelles et al. [2004, 2012a, b], Puelles [2001, 2013], Puelles and Rubenstein [2015], and Puelles [2018]. These molecular studies logically emphasized the description of given gene expression domains consistent with a neuromeric interpretation, providing evidence for a molecularly and causally underpinned Bauplan of the neuraxis, rather than delving on the epiphenomenal neuronal differentiation patterns.

The Updated Prosomeric Model

The prosomeric model has evolved in recent years, particularly in its hypothalamo-telencephalic and hindbrain parts, and has become better substantiated molecularly and experimentally than the simpler model used by Puelles et al. [1987a].

The axial dimension of the neural tube is defined first. Five universal longitudinal zones of reference are present in all vertebrate brains: (1) the floorplate (which is induced up to a rostral end under the mamillary area by the axial notochord [Puelles et al., 2012b; Puelles and Rubenstein, 2015; Puelles, 2018], (2) the roofplate (resulting from the median fusion of the medullary folds limiting peripherally the initial neural plate primordium); its rostral end lies at the anterior commissure [Puelles et al., 1987b; Cobos et al., 2001], (3) the molecular alar-basal boundary (which emerges throughout the neural tube due to an early equilibrium generated by antagonistic dorsoventral interplay of ventralizing floor morphogens against dorsalizing roof morphogens [see Puelles et al., 2012a; Puelles and Rubenstein, 2015; Puelles, 2018]. These first three longitudinal landmarks of the neural wall form at three different dorsoventral positions, and we understand now fairly well their causal mechanisms. They are roughly parallel to each other, and, moreover, they co-define the equally longitudinal but wider (4) basal plate and (5) alar plate of the lateral wall, where most neurogenesis occurs (Fig. 1a).

Fig. 1.

Schemata illustrating three stages in progressive anteroposterior subdivision of the updated prosomeric model. Red = acroterminal domain; green = prospective telencephalic roofplate; blue = hindbrain central field where the overt rhombomeres r2–r6 are seen; light gray = floorplate; dark gray = notochord. a Earliest division into forebrain, hindbrain, and spinal cord tagmata. The transverse isthmic boundary separating the forebrain and hindbrain tagmata is drawn as a double-thick black line. b Subsequent proneuromere subdivisions of the forebrain and hindbrain tagmata; the forebrain tagma divides into secondary prosencephalon (including hypothalamus, telencephalon and optic vesicle), diencephalon, and midbrain proneuromeres; the hindbrain tagma divides into prepontine, pontine, retropontine, and medullary proneuromeres. The added dividing boundaries are marked as thick black lines. c Final neuromeric subdivisions of the proneuromeres. There are hypothalamo-telencephalic prosomeres hp1 and hp2 (SP), diencephalic prosomeres dp1–dp3, and midbrain prosomeres (mesomeres) m1 and m2, all delimited by thin black transverse boundaries orthogonal to the longitudinal zones. In the hindbrain there appear prepontine rhombomeres r0 and r1, pontine units r2–r4, retropontine rhombomeres r5 and r6, and medullary units r7–r11. The schema also indicates which rhombomeres are overt or cryptic, all of them being functionally equivalent. Regionalization increases merely by addition of novel boundaries, always keeping the earlier ones. Note that embryos at the tagmatic or proneuromeric stages do not have the shape shown in the schemata a and b (copied from c), being much smaller and simpler; these schemata, so to speak, emphasize the respective fates.

The rostral end of this quintuple system of longitudinal zones reaches what Puelles et al. [2012a] first called the acroterminal domain of the forebrain (this concept was not available before). This is a transverse linear rostral end of the neural tube primordium. Indeed, the acroterminal domain has a ventrodorsal extent, and reaches from the front of the mamillary body (rostralmost floor) along the rostromedian left-right continuity of the basal and alar plates up to the anterior commissure, the rostralmost roof (acro-term; Fig. 1a). Interestingly, the acroterminal domain is selectively labeled in the mouse at E11.5, E13.5, and E15.5 by the gene Dlk1 (Allen Developing Mouse Brain Atlas; see our Fig. 15a). The acroterminal territory apparently owes part of its singular properties to its unique reception of strong inductive effects from the prechordal plate from early gastrulation stages onwards. The latter is represented by an early axial cell population of the anterior visceral endoderm. The initial static endodermal prechordal plate adheres to the neural ectoderm at the level of the prospective infundibular/tuberal hypothalamus; subsequently some prechordal plate cells undergo an epithelio-mesenchymal transformation and migrate dynamically dorsalwards in front of the acroterminal domain, moving from its floor to its roof levels [see Diaz and Puelles, 2020]. The right and left longitudinal alar-basal boundaries clearly meet at the acroterminal domain under the optic chiasma. It has been postulated [Puelles et al., 2012a; Ferran et al., 2015; Puelles, 2017; Diaz and Puelles, 2020] that the dynamic (migrating) prechordal plate cells plus the static acroterminal domain jointly represent the true source of anteroposterior patterning signals in the closed neural tube (rather than the anterior neural ridge emphasized by literature, since the latter is a roof plate locus, and therefore should have dorsalizing morphogenetic effects).

The updated prosomeric model next postulates a series of transverse (anteroposterior) segments of the neural tube observable caudal to the acroterminal domain and always topologically orthogonal to the five longitudinal zones. Typically, such AP brain parts extend from the floor to the roof of the neural tube, and thus enclose parts of all the dorsoventral longitudinal zones (establishing a metameric DV structural pattern) [Puelles and Rubenstein, 1993, 2003]. Depending of the developmental stage we may describe AP divisions as tagmata, proneuromeres, or neuromeres (Fig. 1a–c) [Puelles, 2018]. In this work, we will discuss neurogenetic heterochrony in the forebrain and hindbrain tagmata. After growth and further AP patterning, three major parts of the forebrain tagma are delimited as proneuromeres (Fig. 1b), namely the secondary prosencephalon (SP; hypothalamus plus telencephalon and eyes), the diencephalon (Di), and the midbrain (Mes). All three share details of dorsoventral patterning and there is evidence for their joint neural induction by signals from the early node. With continuing anteroposterior growth and regionalization, these proneuromeres divide each into neuromeres (Fig. 1c).

