A study has been made of the modifications of the shape of a nerve action potential dependent upon the placement of the two electrodes, always necessary for a lead. In a classic diphasic lead separation of the electrodes brings out, in addition to a separation of the phases, the appearance of a positive deflection traceable to the passage of an impulse between the electrodes. This phenomenon, called the lead separation effect (1.s.e.), must be considered as an expression of a feature of normal nerve fiber biophysics. It regularly appears and it can be analyzed with respect to the position of the sink maximum. Also it cannot be eliminated by a block at the second electrode. The advantage of approximating the leads was shown by the absence of a 1.s.e. following spikes recorded by electronic integration of tangents, which with validity can be derived from threshold fibers. Since tangent leads are not adaptable to recording a spectrum, a block at the second electrode is required. The making of such blocks and the configuration of records obtained with them are described. Conditions for an optimal lead, but not an ideal lead, were delimited. In an optimal lead only two major elevations appear in the spectrum of a skin nerve: those known as alpha and delta. A reference to maps of fiber size analyses shows that the fibers in the delta elevation have velocities of conduction slower than they would have if following in linear sequence the fiber diameters belonging to the alpha elevation.

The positivity following the spike in the action potential of unmedullated nerve fibers of dorsal root origin (d.r.C) has been shown to be homologous with the first positive potential (P 1 ) of other varieties of nerve fibers. Thus it is only through the large size of the positivity that this group of nerve fibers is set apart from other groups. New findings accentuate and make more explicit the difference of d.r.C fiber behavior from that of the sympathetic unmedullated fibers. Support of the conclusion is derived from re-examination of the A fibers as well as from observations on the d.r.C fibers. Increase in size of the P 1 's in a tetanus of the d.r.C fibers can occur if the frequency is high enough; and it does not occur in an A fiber tetanus if the frequency is low enough. Frequency is also critical in the obtainment of increasing P 1 's in a tetanus of sympathetic C fibers. Decrease in the size of the P 1 's in the course of a tetanus is attributable to development of the negative after-potential (N a-p). In rested d.r.C fibers the N a-p is latent. But it appears during a tetanus, develops in size, and after the tetanus leads to a long lasting and clearly defined second positive potential. Absence of a supernormal period during the N a-p of the d.r.C fibers is accounted for. An analysis is made of the apparent increase in size of the spike elevations during a tetanus, for the two subgroups of the C fibers. The difference between the after-potentials of A fibers and of sympathetic C fibers is correlated with the shapes of the curves of cathodal electrotonus of these fibers.

Cross sections of olfactory nerves present a unique appearance. They indicate the presence of large numbers of very small nerve fibers, with a modal diameter of about 0.2 µ and a narrow range for their size variation. From one side of the nasal septum of a pig the yield of fibers was estimated at 6,000,000; the number arising from the turbinates would be considerably larger. The fibers are attached to the membranes of the Schwann sheaths in large bundles through mesaxons longer and more branched than those that have been seen in other nerves. Continuity of the axons between the nerves and the bipolar cells was traced in an examination of the olfactory mucous membrane; and the indication of a one-to-one relationship between cells and axons was reinforced by a comparative count. After the axons leave the bipolar cells they become incased in the central projections of the sustentacular cells. Where the latter come into contact with the basal cells the axons emerge to push back the plasma membranes of the basal cells in the first step in acquiring their nerve sheaths. Later steps are described. When the axons are delivered by the basal cells to the collecting Schwann tubes, they are already aggregated into small bundles with sheaths fundamentally the same as those they will possess until they are delivered to the glia in the olfactory bulb. Some of the aspects of the cytology of the bipolar cells and adjoining sustentacular cells are described. A survey of the physiological properties of olfactory nerve fibers was made in some experiments on the olfactory nerve of the pike. Almost all of the action potential is encompassed within a single elevation, manifesting at its front a conduction velocity of 0.2 m./sec. For a comparison, the last elevation in the C action potential in the sciatic nerve of the frog is cited as an example of conduction at the same velocity. Though expressed through long time constants, the properties of the pike olfactory fibers conform to the generalized schema for properties of vertebrate nerve fibers. This conformity signalizes that they differ from the exceptional properties of the unmedullated fibers of dorsal root origin. An afferent function for unmedullated nerve fibers does not imply that the fibers concerned are alike in their physiological properties.

