Individual Dendrites Research Articles

Choline acetyltransferase immunoreactivity is localized to four types of synapses in the rat interpenduncular nucleus

The cholinergic synapses of the rat interpeduncular nucleus (IPN) were demonstrated by immunostaining that utilized a monoclonal antibody directed against choline acetyltransferase. The rostral, central, intermediate and lateral subnuclei of the IPN each contained a single type of immunoreactive terminal. Immunoreactivity was localized to synaptic vesicle membranes (especially at the contact zones), and longitudinal microtubules in preterminal portions of axons. Terminals were identified by comparison to previous studies of the synaptic organization of the IPN. In the rostral subnucleus, the immunoreactive terminals were characterized by their content of spherical vesicles, 45 nm in diameter, intermixed with moderate numbers of dense-cored vesicles, 75-100 nm in diameter. These terminals formed asymmetrical contacts. They correspond to the more numerous of the two types of axodendritic terminals described in this subnucleus, i.e. those which degenerate after lesions of the habenula. The moderate number of immunoreactive terminals in the lateral subnucleus contained pleomorphic vesicles, 30-45 nm in diameter. Up to three of these formed symmetrical contacts with individual dendrites, which ranged in diameter from 0.35 to 0.55 micron. The other types of axodendritic terminal in this subnucleus, which often contacted the same dendrites, were unstained. These latter terminals have been interpreted as being those which contain substance P. The immunoreactive terminals in the diameter, and formed markedly asymmetrical en passant contacts with small dendritic processes or spines. The immunoreactive terminals in the intermediate subnucleus had the same presynaptic and contact morphology. Many were clearly crest synapses. The remainder appeared to be such, but seen only partially within the plane of section. In the intermediate subnucleus there were up to several hundred immunostained terminals per grid square in some sections. These findings are consistent with the existence of a dense cholinergic projection to the IPN. The habenular region, are shown to crest and S synapses, both of which generate after lesions of the cholinergic innervation of other subnuclei of the IPN increases understanding of the relation of cholinergic to other transmitters localized to various portions of this nucleus.

Integrative mechanisms controlling directional sensitivity of an identified sensory interneuron

Several identified interneurons in the cricket cercal afferent system display directional sensitivity to wind stimuli: the spike frequency of these cells depends on the wind direction with respect to the animal's body. Factors determining the directional sensitivity of one of these identified interneurons (interneuron 10-3) were studied in detail. This cell has 3 dendritic branches that arborize in 3 distinct regions of the terminal abdominal ganglion. Using 2 independent methods, it was demonstrated that the dendrites have different receptive fields to wind stimuli. First, small patches of filiform hairs, whose afferents projected to individual dendrites, were isolated and selectively stimulated. In each case the response of the cell matched the receptive field of the afferents in the patch. Second, a laser beam directed through the stereo dissecting microscope was used to photoinactivate small portions of the cell in situ during intracellular recording. By isolating or ablating individual dendrites, the contributions of each of the 3 dendrites to the overall receptive field were assessed. Although the receptive field of the whole cell could be predicted by a summation of the receptive fields of all 3 dendrites, the precise directional sensitivity of the cell could not be predicted by a simple linear summation of the receptive fields of each dendrite. Two factors were found to account for this nonlinearity of summation. The first factor was polysynaptic inhibition from other interneurons within the terminal abdominal ganglion. Wind directions that activate inhibition in interneuron 10-3 were identified, and the specific classes of filiform afferents that activate the inhibitory pathway were determined. The net effect of the inhibition was to "sharpen" the directional sensitivity of 10-3 by selectively decreasing the cell's response to specific excitatory inputs. The second factor that contributed to directional sensitivity was the complex electroanatomy of the interneuron. The probable location of the spike-initiating zone (SIZ) was determined by using the laser photoinactivation technique. The relative efficacies of synaptic inputs onto the 3 different branches were then interpreted with respect to their different electrotonic distances from the SIZ. On the basis of the data obtained in this report, we present a qualitative model for the basis of directional sensitivity in this cell.