The secondary prosencephalon divides into two hypothalamo-telencephalic prosomeres (hp1, hp2; defined in caudorostral order; Fig. 1c; the hypothalamic portions of these neuromeres are named peduncular and terminal hypothalamus, respectively – PHy, THy) [Puelles et al., 2012a]. The diencephalon divides into three diencephalic prosomeres dp1–dp3 (defined in caudorostral order; often referred to as p1–p3). These contain in their alar domains, respectively, the primordia of the pretectum (PT), the thalamus (TH; plus epithalamus), and the prethalamus (PTh).

Finally, the midbrain divides into a massive rostral m1 mesomere (or mp1 midbrain prosomere) and a slender caudal m2 mesomere (or mp2; note their rostrocaudal order; Fig. 1c). The m1 contains most well-known midbrain structures like the oculomotor nucleus (3), red nucleus, and the superior and inferior colliculi (SC, IC). The tiny m2 mesomere is one of the least known parts of the brain; its existence as an atrophic domain was first postulated by Palmgren [1921; see also Vaage, 1969, 1973], but it was modernly found to have specific derivatives and a differential molecular profile [Hidalgo-Sánchez et al., 2005; Puelles et al., 2012b; Puelles, 2013].

The hindbrain tagma divides first into prepontine, pontine, retropontine, and medullary proneuromeres (PrP, P, RP, Med; Fig. 1b); these subdivide later into 12 rhombomeres (r0–r11; defined in rostrocaudal order; Fig. 1c). The PrP (or isthmocerebellar hindbrain) produces r0 and r1 (the latter can be subdivided into rostral and caudal parts), the P r2–r4, the RP r5 and r6, and the Med r7–r11 [see also Puelles et al., 2013]. Some of these rhombomeres – r2–r6 – are referred to as “overt” (meaning they are delimited by visible outer constrictions) and others – r0–r1; r7–r11 – as “cryptic” (no outer constrictions, but with demonstrable equivalent molecular limits) [Marín and Puelles, 1995; Cambronero and Puelles, 2000; Marin et al., 2008; Tomas-Roca et al., 2016].

Our present analysis reexamines heterochronic differentiation patterns in the chick forebrain and hindbrain more extensively than before in the light of the updated prosomeric model and its checkerboard pattern of AP/DV subdivisions considered as fundamental morphogenetic units, adding a glance at corresponding patterns in the lizard and rat forebrain. Shared heterochronic aspects in the developmental appearance of AChE-positive young neurons observed in the three studied amniote species are clearly consistent with the updated prosomeric model conceived as representing an evolutionarily conserved Bauplan. Such histogenetic Bauplan is held to be underpinned by an evolutionarily conserved system of neuromeric units, comparable delimiting and characterizing gene expression patterns, and related conserved patterning mechanisms throughout vertebrates [Albuixech-Crespo et al., 2017]. Nevertheless, subtle variations of heterochronic pattern were noted between the amniote species studied, which we think may underlie species-specific variant morphogenesis.

Material and Methods

Chick embryos were incubated at 37.8°C in a rotating forced-draft incubator (eggs obtained from a commercial source). This study comprises chick embryo data between stages HH11 and HH26. The specimens were first collected in saline solution, staged [Hamburger and Hamilton, 1951], and transferred into cold fixative solution (10% formaldehyde in 0.1 M, pH 7.2–7.4 phosphate buffer, with 1 mL of stock 0.5% CaCl2 solution per 100 mL). Fixation at 4°C varied between 1 h and 24 h (according to size) without significant change in AChE activity.

The size of the pieces was reduced before histochemistry in order to aid penetration of the histochemical reactives. The youngest embryos (stages 9–12) were incubated whole. Between stages 13 and 18 we separated the forebrain, cutting at the isthmus, as well as the spinal cord. The eye vesicles were discarded in most embryos. Additionally, the forebrains were divided into halves, and the covering skin and mesenchyme were dissected away using fine-tipped watchmaker forceps and sharpened tungsten needles. The limit of the wholemount method occurred after stage 26, when the thickest tissue parts showed at their center a whitish patch devoid of histochemical reaction.

After a washing in distilled water or maleate buffer, the neural tube pieces were incubated at 4°C in the medium of Karnovsky and Roots [1964] with acetylthiocholine in acetate buffer at pH 6. The medium contained 8×10–5 M iso-OMPA, to inhibit pseudo-cholinesterase. Incubation proceeded inside a refrigerator for 8–24 h (or longer, in the largest specimens), with occasional stirring. Controls incubated in the presence of 5×10–5 M BW284C51 or 10–5 M physostigmine were negative. Alternative incubation with butyril-thiocholine as a substrate yielded no significant reaction product.

Detailed study and photographic reproduction usually required that all tissues external to the neural primordium (mesenchyme, meninges) be dissected away while submerged in buffer or saline solution. We used the concavity of a Maximow culture slide to contain the fluid, placing the slide on an operating microscope equipped with an underlying light source (visualizing in this way by transparency the reaction product). An electrolytically sharpened and L-bent tungsten needle fixed to the tip of a Pasteur pipette with molten paraffine, jointly with a fine-tipped watchmaker forceps (for holding the tissue), were used to finish these dissections. After the end of this process, the specimens were dehydrated by steps in an ethanol series, and then cleared in methyl benzoate. Wholemount preparations remained in methyl benzoate for observation, photography, and permanent conservation. Some Paraplast-embedded pieces were sectioned parasagittally up to the midline, and the remaining half (with a sharp midline) was deparaffinated and returned to methyl benzoate as a wholemount.

Given that the curvature of the neural tube halves impeded focussing them whole for microphotography with ×6.3 or ×4 microscope objectives, we positioned the specimens as desired on the concave slope of the Maximow slide concavity, submerged in methyl benzoate. This setup allowed taking “oblique” microphotographs focused selectively on the basal or alar aspects of the neural tube, or on any other portion of interest. In a few cases we resorted to the reconstruction of separately photographed focus planes.

