In a series of articles, Martin has described a method for using induced current (faradic stimulation) to determine thresholds for muscle and sensory processes. He and his co-workers have used this in several researches.2 The same method has been used at the Nutrition Laboratory with apparatus which is practically a duplicate of Professor Martin's. Two researches have been published, one by Dodge and Benedict, the other by Miles, in which the sensory threshold for faradic stimulation was a factor under investigation. According to the theory and formula given by Martin, threshold determinations taken with different resistances in the secondary circuit with the fingers of the subject should fall on a straight line when plotted. In the results which have been obtained at this Laboratory it has seldom been found that the thresholds of a subject fall on this theoretical straight line. Our difficulty may have been subjective, due to lack of careful cooperation and attention of the subject measured. On the other hand, it is not impossible that some of the trouble was instrumental or had to do with the technique. One point in the method seems particularly unsatisfactory, i. c, the determination of the tissue resistance. This was done by means of balancing on a simple Wheatstone bridge. The tissue, usually the finger-tips, together with a known resistance, was placed in one arm of the bridge, and against this a variable resistance was balanced, the final adjustment being made on a slide-wire. A telephone was used as an indicator of the point of balance on this wire. As a matter of fact, a good balance-point which gives anything near silence in the telephone can never be found; the operator has to resort to balancing quality against quality in the two ends of the slide-wire. This is a very difficult proposition with the telephone as an indicator, since the telephone membrane has characteristics which make it respond to certain vibration frequencies with greater degrees of intensity.' When the string galvanometer was used as an indicator in the bridge in place of the telephone, the reading was far from agreeing with that obtained by the telephone. The balance in the bridge was somewhat improved by using an adjustable capacity in the variable arm, but even with this improvement it was quite unsatisfactory as a measurement. According to the best determination, the resistance of the tissue of the finger-tips when immersed to a depth of 2 cm. in salt solution is in the neighborhood of 4,000 to 5,000 ohms.

1 In practice the constant 1.673 for the instrument and the unit width 0.008242 mm. for an arc of 1 second are, by dividing the letter into the former, combined into one factor of 191, which, when divided by the micrometer scale reading, gives directly the angular separation of the bright bands in degree-seconds.

2Martin. Am. Jottm. Physiol.. 1908, 22, p. 110. and 1010.27. p. 226; also. The Measurement of Induction Shocks, New York, 1912. See also. Martin, Porter and Nice, Paychol. Rev . 1912. 29. p. 194.

3 See Meyer and Whitehead. Proc. Am. Institute of Bee. Engineers, 1912, 31, p. 1023; Kennelly and Affel. Proc. Am. Acad. Arte and Set, 1915, 51, p. 419.

An effort was made to include a large resistance in series with the secondary circuit so that the tissue resistance of the fingers would be but a small fractional part and could be assumed without actual measurement. With this sort of a change it is necessary to use a stronger current in the primary coil of the induction apparatus. The sparking at the breaking of the primary circuit becomes very objectionable under this condition and a source of considerable error. Even with the larger coils it is not possible to include 100,000 ohms and still have sufficient strength of shock in the secondary circuit for purposes of stimulation. Larger induction coils were experimented with, such as those used in commercial transformers. One of these proved somewhat more satisfactory in that a large resistance, such as 200,000 ohms, could be included in a secondary circuit with the fingers and still enough strength of shock be provided for purposes of stimulation, without the necessity of increasing the current in the primary circuit to an objectionable amperage. Since the primary and secondary windings had a fixed relation to each other in this coil, the changing of the strength of the induced current would necessitate a change in the current at the primary. Theoretically, it seemed that this would also change the wave-form of the induced current.

An investigation was carried out to determine if the wave-form changed materially within the range of change in strength of primary current, which was desirable for purposes of sensory-threshold determination. In the course of this investigation it was observed that the change in wave-form of the induced current was considerable when the primary current was changed through that range which was necessary for experimental purposes. Another factor also revealed itself, i. e., that the wave-form of the induced current was varied by the amount of resistance in series with the secondary. The larger the resistance placed in series, the steeper was the rise of the current curve and the less the time to the maximum and steeper the fall. In the light of these facts, many of which are well known, it seemed highly desirable to discontinue the use of induced current and to use direct-current stimulation if possible. While most threshold measurements are undoubtedly subject to a certain percentage of error, this does not warrant the use of apparatus with variable factors for which corrections can not be made.

Many arrangements for using direct-current stimulation were tried; finally a simple one was adopted. This is shown in schematic diagram in figure 43. The apparatus consists of six main items: a drop wire D; a voltmeter V; a non-inductive resistance R; a pendulum, indicated by the arrow A, for breaking the circuit at points denoted by S and S'; a control switch C, which is also a pole-changing switch; and non-polarizable electrodes E. The switches S and S' are arranged to be struck open by a pendulum moving from left to right. When struck open they remain open. It is therefore evident from the diagram that with the switches closed, as shown in the figure, the current from the drop-wire has parallel paths P and P'. Since the fingers F of the subject are in series with P', and the tissue of the fingers has a resistance of about 5,000 ohms, the current thus goes through P, where the resistance is negligible as long as S is closed. When S is struck open the current is established in P'. This amounts to the "make" of the circuit. The duration of the shock is regulated by the distance between S and S'. When Sf is opened the shock is ended. This provides a shock which theoretically should have as nearly as is possible a curve showing an instantaneous rise, a square top, and an instantaneous fall. The height of the shock is a function of the voltage. The amount of electrical energy actually delivered at the fingers need not concern us here, as our results are comparative. The threshold values are expressed in terms of voltage, read at the voltmeter V, which had an internal resistance of 33,894 ohms at21°C.