Morphology of HRP-injected spinocervical tract neurons: effect of dorsal rhizotomy

Twenty-five physiologically identified spinocervical tract (SCT) neurons in the sixth lumbar segment of the cat were filled with HRP by intracellular injection. All were reconstructed from sagittal sections using the camera lucida, and a subset (n = 18) was also reconstructed using a computer reconstruction system. Thirteen cells were in intact preparations, nine were in spared root preparations (L5, L6, S1, S2 cut; L7 spared), and three were in preparations with L5 through S2 cut. Analysis of the dendritic tree of these neurons revealed little change in gross morphology after partial deafferentation despite increased proportions sensitive to nociceptive input (Sedivec et al., 1983). The dendrites still largely respected the lamina II-III border, and relatively few dendrites were directed ventrally from the cell body, although the ratio of ventral to dorsal dendrites was greater than normal. The major change was an increase in surface area and volume caused by changes in diameter (but not length) of the dendrites. Larger-than-usual maximum branch order of individual dendritic trees of some cells was also observed after chronic deafferentation. Thus, SCT cells in deafferented segments do not undergo atrophy, but show, rather, limited signs of growth and the possibility of dendritic reorganization. We have also computed correlations between different parameters of these cells (cell body size, number and size of primary dendrites, total area and length of individual dendrites) and have found that, as in motoneurons, diameter of the primary dendrite measured 30 micron from the soma is significantly correlated with total dendritic surface area and length. SCT neurons tend to have more dendrites than spinal alpha-motoneurons, but total surface area is smaller for a given diameter of a proximal dendrite.

Choline acetyltransferase immunoreactivity is localized to four types of synapses in the rat interpeduncular nucleus.

The cholinergic synapses of the rat interpeduncular nucleus (IPN) were demonstrated by immunostaining that utilized a monoclonal antibody directed against choline acetyltransferase. The rostral, central, intermediate and lateral subnuclei of the IPN each contained a single type of immunoreactive terminal. Immunoreactivity was localized to synaptic vesicle membranes (especially at the contact zones), and longitudinal microtubules in preterminal portions of axons. Terminals were identified by comparison to previous studies of the synaptic organization of the IPN. In the rostral subnucleus, the immunoreactive terminals were characterized by their content of spherical vesicles, 45 nm in diameter, intermixed with moderate numbers of dense-cored vesicles, 75-100 nm in diameter. These terminals formed asymmetrical contacts. They correspond to the more numerous of the two types of axodendritic terminals described in this subnucleus, i.e. those which degenerate after lesions of the habenula. The moderate number of immunoreactive terminals in the lateral subnucleus contained pleomorphic vesicles, 30-45 nm in diameter. Up to three of these formed symmetrical contacts with individual dendrites, which ranged in diameter from 0.35 to 0.55 micron. The other types of axodendritic terminal in this subnucleus, which often contacted the same dendrites, were unstained. These latter terminals have been interpreted as being those which contain substance P. The immunoreactive terminals in the diameter, and formed markedly asymmetrical en passant contacts with small dendritic processes or spines. The immunoreactive terminals in the intermediate subnucleus had the same presynaptic and contact morphology. Many were clearly crest synapses. The remainder appeared to be such, but seen only partially within the plane of section. In the intermediate subnucleus there were up to several hundred immunostained terminals per grid square in some sections. These findings are consistent with the existence of a dense cholinergic projection to the IPN. The habenular region, are shown to crest and S synapses, both of which generate after lesions of the cholinergic innervation of other subnuclei of the IPN increases understanding of the relation of cholinergic to other transmitters localized to various portions of this nucleus.