The lizard Lacerta gallotia galloti eggs were collected with permission in the field in Tenerife (Canary Islands, Spain) during several yearly laying seasons. The eggs were incubated further at room temperature in the laboratory, enveloped in slightly humified cotton in periodically ventilated containers. At appropriate intervals, the embryos were fixed and staged according to the developmental tables for Lacerta vivipara [Dufaure and Hubert, 1961] (abbreviated DH) and Lacerta gallotia galloti [Ramos, 1992]. This study comprises data sampled from practically all stages between stages 20 and 32. The rest of the procedure was like in chick embryos.

In the case of rat embryos, our AChE protocol consisted of overnight fixation in a cold phosphate-buffered paraformaldehyde solution, and subsequent wholemount incubation (usually overnight) at 4°C of the partly dissected and bisected embryo forebrains (cut at the isthmus) according to Karnovsky and Roots [1964]. We found that the rat embryonic AChE was more sensitive to paraformaldehyde than chicken AChE. We thus reduced the concentration of paraformaldehyde from 4% to 1%, and, after various empiric trials, selected a pH 5.7 Tris/maleate buffer solution in the incubation solution, instead of the original pH 6 acetate buffer used with chick embryos. The reaction was stopped by buffer washes after its progress was judged sufficient by visual inspection. The forebrains were then halved, and all tissues external to the neuroepithelium were peeled off manually under a magnifying microscope. The clean specimens were dehydrated in an ethanol series and cleared in methyl benzoate.

In some cases, we further dissected the specimens eliminating the telencephalic vesicle to facilitate flat mounting under a coverslip, which was elevated by pieces of coverslip under the corners. Some tears sometimes appeared at this point. Moreover, some reacted specimens were embedded in paraffin and serial 10 μm-thick sections were obtained. The weak staining observed in this material was successfully intensified by overnight exposure of the slides (after deparaffination and hydration) to osmium tetroxide vapor in glass slide racks placed in a closed chamber holding a small amount of 1% OsO4 solution at the bottom.

Results

We will first reexamine the early forebrain data published in Puelles et al. [1987a] (stages 11–18, i.e., up to 2.5 days of incubation), profiting to correct some interpretive errors that we detected in retrospect. Then we proceed with our correlative unpublished chick AChE results for the period of 3–5 days of incubation [Amat, 1986]. Basal and alar plate data will be described separately. Note that each distinct emerging positive cell or cell group is heterochronic relative to surrounding elements appearing earlier or subsequently. Such heterochrony is best noticed initially, at the start of neurogenesis, since it tends to become less distinct as soon as the available space of each structural subunit becomes uniformly covered by neurons. We will next present one after another similar observations on the forebrain from the lizard and rat embryonic AChE material (the latter extracted from Puelles et al. [2015b]). The last section illustrates heterochronic neurogenetic data on the chicken embryonic hindbrain (unpublished material from Amat [1986]), with a glance at selected rat hindbrain data.

B – Chick Forebrain PatternsB.1 – The Basal Forebrain Longitudinal Domain

The earliest forebrain AChE-positive neurons appear caudally at stage 11 in the diencephalic proneuromere (site of prospective dp1 prosomere), roughly at dorsoventral mid-levels of the lateral neural wall (Fig. 2a; we refer to Hamburger and Hamilton [1951] stages). Subsequent development by stages 12–14 increases the number of postmitotic neurons at this dp1 site identified as nflm (nucleus of the medial longitudinal fascicle, a.k.a. interstitial nucleus of Cajal) (nflm; Fig. 2b–d). This patch remains well delimited rostrally and caudally. It soon becomes evident by disproportionate growth of the overlying undifferentiated alar plate that these earliest nflm neurons actually belong to the developing basal plate, which increasingly occupies a relatively more ventral locus (it is unclear whether some amount of dorsoventral displacement of the differentiating cells occurs). It is interpreted that the prospective forebrain alar plate is tiny in the early neural tube, and the prospective basal plate relatively larger, but the alar plate grows differentially due to its exponential mode of proliferation, whereas the basal domain does not expand so much (compare with the proportions at stage 16; Fig. 2f). Remarkably, the precociously differentiated population of the caudal diencephalic basal plate does not expand from this initial dp1 locus either rostralward, or caudalward into the midbrain. Neither midbrain or hypothalamus show any neurogenesis at stages 11 and 12 (Fig. 2a, b).

Fig. 2.

Chick AChE-reacted material at stages HH11-HH16. Red lines separate anteroposterior partitions (proneuromeres or neuromeres). a Sagittal forebrain section at HH11 showing earliest postmitotic neurons of the nflm basal cell group at the caudal diencephalon (Di). M = midbrain; SP = secondary prosencephalon. b Wholemount at HH12. The diencephalon has produced already prosomere dp1, where most differentiated nflm neurons appear, while the larger rostral parencephalic region (par) is not yet divided into dp2 and dp3. c Wholemount at HH13, showing a second incipient basal cell group – abas – at the rostralmost (acroterminal) part of the secondary prosencephalon (SP), apart of increased nflm cells at basal dp1. d Wholemount at HH14. The abas cell group is now much more populated, and the remaining neuromeres have appeared (dp2, dp3 and hp1, hp2), so that it now clearly lies in hp2. Note hp1 contains dorsally the telencephalic primordium, while hp2 contains the eye vesicle (cut at the optic stalk, os). The m1 mesomere remains fully undifferentiated. There appear also some alar neurons in dp1 and dp2. e Wholemount at HH15. Apart of the previous basal cell groups nflm and abas, additional basal groups have appeared: p2tg in dp2, p3tg in dp3, and pbas in hp1. Group abas has started to extend bilaterally caudalward, under the optic stalks (forming the “abas wings”). There remains a gap between abas and pbas. At the basal m1 there appears the oculomotor nucleus (III). Incipient alar neuronal populations are observed in m1 (the Me5 dorsal cell group), dp1, and dp2. f Wholemount at HH16. All the forebrain neuromeres except the retarded m2 have developed basal plate cell groups. In m1, the rtg (rubral tegmental) group has emerged rostrally above the III. The other cell groups have established mutual contact, thus building jointly a basal plate band of neurons, which is still rather sparsely populated in some parts. A blue line was traced indicating the position of the alar-basal boundary. At HH16 the earliest neurons start to appear at the preoptic area (po) within hp2, just dorsal to the optic stalk (os). The epiphysis protrudes out at the center of the dp2 roof, and the telencephalic outpouching is visible.

