[Illustration: FIG. 17.--THINKING OF A BUILDING. Right hand holds pencil, and left hand holds record plate. Subject facing ↑. In tracing I., direction of the attention ↑; in tracing II., direction of the attention ↓. Time of each record, 60 seconds. II. shows respiration records.] [Illustration: FIG. 18.--COUNTING THE STROKES OF A METRONOME. Right hand holds pencil, left hand holds record plate. Direction of the attention from A to B ↑, from B to C →, from C to D ↓, from D to E ←. Time of each portion, 45 seconds.] To eliminate the record of the swaying of the body, we may experiment with the subject seated; we obtain a distinctive record in which certain phases of the fluctuations have almost disappeared, and in which the record approximates to a straight line (tracing II. of Fig. 14). One may also eliminate the record of the swaying by dispensing with the automatograph, and simply holding the recording plate in one hand and the recording device or pencil in the other; for then the plate and pencil sway together, and naturally no record of it is made. The relatively fine movements thus obtained are shown in Fig. 16; the contrast between this record and such records as Figs. 4, 5, 6, is mainly the contrast between a record in which the general swaying of the body is registered, and one from which it has been eliminated. It is interesting to note that in records thus taken, there is but a slight difference in the result when the subject is standing and when he is sitting; which is a further proof that the swaying of the body has been eliminated. (Compare these with Fig. 14.) Traces of periodic oscillations are noticeable in Fig. 16; these are due to movements of respiration, and in tracing II. of Fig. 17, they are unusually distinct and regular, about twenty to the minute. In this case the forearm of the hand holding the record plate was braced against the body, while the recording hand was held free from it; and thus the abdominal movements were registered. The movements toward the object of attention appear throughout. Fig. 17 shows a movement towards the rear of the subject, as well as towards the front; which again shows that under suitable conditions, involuntary movements may be recorded in one direction as readily as in another. Fig. 18 presents a most beautifully regular movement in all four directions. As the metronome, the strokes of which the subject was counting, was carried from one corner of the room to another and so on around the room, the hand involuntarily followed it and recorded an almost perfect square. So striking and regular and so varied an involuntary movement, in conformity with changes in the direction of attention, one can expect to secure but seldom, and then only with a good subject.