An electron microscopic study of the development of axons and dendrites by hippocampal neurons in culture. II. Synaptic relationships

We have studied by electron microscopy the synaptic relationships which develop between hippocampal neurons in dissociated cell cultures. Neurons, obtained from fetal rats at 18 to 20 days of gestation, were plated at high cell density onto polylysine-treated coverslips and maintained in serum-free medium. After 3 to 4 weeks, the cells were interconnected by an extensive network of processes which made frequent synaptic contacts with one another. Certain of the processes could be readily identified as dendrites by their branching pattern and content of polyribosomes. Often individual dendrites could be followed for 100 to 200 microns from their cells of origin. In every instance observed, the dendrites were postsynaptic. The presynaptic processes were quite different in appearance; they lacked ribosomes, their microtubules were spaced more closely together, and they were thinner than even the distal dendrites except at synaptic sites where they formed varicosities. Because of their small diameter, the presynaptic processes could not be traced through the dense neuropil to their origins, but in all other respects they resembled the axons identified in younger cultures. There were differences in the features of the synapses present on different portions of the cell. The great majority of synapses on dendritic spines made asymmetric junctions, whereas those on cell bodies made symmetric junctions. Both types were observed on dendritic shafts, but asymmetric junctions were predominant. These results show that synapses do not form indiscriminately between the processes which develop in culture, but rather that axons and dendrites acquire distinct synaptic polarities, just as when they develop in situ.

Electrophoretic migration due to postsynaptic potential gradients: Theory and application to autonomic ganglion neurons and to dendritic spines

A postsynaptic mechanism for transforming electrical activity at a synapse into structural modifications near the synapse is proposed. The firing of a synapse produces a postsynaptic potential which is attenuated as it spreads electrotonically into other parts of the neuron. The potential difference so generated gives rise to an intraneuronal electric field, and permits the electrophoretic migration of charged metabolites, both along the inner surface of the membrane and in the cytoplasm. Estimates of the sizes of these fields suggest that they can play a significant role in locally organizing the distribution of the charged species, thus allowing differential biochemical activity at different synaptic locations. The mechanism is applied first to the innervation pattern found in ciliary ganglion neurons by Purves and Hume [(1981) J. Neurosci. 1, 441–452]. They showed that although ganglionic neurons are multiply innervated in early postnatal life, adult neurons are innervated such that the number of separate pre-ganglionic axons is approximately proportional to the number of major ganglionic dendrites. It is demonstrated how electrophoretic competition can result in individual dendrites establishing individual domains of innervation. The electrophoretic mechanism is next applied to dendritic spines. Theoretical calculations are presented which show that synaptic activity at a spine head can establish large electric fields along the spine, whereas synaptic activity at the base of the spine generates very small electric fields in the spine. The thinner the spine, the stronger the fields. If one assumes that these electric fields can lead to electrophoretic migration of charged metabolites necessary for either synaptic stabilization or enhancement, significant accumulation should occur within spines. Because the electric fields are small along the spine when electrical activity is generated elsewhere in the neuron, charged metabolites should not easily be drawn out of a spine. As a result, it is proposed that dendritic spines provide an anatomical means to reduce competition between neighboring afferent inputs. It is also shown how two other sets of experimental observations illustrate the electrophoretic effect. They are calcium action potentials in growth cone areas and membrane heterogeneity in myelinated axons.

Dendritic reorganization in the denervated dentate gyrus of the rat following entorhinal cortical lesions: A Golgi and electron microscopic analysis

AbstractThe organization of the dendritic tree and the morphology of individual dendrites of the dentate granule cell were analyzed qualitatively and quantitatively by means of the rapid Golgi technique 2, 4, 8, 10, 14, 30, 60, and 250 days after unilateral lesions of the entorhinal cortex (EC). Three dendritic field parameters were analyzed from camera lucida drawings of de‐nervated granule cells at each survival time: (1) dendritic field spread, (2) dendritic length, and (3) dendritic branching. Cells in the contralateral dentate gyrus served as controls. Spines were counted at each postlesion interval from material stained with a modification of the Golgi‐Kopsch method. The amount of tissue occupied by dendritic shafts at different postlesion intervals was also evaluated in samples of the ventral leaf of the dentate gyrus prepared for electron microscopy. After unilateral lesions of the EC, dendrites of the granule cells undergo modifications which appear to represent deterioration and recovery. When the dendrites reach the denervated zone, they abruptly change their orientation and tend to follow a course parallel to the granule cell layer. In contrast to normal dendrites, those in the denervated neuropil only occasionally reach the outer boundaries of the molecular layer. At the time of maximal denervation there is often a sudden reduction in dendritic diameter as the dendrite enters the denervated zone. Varicosities are also prominent. The alterations in individual dendrites are not evident 2 days after the lesion, are fully developed 10 days after deafferentation; and disappear for the most part by 30 days postlesion. The quantitative analysis of Golgi‐stained granule cells reveals that there is a 40% reduction in the total length of the granule cell dendritic tree. Electron microscopic analysis confirms the Golgi observations, indicating that the amount of neuropil occupied by dendritic shafts in the denervated zone is initially reduced and later recovers to values close to those observed in control animals. While the apparent dendritic loss is mainly restricted to the denervated zone, significant modifications occur in the inner npnde‐nervated molecular layer; there is an increase in the length of primary dendrites as evidenced by an increase in the distance to the first branch point. There is also a polarized redistribution of 1st, 2nd, 3rd, and 4th branch points toward the middle and outer molecular layer that persists even 30 days after the lesion. The results are discussed in terms of the capabilities of the granule cells of the dentate gyrus to reorganize and remodel their dendritic surface after partial deafferentation.