At stage 13 we first see a new AChE-positive cell group – abas – which emerges tenuously in the hypothalamic proneuromere (SP) at median acroterminal basal plate level of the prospective hp2 (i.e., far apart from nflm). Subsequently, at stage 14, this population appears better developed (abas; Fig. 2c, d). This cell patch was originally identified as “arch” (retrochiasmatic area) [Puelles et al., 1987a], but Puelles et al. [2012a] renamed it anterobasal nucleus or area (abas), following the apt term introduced by Altman and Bayer [1978, 1988]. The midbrain remains devoid of any neurogenesis at stages 13–14 (m1; Fig. 2c, d).

Nevertheless, a few isolated AChE-positive neurons can be distinguished at basal dp2 and dp3 diencephalic levels already at stage 14, which we now interpret as p2 tegmentum (p2tg) and p3 tegmentum (p2tg; p3tg; Fig. 2d; these substitute the vague or inaccurate older names “atp” – area of the tuberculum posterior – and “arm” – retromamillary area) [Puelles et al., 1987a]. All these basal sites appear more abundantly populated by stage 15. At stage 16 the dp2 cell group p2tg shows a prominent extension ventralwards (compare p2tg in Fig. 2e, f), and the group abas expands bilateral wings oriented caudalwards under the optic stalks, roughly up to the limit with hp1, which appears free of differentiation at stage 15, but shows earliest neurons of the prospective posterobasal group – pbas – at stage 16 (abas; pbas; Fig. 2e, f). The pbas notion also derives from Puelles et al. [2012a]; it had been misinterpreted as “amaml” (lateral mamillary area) in Puelles et al. [1987a].

Moreover, at stage 15 the first ventral basal cell group appears in the m1 midbrain unit, which we identify as the anlage of the oculomotor nucleus (3); it appears more distinctly developed at stage 16 (3; Fig. 2e, f). We also observed at stage 16 another incipient postmitotic population at the rostral end of midbrain basal m1, not well delimited from nflm, and located just above the 3. This forms an incipient population that we now identify more precisely as the rubral tegmental area or rtg (rtg; Fig. 2f); previously we called it imprecisely “ateg,” or “tegmental area”). Note that all early neuromeric basal populations identified so far in the midbrain, diencephalon, and hypothalamus will later occupy a dorsal part of the adult basal plate, whereas the corresponding void ventral or paramedian basal domains of early embryos will become populated subsequently, often producing different basal neuronal types (e.g., mesodiencephalic dopaminergic neurons – an evidence of forebrain tagmatic mesodiencephalic similarity, or hypothalamic tuberal and mamillary neurons – hp2 –, and retrotuberal and retromamillary neurons – hp1). It thus seems that the modular basal plate domain is heterochronic and uniformly starts to develop at its dorsal border with the alar plate and thereafter expands ventralwards towards the underlying paramedian basal subdomain that shows relatively retarded neurogenesis next to the floor plate.

During the following stages 17–19 the different basal cell groups (abas, pbas, p3tg, p2tg, nflm, 3, and rtg) previously emerged in a disjoint order along hp2, hp1, dp1, dp2, dp3, and m1 cohere gradually by further intercalated addition of neurons forming an apparently continuous columnar arrangement known as the basal plate band (Fig. 3a–c). Note this band is not continuous caudally with the hindbrain, due to the largely undifferentiated m2 neuromeric unit. This is part of the evidence (apart from various shared vs. differential molecular patterns) indicating that the midbrain belongs developmentally to the forebrain tagma, contrary to the classic neuroanatomic notion that it is a separate vesicle intercalated between forebrain and hindbrain. In contrast with the much retarded caudal midbrain unit (m2), the rostral hindbrain progresses separately in its neurogenetic program (in our forebrain specimens we occasionally see attached to the caudal midbrain the AChE-positive anlage of the isthmic trochlear motor nucleus (4) (e.g., Fig. 3c). At stages 17–20 the initially unpopulated ventral or paramedian parts of the different neuromeric basal plate areas increasingly display clear-cut ventralward extension of the basal populations, particularly at the hypothalamic hp2 abas unit (Fig. 3a–c, 4a, b).

Fig. 3.

AChE-reacted chick forebrain wholemounts at HH17 (a, b) and 18 (c). Red lines separate anteroposterior partitions (neuromeres). A blue line enclosing the basal plate band separates basal plate from alar plate. a, b These specimens were oriented for preferent phocus on the basal plate. Some basal cell groups extend ventralwards towards the floor plate across the interposed undifferentiated paramedian basal domain (e.g., p2tg and p3tg). The alar pretectal cells in dp1 appear subdivided into the molecularly distinct commissural and precommissural pretectal areas (cpt; pcpt). Thalamic alar cells largely concentrate at the incipient anterobasal area (ab), which contacts both the basal plate and the emergent and increasingly AChE-positive zona limitans intrathalamica separating th/dp2 from pth/dp3 (ZLI; not seen in b). There is also a small group of prethalamic alar neurons. c At stage HH18 the basal plate band is much more compact, though the individual neuromeric modules are still partially identifiable. The alar midbrain shows a well-developed dorsal Me5 population, but the tectal domain is still essentially undeveloped. Only rostrally, behind the pretectum, a rostral m1 alar domain starts to develop, identified as the tectal grey area (tg). The pretectal cpt subarea is more populated than its pcpt counterpart (alar dp1). At the thalamus (th; alar dp2) we still see only the ab cell group. The ZLI shows strong neuroepithelial AChE activity. At the prethalamus (alar dp3) most alar neurons concentrate in a posterobasal area in front of the ZLI (pb). The alar part of hp1, found behind the optic stalk, shows the earliest paraventricular and subparaventricular cell groups (pa; spa), which later will expand rostrally into alar hp2, dorsally to abas. The preoptic area shows at this stage a distinct neuronal population (po). The telencephalon remains wholly undifferentiated.

