Search

To identify the forces directing organismal evolution, a general analytical system is developed to synthesize structuralist and Darwinian traditions in an explicitly historical framework. Morphological features of lungless salamanders are examined to identify hierarchical systems of developmental and functional constraint on evolution and their interactions with processes at the genic and populational levels. Characteristic patterns of change occur repeatedly in the evolutionary history of this group.

We present an HRP study of the sensory tracts and motor nuclei associated with feeding (especially use of the tongue) in plethodontid salamanders (mainly Batrachoseps attenuatus, Bolitoglossa subpalmata, Desmognathus ochrophaeus, Eurycea bislineata, and Plethodon jordani). The nerves studied are VII (ramus hyomandibularis only), IX, X, XI, the first spinal nerve (hypoglossus), and the second spinal nerve. Two types of sensory projections are universally found in the brainstem: superficial somatosensory projections of VII, IX, and X, and deeper visceral sensory projections of IX and X to the fasciculus soltarius. The first spinal nerve and the spinal accessory nerve (XI) have no sensory projections, but the second spinal nerve has typical projections along the dorsal funiculus of the spinal cord. The motor nuclei of VII ramus hyomandibularis, IX, and X form a combined nucleus situated at the level of the IX/X root complex. The nucleus of the first spinal nerve is well separated from the combined nucleus and is situated rostral and caudal to the obex. The rostral part of the motor nucleus of the second spinal modestly overlaps that of the first. The motor nucleus of the spinal accessory nerve is more or less restricted to the region of the second spinal nerve. Its fibers leave the brain through the last root of the IX/X complex and the related ganglion. Bolitoglossine and nonbolitoglossine differ in the architecture of the spinal nuclei. Two distinct types of motor neurons occur in spinal nuclei of nonbolitoglossine species--some of those with tongue projection--but only one type is found among the tongue-projecting bolitoglossine group.

We present an analysis of the genetic structures of 22 species of salamanders, with regard to levels of gene flow among populations. We estimate the gene flow parameter, Nm (the product of the effective population number and rate of migration among populations) using two alternative methods described by Wright and Slatkin. For most species, these two methods give approximately congruent estimates of Nm; when estimates differ, the method of Wright produces values slightly larger than those derived by the method of Slatkin. We analyze these results in light of independently derived historical inferences of the fragmentation of populations. This analysis suggests that the Nm values calculated from protein polymorphisms may contain information more relevant to historical patterns of gene exchange than to the current population dynamics; moderately large values of Nm may be calculated for species containing populations known to be no longer exchanging genes. Application of a method for estimating the maximum possible rate of gene exchange among populations indicates that, for most species studied here, gene flow among populations probably is no greater than the mutation rate. We suggest that most plethodontid species cannot be viewed as units whose cohesion is maintained by continuing gene exchange. Furthermore, we suggest that phenotypic uniformity among populations is not easily explained by hypotheses of continual stabilizing selection and propose that future work concentrate upon clarification of the genetic and epigenetic factors conferring self-maintenance or autopoietic properties on living systems.

Innervation of the tongue and associated musculature in plethodontid salamanders was studied using Palmgren stained sectioned materials, fresh dissection, and whole mounts of experimental specimens treated with horseradish peroxidase (HRP). Species studied were chosen to represent modes of tongue projection recognized by Lombard and Wake ('77). Special attention was given to species of the genera Plethodon, Batrachoseps, Pseudoeurycea, and Hydromantes, but representatives of other genera were investigated. As expected we found that cranial nerves IX and X and spinal nerve 1 supplied the muscles involved in tongue movement. The peripheral courses of the nerves were traced, and both functionally related and phylogenetically determined routes were found. As relative projection length increases, the nerves supplying the tongue tip also increase in length. When the tongue is at rest the long nerves are stored in coils. The coil of ramus lingualis lies between the ceratobranchials, but that of ramus hypoglossus is more variable, although constant within a species. Ramus hypoglossus bifurcates into separate branches to tongue and anterior musculature of the floor of the mouth. In generalized, presumably primitive, modes the bifurcation and coiling are far anterior. In most of the tongue projection modes bifurcation is relatively posterior, but in one, bifurcation is anterior, but coiling is relatively posterior in position. The most unusual condition is in Hydromantes, in which bifurcation is relatively posterior and a coiled ramus hypoglossus joins a coiled ramus lingualis to form a unique, coiled common ramus to the tongue tip. Hydromantes has the greatest projection distance of any salamander.

The organization of the motor nuclei of cranial nerves V (including mesencephalic nucleus), VI, VII, IX, and X is described from HRP-stained material (whole mounts and sections) for 25 species representing five families of salamanders, and the general topology of the brainstem is considered. Location and organization of the motor nuclei, cytoarchitecture of each nucleus, and target organs for nuclei and subnuclei are described. The trigeminal nucleus is separated distinctly from the facial and abducens nuclei and consists of two subnuclei. The abducens nucleus consists of two distinct subnuclei, one medial in location, the abducens proper, and the other lateral, the abducens accessorius. The facial nucleus has two subnuclei, and in all but one species it is posterior to the genu facialis. The facial nucleus completely overlaps the glossopharyngeal nucleus and partially overlaps that of the vagus. In bolitoglossine plethodontid salamanders, all of which have highly specialized projectile tongues, the glossopharyngeal and vagus nuclei have moved rostrally to overlap extensively and intermingle with the anterior and posterior subnuclei of the facial nerve. In the bolitoglossines there is less organization of the cells of the brainstem nuclei: dendritic trunks are less parallel and projection fields are wider than in other salamanders. Some aspects of function and development are discussed; comparisons are made to conditions in anurans; and phylogenetic implications are considered.