Estimation of the drag forces acting on particle dendrites

Aerosol particles depositing in fibrous media form chainlike agglomerates (dendrites) on the fibers, and this process leads to substantial increases in filtration efficiency and pressure drop. The object of the present work is to develop an expression for the determination of the drag forces exerted on individual dendrites. This problem is central to the determination of pressure drop increase caused by particle deposition. A method for the approximate estimation of the drag force is developed, based on the concept of the ideal dendrite, which is defined as one composed of a chain of coaxial particles. The drag on an ideal dendrite is set equal to the drag that would be exerted on a half prolate spheroid of appropriate dimensions, placed in the position of the dendrite. Both tangential and normal drag components are obtained. The tangential component is derived by considering linear shear flow, whereas the normal component is derived by considering stagnation flow. The results are expressed in terms of correction factors for the Stokesian contributions of the particles that compose a dendrite. These correction factors can be used to estimate drag forces on dendrites deposited on plane or on curved surfaces. The theory is in good agreement with existing data.

Quantitative analysis of the dendritic branching pattern of granule cells from adult rat dentate gyrus

The three-dimensional structure of Golgi-impregnated cells of the hippocampal dentate gyrus was studied using modern data-collecting techniques. Branch length and branch angle distributions were examined and found to have a wide range of observed values. The distributions imply a stochastic design for the branching pattern of dendrites. Examination of the ordered branch lengths suggests that the local dendritic branching probability may be dependent on the density or specificity of afferent projections. Analysis of the total branch lengths and number of branch nodes strongly suggests that quantitative parameters related to the total dendritic arborization may be more important in the structural design than parameters related to the individual dendrites.

Direct observations on the contacts made between Ia afferent fibres and alpha-motoneurones in the cat's lumbosacral spinal cord.

1. The enzyme horseradish peroxidase was injected into identified lumbosacral alpha-motoneurones and Group Ia afferent fibres in cats anaesthetized with chloralose and paralysed with gallamine triethiodide. Subsequent histological examination allowed the determination of (a) the extent of the motoneuronal dendritic trees, (b) the number and location of Ia synapses upon the motoneurones. 2. alpha-motoneurones had seven to eighteen primary dendrites and each produced daughter branches up to the fourth to the sixth order. At dendritic bifurcations Rall's 3/2 Power Law was obeyed. There was little or no dendritic tapering up to about 800 micrometers from the soma. Beyond this distance, however, there was considerable tapering. 3. Horseradish peroxidase injections revealed that motoneuronal dendrites are much longer than previously thought. Individual dendrites could be traced for up to 1600 micrometers from the soma and dendritic trees were usually 2-3 mm from tip to tip. Nearly all the motoneurones had dendrites that entered the white matter of the cord. Dendrites could also reach as far dorsally as laminae V and VI. 4. Ia synapses upon motoneuronal somata were examined in cords counterstained with cresyl violet or methylene green. About 10% of Ia boutons in lamina IX were on somata and each Ia collateral terminated on 3.66 motoneuronal somata or the most proximal (30 micrometer) dendrites, with on average about two contacts per motoneurone. 5. Ten Ia afferent fibre-motoneurone pairs were injected with horseradish peroxidase. The following conclusions were drawn: (i) only one collateral of any given Ia axon makes contact with a motoneurone even though other collaterals from the same axon might pass through the dendritic tree, (ii) usually all contacts made between a Ia fibre and a motoneurone are at about the same geometrical distance from the soma, even when on different dendrites, (iii) between two and five contacts are made upon the dendritic tree (average 3.4) at distances of between 20 and 820 micrometers from the soma. 6. The results are discussed in relation to previous anatomical and electrophysiological work.