Fig. 4.

AChE-reacted chick forebrain wholemounts at HH19 (a) and HH20 (b). The basal plate band continues to consolidate at these stages, with added expansion ventralwards into the retarded paramedian basal territory. a At stage 19 there appear dispersed pioneering neurons in the m1 tectal area (tect), caudally to the more precocious tectal grey area (tg). Caudally the prospective torus area remains unpopulated, as occurs with the preisthmic area in m2 (tor; preisth). The alar dp1 (pretectum) shows no significant change, whereas the alar dp2 (thalamus) displays at mid-dorsoventral level the first neurons of the principal thalamic area (th), caudal and above the ab cell group. A few neurons appear at the habenular thalamic area (hb). The epiphysis (ep) starts to show some neuroepithelial AChE reaction. The alar dp3 (prethalamus) has developed a distinct population ventrally, corresponding to the future “subcentral” incertal area (SCe), as well as early neurons at the future “eminential” area (E). In the alar hypothalamus the pa and spa areas start to expand into hp2. The telencephalic subpallium (spall; hp1) starts to have postmitotic neurons, whereas the preoptic area (po; hp2) increases its population. b At stage 20 the hypothalamic basal plate shows novel paramedian differentiated areas identified as periretromamillary (prm) and perimamillary (pm) in hp1 and hp2, respectively; these lie just over the still undifferentiated sites of the prospective retromamillary and mamillary areas (rm; m). The alar m1 domain expands distinct tg and tect populations and shows now also some toral area cells (tor). The Me5 population is still distinct dorsally and clearly reaches also the m2 unit. There appears some neuroepithelial background AChE activity at the rtg (m1) and the cpt (dp1). The epiphysis (ep; dp2) shows strong neuroepithelial AChE activity, and so does the dorsal end of the ZLI (compare with a). The alar hypothalamic spa and pa areas (hp1, hp2) as well as the telencephalic spall and po areas (hp1, hp2) are more developed. The pa area in hp1 typically extends also dorsally into the telencephalic stalk, caudally to the spall (future hypothalamo-amygdalar corridor).

At stages 19 and 20 the basal plate band of AChE-positive neurons is thicker and better developed, though the individual neuromeric basal plate components can still be roughly identified, particularly abas, pbas, p3tg, p2tg, and nflm, due to their advancing front at the ventral border of the basal band, where each of these populations extends ventralward separately with an irregular spike of cells (Fig. 4a, b). Remarkably, the basal plate of m2 still remains devoid of neurons, appearing as a neuron-less unstained domain intercalated between the 3 in m1 and the 4 within the isthmus (Fig. 4a, b). At stage 22 the growing rtg midbrain cell group shows a gradiental distribution of its population, decreasing caudalwards in its cellular density (rtg; Fig. 5a).

Fig. 5.

AChE-reacted chick forebrain wholemounts at HH22 (a, b). a An overview shows advanced alar development, particularly in the pretectum (cpt, pcpt; dp1), thalamus and habenula (ab, th, hb; dp2), and prethalamus (pb, SC, C, E; dp3), as well as in the alar hypothalamus and the subpallial and preoptic telencephalon (spall, po; hp1, hp2). b Higher magnification detail showing in particular the thalamic ab area and the prethalamic pb area, both of which relate intimately to the ZLI secondary organizer, whose (Shh-positive) core is marked by the red line (compare Fig. 15a).

At stages 24 and 26 the basal plate band has developed further but is now too thick to admit efficient penetration of the histochemical reactives, so that we lack an accurate AChE image (Fig. 6a, b).

Fig. 6.

AChE-reacted chick forebrain wholemounts at HH24 (a) and HH26 (b). These images show mainly the alar diencephalon, distinctly oversize now compared with the underlying basal plate (blue line). a At stage 24 (4 days of incubation) the pretectum (alar dp1) starts to show its underlying molecular division into the commissural, juxtacommissural, and precommissural domains (cpt, jxpt, pcpt). The thalamus (alar dp2) displays a massive central ovoid mass (th) above the ab area (indistinct here), and a clear-cut habenular area (hb), not yet fully expanded dorsalward. The prethalamus (alar dp3) is now uniformly covered by neurons (SCe, C), which show low AChE activity (similarly as the pcpt population), except the molecularly distinct eminential subpopulation limiting caudally the interventricular foramen, which displays higher AChE (E). Both thalamus and prethalamus show their AChE-negative chorioidal roofplate which starts in front of the epiphysis (the habenular commissure is not formed yet). The alar hypothalamic pa and spa areas (a.k.a. “supraopto-paraventricular area”) are also well developed across hp1 and hp2, and the dorsal spike of the pa area clearly enters dorsalward the floor of the interventricular foramen, forming the hypothalamo-amygdalar corridor (HyA). b At stage 26 (5 days of incubation), the alar diencephalon appears uniformly covered by neurons, except at the ZLI cell-poor central gap (representing a radial glia palisade). The cpt, jcpt, and pcpt subdomains of the pretectum are distinguishable (note the fibers of the posterior commissure – pc – running ventralwards strictly along cpt). The thalamic ab, th, and hb subdomains are also distinct (note a dorsalward expansion of hb when compared to that in a). The different prethalamic domains (pb, SCe, C, E) also have reached their full areal extent, particularly by dorsal expansion of both C and E (the stria medullaris tract courses longitudinally through E, ZLI, and hb towards the habenular commissure that forms behind the caudal end of the diencephalic chorioidal roofplate).