As an aid in the interpretation of the physiological properties of unmedullated nerve fibers, particularly those having their cells of origin in the dorsal root ganglia, more precise information about their morphology has been acquired through employment of the electron microscope. The appearance of the fibers in the skin nerves is described, with special reference to the structure of their sheaths; and a notation is made about the bearing of the axon-sheath relationship on the biophysical mechanism of conduction (p. 714). There is no basic difference between the sheath systems of the d.r.C and the s.C fibers. Attention is called to a point of similarity between the sheaths of unmyelinated and myelinated axons (p. 715). An assessment was made of the likelihood of interaction between the fibers. In action potentials showing temporal dispersion at several distances, the elevations appeared in their calculated positions. A model of a group of Schwann sheaths was constructed from successive electron microscope sections, showing that the lengths of the sheath branches are short in comparison with the wave lengths of the action potentials. Supported by these and other considerations, the argument is strongly in favor of the conclusion that among d.r.C fibers, as in other fibers, there is no cross-excitation between the axons. A new analysis of the size distribution of the fibers in a sural nerve was made from electron microscope pictures; and from the measurements the action potential was constructed. The result confirmed the view, previously expressed, that the velocities of conduction in the fibers can be precisely accounted for by multiplying the diameters by a constant. In the dorsal roots, the striking change that takes place in the appearance of the fibers and their disposition in the Schwann sheaths can be seen in Fig. 11. The axons partake of the special properties of the peripheral branches, which necessitated the creation of the subdivision of d.r.C fibers. But, their diameters are much smaller. At a set of reduced conduction velocities the configuration of the compound action potential in the nerves is repeated in the roots, with the root velocities still conforming to the size-velocity rule derived from nerve axons.

The compound action potential of the unmedullated fibers arising from dorsal root ganglia, as recorded in cat skin nerves after conduction of simultaneously initiated impulses, shows among its components a temporal dispersion corresponding to velocities between 2.3 and 0.7 M.P.S. The maximum representation of the component velocities is at about 1.2 M.P.S. On both sides of the maximum the representation falls off irregularly, in such a way that groupings in the distribution produce in the action potential a configuration in which successive features appear always in the same positions at a given conduction distance. Through this demonstration of a characteristic configuration the system of the unmedullated fibers is brought into analogy with that of the medullated fibers. The unmedullated fibers originating in the dorsal root ganglia have distinctive physiological properties, among which is a large positive potential which reaches its maximum immediately after the spike and decrements to half relaxation in about 50 msec., at 37°C. The positive phases of the unit potentials in the compound action potential, owing to their duration, sum to a much greater extent than the temporally dispersed spikes; and, since they have sizes such that one equivalent to 25 per cent of the spike height would not be at the limit, in the summation process the major portion of the compound action potential is caused to be written at a potential level positive to the starting base line. The position of the spikes in the sequence can be seen in the analyses in Section III. The course of the activity in unit fibers is subject to variation in ways affecting the positive potential. Preliminary descriptions, based on orienting experiments, of how these variations are conditioned are given in Section I. Two of the findings are particularly noteworthy. One is the high sensitivity of the dimensions of the postspike positivity to temperature in the range of temperatures at which skin nerves may be expected to function, even when the environmental temperatures of an animal are moderate. The other is the high sensitivity to conditioning by previous activity. The positivity is first decreased, then replaced by a negative potential of similar duration. Reasons have been given why it is inadvisable at the present time to call the postspike potential an after-potential. A comparison has been made of the properties of the unmedullated fibers arising from dorsal root ganglia with those of fibers arising from sympathetic ganglia. The differences are so great that, in the interest of precision in designation, a division of the C group of fibers into two subgroups is indicated. It is suggested that the two subgroups be named respectively d.r.C and s.C. Measurements have been made of the diameters of the d.r.C fibers in a saphenous nerve stained with silver. Graphs showing the number of fibers at each diameter are presented in Section II. In Section III there are shown constructions, from histological data, of the action potential as it would appear, after 3 cm. of conduction, with the correlation between diameter and velocity in strict linearity. The degree of fit between the constructed and recorded potentials can be seen in Fig. 18.