Monosynaptic cortical input and local axon collaterals of identified striatonigral neurons. A light and electron microscopic study using the Golgi-peroxidase transport-degeneration procedure.

Following the injection of horseradish peroxidase into the ipsilateral substantia nigra, 36 retrogradely labelled neurons in the striatum were characterized (in three rats) by Golgi staining and gold toning: each neuron was of the medium-size, densely spinous type. Prior to the injection of horseradish peroxidase, two of the rats had had lesions placed in the ipsilateral motor cortex, the third rat had had a lesion placed in the ipsilateral frontal and prefrontal cortex. In the electron microscope, degenerating boutons of cortical neurons were found in asymmetrical synaptic contact with the spines of proximal and distal dendrites of all six of the identified striatonigral neurons that were studied. Some of the degenerating boutons were small (diameter 0.1-0.3 micron), while others were larger (1-2 microns). An individual dendrite of a striatonigral neuron was in symaptic contact with very few degenerating boutons. Local axon collaterals in the striatum could be traced from two of the identified striatonigral neurons that received degenerating cortical boutons. These were studied in the electron microscope; their boutons formed symmetrical synapses with spines or dendritic shafts of other striatal neurons. The synaptic boutons contained large, clear, round and pleomorphic vesicles. The postsynaptic targets of these boutons morphologically resemble the dendrites of medium-size spiny neurons. It is concluded that afferents from the cortex make monosynaptic contact with the dendritic spines of medium-size spiny striatonigral neurons and that such neurons have local axon collaterals in the striatum that form synapses with other spiny neurons.

The direct observation of unidirectional solidification as a function of gravity level

A saturated solution of NH 4 Cl-H 2O was solidified unidirectionally on Earth at 1 g and in a suborbital rocket at 10 -5 g. In the 1 g experiment, extensive fluid flow was apparent, creating plumes which carry numerous crystallites vertically above the interface in a manner usually attributed with causing freckling in castings. During growth the interface grew into the liquid at the same rate as the individual dendrites. When solidified in low g the interface grew at a slower rate than the individual dendrites. No plumes of liquid carrying crystals were seen, but several crystals did appear and grow ahead of the interface. Growth rates, dendrite morphologies and preferred crystallographic orientations are compared for the two experiment conditions. Results indicate that solidification in low gravity produces more uniform dendrite arm spacings, and tends to be controlled by diffusion and crystallographic orientation.

Early postnatal development of entopeduncular neurons and their neostriatal connections

The early postnatal development of dendrites and axodendritic connectivity in the entopeduncular nucleus of kittens was studied by light and electron microscopy. We focused on the fine structure of neuropil and the organization of synaptic elements at several ages during the first weeks of life. We were able to define, in newborn to 1-week-old kittens, several of the distinct types of nerve endings seen previously by us in adult animals. They were found singly or in small clusters widely spaced along surfaces of dendrites and cell bodies. Most endings were small, contained 45 to 50-nm synaptic vesicles, and formed symmetrical synaptic junctions. Fewer large endings, filled with 30- to 35-nm vesicles, formed asymmetrical synapses. By the end of the second week, most dendrites and much of the cell body surfaces were ensheathed by these endings. Astroglial elements apposed nonsynaptic surfaces and enveloped arrays of axodendritic synapses. We used Fink and Heimer staining of degenerating axons to show that much of this early connectivity originated in the neostriatum. These morphologic studies support physiologic data that striopallidal pathways are functional at birth and suggest that many of the age-related changes in response parameters reflect a quantitative increase in axodendritic connectivity. To assess this notion, we initiated a computer-assisted study of Golgi-impregnated neurons and measured dendritic growth during this period. Preliminary data suggested dendritic fields are extensive in 3-day-old kittens (mean total dendritic length of 2800 μm) and individual dendrites radiate for long distances and branch sparsely (mean total branch length of 700 μm). By 18 days of age, these parameters had increased by about 40% to 4800 and 1200 μm, respectively.