B.2 – The Alar Forebrain Longitudinal Domain

The earliest alar postmitotic neurons of the forebrain tagma that develop AChE reaction were seen at stage 13 in the midbrain (Fig. 2c) [see also Puelles et al., 1987a; their Fig. 5a, b]. These alar cells surprisingly lie along the midbrain roof plate and adjacent dorsalmost alar plate. They represent the singular Me5 population (mesencephalic trigeminal nucleus), which appears simultaneously along both m1 and m2. There is a sharp caudal limit of Me5 at the isthmo-preisthmic (r0/m2) boundary. These neurons later have typological, hodological, and molecular features of sensory ganglion neurons, which normally derive from neural crest or placodal sources. There is a theory that neurulation occurs at midbrain levels so rapidly that part of the neural crest material does not separate from the alar and roof plates and remains locked inside the midbrain after the closure of the roof. The derived sensory ganglion cells thus differentiate at the dorsal locus where one would expect to see any neural crest remnants, that is, close to the roof plate. In some chicken breeds, other neural crest derivatives such as melanocytes are found mixed with the Me5 population [Puelles and Gil, 1978]. In a way, therefore, considered as an hypothetic ectopic neural crest derivative Me5 does not truly belong to the midbrain alar plate.

Properly alar neurons are first seen dispersed at pretectal (dp1) level at stage 14, and they grow in number at stages 15–16, becoming very obvious at stage 17, divided into commissural and precommissural pretectal microzones (cpt; pcpt; Fig. 2d–f, 3a, b). At stage 17, the thalamic alar domain (dp2) has only very few differentiated neurons, mainly found ventrally in an area known as “anterobasal progenitor area” (ab), next to the incipient zona limitans intrathalamica, a glial palisade with secondary organizer functions (ab; ZLI; Fig. 3a, b) [Martínez-de-la-Torre et al., 2002]. Alar cells also start to emerge in the alar prethalamus (dp3), also mainly ventrally (pth; Fig. 3a, b); this precocious microzone corresponds to the prospective prethalamic zona incerta [Puelles et al., 2012a, 2021].

The rostralmost part of the alar optic lobe starts to show AChE-positive cells of the prospective tectal gray retinorecipient center [García-Calero et al., 2002; Puelles, 2019; Puelles, 2022] at stages 17–18 (tg; Fig. 3a–c). The precocious midbrain tg is next followed caudalwards by the more retarded optic tectum and auditory torus semicircularis microzones, both sparsely populated over stages 19–20 (tect; tor; Fig. 4a, b) [Puelles et al., 1994, 2019a].

At stage 19 there appears a well-delimited emergent neuronal group occupying selectively an intermediate dorsoventral sector of the dp2 thalamic alar domain (th; Fig. 4a), probably corresponding to an early born superficial thalamic nucleus, the superficial magnocellular nucleus of Rendahl [1924], studied in Puelles et al. [1991] and Martínez et al. [1991]. The primordium of the epiphysis at the dp2 roof also starts to show AChE activity at stage 19, though it is devoid of neurons (ep; Fig. 4a, b). We also found from stage 18 onwards the earliest hp1 and hp2 hypothalamic alar neurons outside the eye and telencephalon, namely at the regions later occupied by the subparaventricular and paraventricular areas, which later expand into hp2 over and under the optic stalk (spa; pa; Fig. 3c, 4a, b) [Puelles et al., 2012a]. Earliest neural retina ganglion cells differentiate from stage 15 onwards [Prada et al., 1981; Puelles, 2009; L.P., unpublished AChE data]; the whole eye evaginates early on out of an acroterminal alar hp2 subarea lying under the prospective preoptic area; this is represented in our material by the cut optic stalk (os). Other secondary prosencephalic alar neurons emerge gradually at the preoptic area (telencephalic alar subdomain of hp2) during stages 17–20 (po; Fig. 3a, c, 4a, b).

The alar populations advance significantly in cell numbers at stages 19 and 20. Remarkably, the preisthmic m2 domain continues undifferentiated, excepting its participation in the dorsal Me5 singularity (Fig. 4a, b). The tect and tor m1 alar subregions lying caudal to the precocious tg microzone now show likewise a dispersed population of neurons, slightly more numerous ventrally than dorsally; a rostrocaudal gradient is not observed (tect; tor; Fig. 4a, b). These are possibly just the earliest-born tectal neurons previously identified autoradiographically and histochemically as “solitary magnocellular neurons” of the tectal stratum griseum centrale [Martínez-de-la-Torre et al., 1987].

In the diencephalon, the alar pretectum (dp1) shows a caudal commissural pretectal domain (where the fibers of the posterior commissure course into the basal plate), which is most populated, as well as a rostral less populated precommissural pretectal domain that pushes the pretecto-thalamic boundary slightly rostralwards (Fig. 4a, b). We have shown that the pretectal alar domain is actually divided molecularly in anteroposterior direction into three progenitor areas named commissural, juxtacommissural, and precommissural areas, each producing a number of specific pretectal nuclei [Ferran et al., 2007, 2008, 2009]. It is possible that the incipient slender intermediate juxtacommissural area cannot be distinguished at these stages from the denser commissural area (cpt; pcpt; Fig. 4a, b).

The alar thalamus (dp2) shows at stage 22 a higher number of AChE-positive neurons, including some at the dorsocaudal habenular subdomain (ab; th; hb; Fig. 5a, b). The precocious boomerang-like anterobasal area – ab – next to the ZLI, extends also along the alar-basal border [see Martínez-de-la-Torre et al., 2002]. It continues to be the most populated thalamic area, followed by the dorsocaudal th area (with a gradient in dorsoventral direction). We have reported about a thalamic model in which three dorsoventral tiers (dorsal, intermediate, and ventral) are distinguished as thalamic pronuclei (primordia of various definitive thalamic nuclei), apart from the overlying, slightly caudal habenular area lying next to the thalamic neural roofplate (the latter is represented by the epiphysis – ep –, its stalk, and the chorioidal roofplate). The dorsoventral differences in the neuronal densities observed within the thalamic alar area at stages 20–22 suggest the postulated tier structure [Díaz et al., 1994; Yoon et al., 2000; Redies et al., 2000; Dávila et al., 2000; Puelles, 2001; Martínez-de-la-Torre et al., 2002; González et al., 2002].