The distribution and cytoarchitecture of motor nuclei of the cervical spinal cord were studied by using HRP techniques (whole mounts and sections) in 22 species of salamanders (families Hynobiidae, Dicamptodontidae, Ambystomatidae, Salamandridae, and Plethodontidae) representing a wide variety of life histories and functional modes of feeding. The nucleus of the first spinal nerve extends from the level of, or slightly caudad to, the root of the tenth cranial nerve, almost to the ventral root of the second spinal nerve. Approximately one-half of this nucleus is situated in the brainstem. This anterior extension is longest in bolitoglossine plethodontids. The nucleus of the second spinal nerve extends from the root of the first spinal nerve to the dorsal root of the second spinal nerve. The nuclei of the first and second spinal nerves in all species except bolitoglossines have motor neurons arranged in two columns: a lateral one containing large spindle-shaped cells and a medial one containing pear-shaped or polygonal smaller cells. The primary dendrites of these lateral and medial cells are parallel and their arborization is relatively narrow. In contrast, bolitoglossines lack the lateral motor column. The nucleus of the first spinal nerve consists only of a medial band of pear-shaped and sometimes polygonal cells, and the nucleus of the second spinal nerve is a wider band of pear-shaped and polygonal cells which are always situated inside the periventricular gray matter. The arrangement of the somata in bolitoglossines is less organized and the primary dendrites are less parallel and have a broader arborization than in other salamanders. In all species, cells in the second spinal nucleus are arranged in a less orderly manner than those in the first. All salamanders studied possess a spinal accessory nerve whose motor neurons are located in the cervical spinal cord; the axons leave the brainstem with fibers of the vagus nerve. The rostrocaudal extent of this nucleus differs markedly among species. In bolitoglossines the nucleus is more or less restricted to the region of the nucleus of the second spinal nerve. In all other species studied, the accessory nucleus extends from the obex to the caudal end of the nucleus of the third spinal nerve. In the tribe Plethodontini the cytoarchitecture of the accessory nucleus is similar to that of the second spinal. In desmognathine and hemidactyliine plethodontids as well as in all nonplethodontid species studied the nucleus consists of pear-shaped and cone-shaped cells.(ABSTRACT TRUNCATED AT 400 WORDS)

Organisms are self-producing and self-maintaining, or "autopoietic" systems. Therefore, the course of evolution and adaptation of an organism is strongly determined by its own internal properties, whatever role "external" selection may play. The internal properties may either act as constraints that preclude certain changes or they open new pathways: the organism canalizes its own evolution. As an example the evolution of feeding mechanisms in salamanders, especially in the lungless salamanders of the family Plethodontidae, is discussed. In this family a large variety of different feeding mechanisms is found. The authors reconstruct this evolutionary process as a series of "bifurcation points" of either constraints or opportunities forming a sequence of preconditions for the formation of a high-speed projectile tongue characteristic of tropical salamanders. Furthermore, it is shown how parallel evolution of seemingly unrelated domains within an organism such as respiratory physiology, life history biology and pattern of ontogeny has rather direct relevance to the feeding biology, thus demonstrating that organisms always evolve as wholes.

Plethodontid salamanders capture prey by projecting the tongue from the mouth. An analysis of theoretical mechanics of the hyobranchial skeleton is used to formulate a working hypothesis of tongue movements. Predictions that the skeletal elements of the tongue are included in the projectile and that the hyobranchial skeleton is folded during projection are central to the analysis. When decapitated in a particular way, salamanders project the tongue, and it is not retracted. When these heads are fixed and sectioned, examination confirms the predications. In turn, these observations are used to refine the working hypothesis and to generate a general model of tongue dynamics for plethodontids. Muscles performing the major roles of projection (subarcualis rectus I) and retraction (rectus cervicis profundus) are identified. The skeleton is folded passively along a morphological track having the form of a tractrix. Predictions concerning the shape of the track and the exact configuration of the folded skeleton are confirmed by study of sectioned material. The skeleton unfolds along the track during retraction and is spread into the resting state. The model developed herein will be used as a basis for predictions concerning selection patterns in the family and for analytical purposes in comparative and evolutionary studies.

Study of the innervation of the musculature related to feeding behavior in plethodontid salamanders by means of the horseradish peroxidase (HRP) technique has demonstrated the existence of a true spinal accessory nerve which innervates neck musculature, enters the brain via the ganglion of IX/X cranial nerves and has its motor neurons within the nucleus of the second spinal nerve. Further, it has been shown that the first spinal nerve, being strictly motor, alone constitutes the ramus hypoglossus and is, therefore, homologous to the hypoglossus of amniotes.