Early lesions of mouse vibrissal follicles:: their influence on dendrite orientation in the cortical barrelfield.

There is a statistically significant order in the tangential orientation of stellate cell dendrites, both spiny and smooth, in layer IV of the barrelfield of the mouse parietal cortex. Neurones situated in a barrel side have most of their dendrites oriented towards the barrel hollow; those situated in the hollow preferentially orient their dendrites parallel to the long axis of the barrel. A quantitative measure of the orientation of individual dendrites in barrelfields of 60-day old mice was obtained using a semi-automatic computer-microscope and a minicomputer. In the same manner, the dendrite orientation of layer IV stellate cells was determined in barrelfields, whose (cytoarchitectonic) pattern had been experimentally altered through lesions of the middle row of the mystacial vibrissal follicles at birth. The dendrites of these cells are oriented in harmony with the novel parcellation of the cortex. Consequently, for cells in the altered areas of the barrelfields, the dendrite orientation is different from that of cells with identical positions in a normal field (see Fig. 8). We tentatively interpret these findings as an adaptation of dendrite orientation to an altered pattern of thalamic input to layer IV that, in turn, is a consequence of the peripheral manipulation.

Normal dendritic morphology of frog spinal motoneurons: a Golgi study.

Normal dendritic morphology of frog (Rana pipiens) lumbar motoneurons was studied using Golgi silver impregnation. Branching characteristics and quantitative measurements of dendrites were obtained using computer-aided serial reconstruction of a typical lumbar motoneuron over seven adjacent 80-micrometer transverse sections. Dendrites were classified based upon site of dendrite origin from the soma and distribution of the dendritic array within the spinal cord. Eight possible sites of dendritic origin from the soma were identified. Two dendrites, D1 and D2, are planar dendrites which arise from the dorsal aspect of the soma. They are moderately complex, reaching branch order 5-6, and are oriented predominantly in the transverse plane. Input to these dendrites is primarily segmental via dorsal root projections. Three dendrites, D3, D4, and D5, arise laterally from the soma and extend through the lateral funiculus toward the subpial region. Two dendrites, D6 and D7, arise ventrally. D6 extends ventrolaterally and is a simple dendrite reaching branch order 3-4. D7 aborizes extensively in the ventral funiculus and in the central gray, reaching a branch order of 8-9. This dendrite extends rostrally and caudally over a distance of at lest 560 micrometer. Another dendrite (D8) arises from the medial aspect of the soma and projects toward the central canal. Four sites (D1, D2, D6, and D7) almost invariably give rise to dendrites. Dendrites arise at D4 in 66% of the cells examined. Dendrites are found at D3, D5, and D8 much less frequently (6-21%). Total dendritic length (12,043 micrometer) and lengths of the individual dendrites, branch length versus branch order, and number of branches at increasing radii were examined, and Sholl analysis was performed.

Experimental investigation of dendritic deposition of aerosol particles

Deposition of monodisperse aerosol particles on a single metal fiber was studied experimentally under conditions of dominant inertial impaction and interception. Formation of dendrites was observed, in accordance with previous experimental work. During each run, the deposition process was interrupted at regular intervals and the deposits on the same area on the fiber surface were examined; each time, the angular positions of the individual dendrites and their sizes and configurations were recorded and photographed. Thus, the growth of several individual dendrites was followed as a function of angular coordinate and time. These data can be used to evaluate theoretical models of aerosol particle deposition with dendritic pattern. In the present work the model developed by Payatakes and Gradón (1979) is compared with the data. The theoretical rates of dendrite growth are found in good agreement with experiment. Systematic discrepancies between the theoretically predicted dendrite number distribution and the experimentally observed one suggest that the “shadow effect” and skidding of particles (due to lubrication) prior to deposition on the fiber are significant.