The prethalamic domain (dp3) also shows at stage 22 an extensive alar population stretching with a subtle cell density gradient into the roofplate (pth; Fig. 5a, b). We recently examined prethalamic genoarchitectonic subdivisions [Puelles et al., 2021]. Dorsoventrally we identified three parts: prethalamic eminence (E), central prethalamus (C), and subcentral prethalamus (SC), the latter corresponding to the classic zona incerta. Apart the unitary dorsal prethalamic eminence (which reaches the insertion of the roof chorioidal tela – ch), the central and subcentral domains are divided each in three distinct anteroposterior portions. There also exists a prethalamic mirror image of the thalamic anterobasal area (but placed rostrally to the ZLI) which we accordingly identified here as posterobasal area (pb; Fig. 5a, b).

As regards the alar hypothalamus, at stages 20–22 the alar populations spa and pa of hp1 appear well delimited from the neighboring alar prethalamic area (dp3) and the less populated alar terminal hypothalamus (hp2). The latter was partly lost during the dissection of the optic vesicle, which derives from its acroterminal subregion (PHyA; THyA; Fig. 5b). The previously described precocious telencephalic preoptic area seems now accompanied by parts of the hemispheric subpallium at stages 20–22 (po; spall; Fig. 4b, 5a). We distinguish in principle preoptic, diagonal, pallidal, and striatal subdivisions of the subpallium, all of which converge dorsalward upon the subpallial septum [Puelles et al., 2000; Bardet et al., 2010; Puelles et al., 2013]. Results at stages 20–22 suggest that postmitotic neurons have appeared at least at preoptic and diagonal subregions, and perhaps are incipient at the pallidal subregion, but are still absent from the more retarded striatum.

At stages 24 and 26 (4 and 5 days of incubation) the diencephalic alar plate shows a well-developed mantle layer at pretectal (dp1), thalamic (dp2), and prethalamic (dp3) levels (Fig. 6a, b). The pretectum is neatly limited dorsally from the more dorsally prominent habenular area of the thalamus. The pretectal boundary with the underlying tiered egg-shaped main part of the thalamus (dark th area in Fig. 6a at stage 24) is unclear in the stage 26 whole mount (Fig. 6b). The prethalamus is separated from the thalamus by the cell-poor ZLI boundary, best visualized at stage 26 (ZLI; Fig. 6b). The wholemount shows at the top of the populated parts of thalamus and prethalamus a transparent membrane, which is the diencephalic roof chorioidal tela, which is inserted caudally in front of the prospective habenular commissure (apparently not yet formed at these stages) which lies in front of the epiphyseal stalk (not seen).

Finally, the stage 26 wholemount also shows distinctly the superficial anterobasal area population – ab – that borders both the ZLI and the underlying basal plate, in a boomerang shape (ab; Fig. 6a, b).

C – Lizard Forebrain Patterns

We next illustrate selected embryonic specimens from Lacerta gallotia galloti between stages DH25 and DH30 (Fig. 7), emphasizing similarities and differences with regard to chicken material. One peculiarity of the lizard embryos is that some forebrain neuroepithelial sites that showed marked AChE reaction in the chick (e.g., the zona limitans intrathalamica or ZLI, or the pretectal commissural area or cpt) did not do so in the lizard, whereas other neuroepithelial sites such as the isthmic organizer (IO; Fig. 7a, b, f) and the anterobasal hypothalamic region within THy (abas; Fig. 7a–g) showed a significant neuroepithelial reaction (in the latter case possibly coinciding with local differentiating neurons mixed with the ventricular cells, rather than in a mantle layer, at least initially).

Fig. 7.

AChE-reacted wholemount forebrain preparates of lizard embryos at different stages (a–g). The interneuromeric boundaries are traced in red, while the alar-basal boundary is represented by a blue line. Note that, in contrast to chick embryos, lizard embryos show strong neuroepithelial AChE activity at the locus of the isthmic organizer (IO; a site where FGF8 is released), but do not show neuroepithelial AChE signal at the diencephalic ZLI (a site where SHH is releases; compare Fig. 15b). a At stage HD25 most basal plate areas are partly populated (rtg, nflm, p2tg, p3tg, abas), but pbas still shows sparse neurons. There is no alar differentiation. b At stage HD25+ the forebrain basal plate modules are all visible, and alar differentiation has started at the m1 tect and Me5 areas, as well as at the cpt area (alar dp1) and the thalamic incipient ab area (dp2). See also wholemount AChE staining at pontine and prepontine parts of the hindbrain down to r4 in b. c, d Between stages HD26 and 26+ the AChE image hardly changes. Within m1, additional dispersed cells appear at the tect and Me5 areas, whereas the neighboring tg or tor areas are retarded or undistinct. Within the alar diencephalon, the prethalamic component (pth) is most advanced. The first positive cells emerge at the alar hypothalamic Pa area. The preoptic area (po) remains unpopulated. e The stage HD28 specimen shows considerable progress in its alar cell groups. The midbrain m1 and m2 units both show abundant AChE-positive cells. Earliest tg and tor cells have been added to the pre-existent tect and Me5 ones. The tor has a peculiar shape, here traced by a green dash line. The pretectum (alar dp1) shows more cpt cells, and less numerous pcpt cells. The thalamus (alar dp2) has a better developed ab cell group and some dispersed th cells dorsal to it. The prethalamus (alar dp3) is still the most developed subregion (pth), showing abundant positive cells in a ventrodorsal gradient, well delimited from the neighboring units. Its dorsalmost cells next to the telencephalic part of hp1 probably represents the anlage of the prethalamic eminence (E), whereas the rest must contain the prospective central and subcentral prethalamus subdomains (compare g). The alar hypothalamic areas Pa/SPa are retarded with respect to the alar prethalamus. Sparse preoptic neurons are present (po; this contrasts with precocious local neurogenesis in the chick), as well as sparse subpallial telencephalic cells (spall). f At stage HD29, the m2 midbrain unit shows a distinct preisthmic alar population (preisth), just behind the m1 torus (tor; traced with a green dash line). The pretectal cpt domain shows AChE-positive fibers of the posterior commissure (pc). The thalamic ab area is now much better developed and notably ascends behind the ZLI. There are otherwise still very few central thalamic cells (th), and no habenular differentiation. The prethalamus (alar dp3) can now be subdivided into its subcentral (SCe; prospective incertal), central (C), and eminential (E) portions; the latter appears well delimited from the rest and starts to bulge behind the prospective interventricular foramen. The alar hypothalamic areas SPa and Pa which first emerged in alar hp1 are distinct and have expanded rostrally into hp2. The hp1 Pa component has expanded into the floor of the interventricular foramen, forming the hypothalamo-amygdalar corridor (HyA). Preoptic area and subpallium are better populated (spall, po). g At stage HD30 the forebrain alar plate is fully covered by neurons, excepting the cell-poor gap of the ZLI and pallial parts of the telencephalon. In m1 the torus (tor; green dash limit) has elongated, theoretically by addition of more caudal elements next to the preisthmus. The three anteroposterior subdomains of the pretectum were delimited by white dash lines (cpt, jcpt, pcpt), similarly as the main dorsoventral parts of the prethalamus (SCe, C, E). In the hypothalamus, the SPa alar area extends rostrally intercalated between the abas area and the rostral part of the pa area on top of the optic stalk.

