as the interval of time most easy to catch and reproduce Odder still, both Estel and Mehner found that multiples of this time were more accurately reproduced than the timeintervals of intermediary length ;* and Glass found a certain periodicity, with the constant increment of 1.25 sec., in his observations. There would seem thus to exist something like a periodic or rhythmic sharpening of our time-sense, of which the period differs somewhat from one observer to the next.

Our sense of time, like other senses, seems subject to the law of contrast. It appeared pretty plainly in Estel's observations that an interval sounded shorter if a long one had immediately preceded it, and longer when the opposite was the case.

Like other senses, too, our sense of time is sharpened by practice. Mehner ascribes almost all the discrepancies between other observers and himself to this cause alone.t

Tracts of time filled (with clicks of sound) seem longer than vacant ones of the same duration, when the latter does not exceed a second or two. This, which reminds one of what happens with spaces seen by the eye, becomes reversed when longer times are taken. It is, perhaps, in accordance with this law that a loud sound, limiting a short interval of time, makes it appear longer, a slight sound shorter. In comparing intervals marked out by sounds, we must take care to keep the sounds uniform.§

There is a certain emotional feeling accompanying the intervals of time, as is well known in music. The sense of haste goes with one measure of rapidity, that of delay with another; and these two feelings harmonize with different mental moods. Vierordt listened to series of strokes performed by a metronome at rates varying from 40 to 200 a

* Estel's figures led him to think that all the multiples enjoyed this priv flege; with Mehner, on the other hand, only the odd multiples showed diminution of the average error; thus, 0.71, 2.15, 3.55, 5, 6.4, 7.8, 9.8, ang 10.65 second were respectively registered with the least error. Cf. Phil Studien, II. pp. 57, 562-565.

+ Cf. especially pp. 558-561.

Wundt: Physiol. Psych., II. 287. Hall and Jastrow: Mind, xi. 62 § Mehner: loc. cit. p. 558.

minute, and found that they very naturally fell into seven categories, from 'very slow' to 'very fast.'* Each category of feeling included the intervals following each other within a certain range of speed, and no others. This is a qualitative, not a quantitative judgment-an æsthetic judgment, in fact. The middle category, of speed that was neutral, or, as he calls it, adequate,' contained intervals that were grouped about 0.62 second, and Vierordt says that this made what one might almost call an agreeable time.†

The feeling of time and accent in music, of rhythm, is quite independent of that of melody. Tunes with marked rhythm can be readily recognized when simply drummed on the table with the finger-tips.

WE HAVE NO SENSE FOR EMPTY TIME.

Although subdividing the time by beats of sensation aids our accurate knowledge of the amount of it that elapses, such subdivision does not seem at the first glance essential to our perception of its flow. Let one sit with closed eyes and, abstracting entirely from the outer world, attend exclusively to the passage of time, like one who wakes, as the poet says, " to hear time flowing in the middle of the night, and all things moving to a day of doom." There seems under such circumstances as these no variety in the material content of our thought, and what we notice appears, if anything, to be the pure series of durations budding, as it were, and growing beneath our indrawn gaze. Is this really so or not? The question is important, for, if the experience be what it roughly seems, we have a sort of special sense for pure time-a sense to which empty. duration is an adequate stimulus; while if it be an illusion, it must be that our perception of time's flight, in the experiences quoted, is due to the filling of the time, and to our memory of a content which it had a moment previous, and which we feel to agree or disagree with its content now.

It takes but a small exertion of introspection to show

*The number of distinguishable differences of speed between these limits is as, he takes care to remark, very much larger than 7 (Der Zeitsinn, p. 187). †P. 19, § 18, p. 112.

that the latter alternative is the true one, and that we can no more intuit a duration than we can intuit an extension, devoid of all sensible content. Just as with closed eyes we perceive a dark visual field in which a curdling play of obscurest luminosity is always going on; so, be we never so abstracted from distinct outward impressions, we are always inwardly immersed in what Wundt has somewhere called the twilight of our general consciousness. Our heart-beats, our breathing, the pulses of our attention, fragments of words or sentences that pass through our imagination, are what people this dim habitat. Now, all these processes are rhythmical, and are apprehended by us, as they occur, in their totality; the breathing and pulses of attention, as coherent successions, each with its rise and fall; the heartbeats similarly, only relatively far more brief; the words not separately, but in connected groups. In short, empty our minds as we may, some form of changing process remains for us to feel, and cannot be expelled. And along with the sense of the process and its rhythm goes the sense of the length of time it lasts. Awareness of change is thus the condition on which our perception of time's flow depends; but there exists no reason to suppose that empty time's own changes are sufficient for the awareness of change to be aroused. The change must be of some concrete sort-an outward or inward sensible series, or a process of attention or volition.*