Non-equilibrium solidification of hyperperitetic Al-Zr alloys

The transformation products formed during rapid casting of two alloys with composition Al-0.8 wt % Zr and Al-1.5 wt% Zr have been examined by transmission electron microscopy. The different phases to be nucleated during freezing are; (i) the equilibrium Al 3Zr phase, (ii) the metastable cubic Al 3Zr phase and (iii) supersaturated Al-Zr solid solution. The metastabie phase forms dendritic configurations up to 40 μm in diameter with the individual dendrite arm segments having a needle form about 10 nm in diameter and oriented in 〈110〉 directions. Intermittent nucleation of supersaturated α-Al was found to occur during dendritic growth of the metastable phase, resulting in symmetric dendrite-free zones inside the dendritic configuration. In the Al-1.5wt%Zr alloy, also, large solid particles of the metastable phase were formed in the melt. The observations are interpreted within the contect of the stable and metastable phase diagrams.

The direct observation of solidification as a function of gravity level

The solidification of a metal-model material, NH4Cl−H2O, was directly observed on Earth at 1 g and at 10−5 g on a suborbital rocket flight. In the 1 g experiments, nucleation started at the cold walls and then dendrites and dendritic debris were swept into the central region by fluid flow. The numerous crystals in the central zone created an equiaxed zone. Secondary dendrite arms were oriented toward the cold wall with suppressed arm growth in the direction of the flow pattern. The necking and fragmentation of secondary arms were observed. The variation in secondary and tertiary arm spacing ranged from 27 to 38 pct. Individual dendrites grew at similar rates to interface fronts. When solidified in low g, only four nuclei grew to form the complete casting. There were no free floating crystals or visible dendrite remelting. Symmetrical dendrite growth into the fluid and some necking of secondary arms occurred but no coarsening or fragmentation resulted. The growth rate of interfaces was less than that of individual dendrites. Total growth was columnar with no equiaxed zone being formed.

Superdendritic growth: I. A qualitative analysis

Superdendritic growth is identified as a general cooperative phenomenon involving groups of 30–50 dendrites that collectively present a millimeter-scale saw-tooth front. Limited experimental results are presented to show that near-zero temperature gradients in an alloy melt and gradually decreasing temperature gradients or growth velocities are prerequisites for superdendrites. The rest of the presentation is an attempt to explain qualitatively (1) the origin of superdendrites, (2) the temperature gradient variations required, and (3) the superdendrite spacing or size which seems relatively independent of the particular alloy. This reasoning is successful enough that it is extended to describe “what happens after superdendrites” since they are not regarded as steady-state growth forms. It is argued that individual large dendrites will evolve and reference is made to observations of such singular dendrites.

An on-line digital-computer system for the semiautomatic analysis of Golgi-impregnated neurons.

Rapid and accurate measurements of neuronal processes in Golgi preparations are possible with the aid of a small computer (Glaser and Van der Loos, 1965). This paper describes a system for doing this, using a small digital computer which controls stepping motors (0.5-?m steps) attached to the stage (x, y axes) and fine focus (z axis) of a microscope. The observer tracks the processes, topological information, such as the location of the soma, dendritic origins, branch points, and ends of processes, is signaled to the computer by special controls, and the x, y, and z coordinates of each are stored in digital format. The computer printout yields 1) individual x, y, and z coordinates with associated topological identifiers; 2) quantitative data for each dendritic segment in order (i.e., primary and secondary branches, etc.); and 3) computes the actual linear dimensions of each segment in microns. An associated oscilloscope display can 1) display the whole neuron, or individual processes, by connecting recorded points by vectors; 2) identify each dendrite; 3) rotate (continuously or by a specified angle) the whole cell, or individual dendrites, around any selected point or axis; and 4) indicate spatial relationships by dynamic rotation, intensity modulation, or stereo pairs. The performance of this system has been evaluated for accuracy and the repeatability of measurements.