C.1 – Basal Plate

At stage DH25/25+, AChE-positive cells are already present along the basal plate modules of m1, dp1–dp3, hp1 and hp2 (rtg, nflm, p2tg, p3tg, pbas, abas), though the dp3 and hp1 components (p3tg, pbas) are still very weakly populated (Fig. 7a, b). The m2 remains unpopulated, as we saw in the chick. The oculomotor nucleus is clearly observable within basal m1 at stages DH26/26+ (3; Fig. 7c, d). The initially retarded basal parts are more advanced at stage DH25+, at which a complete but not yet compact basal plate band is visible, further amplified at stage DH26 (Fig. 7b, c). The paramedian tuberal hypothalamic area remains largely undifferentiated at these stages (tu; Fig. 7a–d). Basal plate compaction advances notably at subsequent stages (Fig. 7c–f), though there remains a noticeable gap at the level of the incipient zona limitans cell-poor transversal core (dp2/dp3 limit) as of stage DH28 (ZLI; Fig. 7e–g; note lack of neuroepithelial AChE at the ZLI); this gap was not seen in the chick. Likewise, the periretromamillary and perimamillary ventrobasal hypothalamic areas next to the retarded retromamillary and mamillary areas of hp1 and hp2 start to be populated as of stage DH28, while the tuberal area remains retarded in this aspect (prm, pm; Fig. 7e, f). Midbrain and diencephalic basal areas also expand into the retarded paramedian domain (Fig. 7e, f). At stage DH30 the thickness of the basal plate starts to hinder efficient penetration of histochemical reactives, so that less detail could be observed basally (Fig. 7g).

C.2 – Alar Plate

As regards the reptilian forebrain alar plate, the dorsal Me5 population first appeared at stage DH25+, accompanied already by sparse tectal AChE-positive cells (Me5; tect; Fig. 7b). The latter increase in number at stages DH26, 26+, and 28 (tect; Fig. 7c–e). Though they initially are distributed rather homogeneously within alar m1, starting at stage DH28 into DH30 two subareas with increased cell density are observed, thought to correspond to the incipient rostrally placed “tectal grey domain” (tg; Fig. 7e–g), and the caudoventrally placed “torus semicircularis”; we surrounded the latter by a dashed green limit (tor; Fig. 7e–g; the torus is the sauropsidian homolog of the mammalian inferior colliculus). The m2 alar plate (or preisthmus) remains largely undifferentiated up to stage DH26+ but shows an incipient population at stages DH28–30 (preisth; Fig. 7e–g). The midbrain alar plate thus seems to progress in neurogenesis relatively earlier than the diencephalic counterpart, which contrasts with the opposite pattern in chicken.

The diencephalic alar plate seems largely undifferentiated at stage DH25 (Fig. 7a), but shows a group of caudal cells at stage DH25+; these cells correspond to the commissural pretectum (alar dp1), since they coincide with the caudal locus where the fibers of the posterior commissure are going to course subsequently (cpt; Fig. 7b). This population is not more developed in our stage DH26/26+ specimens (Fig. 7c, d), but is clearly more abundant at stage DH28 (cpt; Fig. 7e), and is now accompanied by additional cells in the precommissural pretectum domain (pcpt; Fig. 7e). At stages DH29/30 the pretectum shows higher cellularity, always with predominance of cpt over pcpt, and the cpt area appears covered by AChE-positive fibers of the posterior commissure (cpt; pcpt; pc; Fig. 7f, g). We indicated by dashed white lines the approximate location of the initially less distinct intermediate juxtacommissural pretectal domain, as identified in the adult lizard (jcpt; Fig. 7g) [Martínez-de-la-Torre, 1985; Medina et al., 1992, 1993; see also Ferran et al., 2007, 2008 for chick and mouse].

The thalamic alar plate (dp2) is slow in starting neurogenesis (Fig. 7a–d). At stages DH28/29 we see a distinct differentiating rostroventral cell group, probably the primordium of the anterobasal thalamic subdomain (ab; Fig. 7e, f). This enlarges subsequently both caudalward (ventrally) and dorsalward (rostrally, behind the ZLI; ab; Fig. 7g). In Figure 7g we separated with a dashed white line this subdomain from the rest of the thalamus (the larger caudodorsal – cd – histogenetic subdomain, site of th); at this stage the AChE-negative retroflex tract (rf) courses dorsoventrally just in front of the dp1/