*I leave the text just as it was printed in the Journal of Speculative Philosophy (for Oct. 1886') in 1887. Since then Münsterberg in his masterly Beiträge zur experimentellen Psychologie (Heft 2, 1889) seems to have made it clear what the sensible changes are by which we measure the lapse of time. When the time which separates two sensible impressions is less than one third of a second, he thinks it is almost entirely the amount to which the memory-image of the first impression has faded when the second one overtakes it, which makes us feel how wide they are apart (p. 29). When the time is longer than this, we rely, he thinks, exclusively upon the feelings of muscular tension and relaxation, which we are constantly receiving although we give to them so little of our direct attention. These feelings are primarily in the muscles by which we adapt our sense-organs in attending to the signals used, some of the muscles being in the eye and ear themselves, some of them in the head, neck, etc. We here judge two timeintervals to be equal when between the beginning and end of each we feel exactly similar relaxations and subsequent expectant tensions of these

And here again we have an analogy with space. The earliest form of distinct space-perception is undoubtedly that of a movement over some one of our sensitive surfaces, and this movement is originally given as a simple whole of feeling, and is only decomposed into its elements-successive positions successively occupied by the moving body— when our education in discrimination is much advanced.

muscles to have occurred. In reproducing intervals ourselves we try to make our feelings of this sort just what they were when we passively heard the interval. These feelings by themselves, however, can only be used when the intervals are very short, for the tension anticipatory of the terminal stimulus naturally reaches its maximum very soon. With longer intervals we take the feeling of our inspirations and expirations into account. With our expirations all the other muscular tensions in our body undergo a rhythmi cal decrease; with our inspirations the reverse takes place. When, therefore, we note a time-interval of several seconds with intent to reproduce it, what we seek is to make the earlier and later interval agree in the number and amount of these respiratory changes combined with sense-organ adjustments with which they are filled. Münsterberg has studied carefully in his own case the variations of the respiratory factor. They are many; but he sums up his experience by saying that whether he measured by inspirations that were divided by momentary pauses into six parts, or by inspirations that were continuous; whether with sensory tension during inspiration and relaxation during expiration, or by tension during both inspiration and expiration, separated by a sudden interpolated relaxation; whether with special notice taken of the cephalic tensions, or of those in the trunk and shoulders, in all cases alike and without exception he in. voluntarily endeavored, whenever he compared two times or tried to make one the same as the other, to get exactly the same respiratory conditions and conditions of tension, all the subjective conditions, in short, exactly the same during the second interval as they were during the first. Münsterberg corroborated his subjective observations by experiments. The observer of the time had to reproduce as exactly as possible an interval between two sharp sounds given him by an assistant. The only condition imposed upon him was that he should not modify his breathing for the purposes of measurement. It was then found that when the assistant broke in at random with his signals, the judgment of the observer was vastly less accurate than when the assistant carefully watched the observer's breathing and made both the beginning of the time given him and that of the time which he was to give coincide with identical phases thereof.-Finally, Münsterberg with great plausibility tries to explain the discrepancies between the results of Vierordt, Estel, Mehner, Glass, etc., as due to the fact that they did not all use the same measure. Some breathe a little faster, some a little slower. Some break their inspirations into two parts, some do not, etc. The coincidence of the objective times measured with definite natural phases of breathing would very easily give periodical maxima of facility in measuring accurately.

But a movement is a change, a process; so we see that in the time-world and the space-world alike the first known things are not elements, but combinations, not separate units, but wholes already formed. The condition of being of the wholes may be the elements; but the condition of our knowing the elements is our having already felt the wholes as wholes.

In the experience of watching empty time flow-'empty * to be taken hereafter in the relative senso just set forthwe tell it off in pulses. We say 'now! now! now!' or we count more! more! more!' as we feel it bud. This composition out of units of duration is called the law of time's discrete flow. The discreteness is, however, merely due to the fact that our successive acts of recognition or apperception of what it is are discrete. The sensation is as continuous as any sensation can be. All continuous sensations are named in beats. We notice that a certain finite 'more' of them is passing or already past. To adopt Hodgson's image, the sensation is the measuring-tape, the perception the dividing-engine which stamps its length. As we listen to a steady sound, we take it in in discrete pulses of recognition, calling it successively the same! the same! the same!' The case stands no otherwise with time.

After a small number of beats our impression of the amount we have told off becomes quite vague. Our only way of knowing it accurately is by counting, or noticing the clock, or through some other symbolic conception.* When the times exceed hours or days, the conception is absolutely symbolic. We think of the amount we mean either solely as a name, or by running over a few salient dates therein, with no pretence of imagining the full durations that lie between them. No one has anything like a perception of the greater length of the time between now and the first century than of that between now and the tenth. To an historian,

* "Any one wishing yet further examples of this mental substitution will find one on observing how habitually he thinks of the spaces on the clock-face instead of the periods they stand for; how, on discovering it to be half an hour later than he supposed, ne does not represent the half hour in its duration, but scarcely passes beyond the sign of it marked by the finger." (II. Spencer: Psychology, § 336.)