This is done by pressing a stud, and sending a current at each transit; so that we shall have a dot in every other space between the clock dots, supposing the wires to be two seconds in time apart; supposing them to be three seconds apart, our dots will be in every third space; supposing them to be four seconds apart, our dots will be in every fourth space, and so on; and tenths and hundredths of seconds are estimated, by the position of each transit dot between those which record the seconds. In this way one sees that we have on the barrel an absolute record, by one of the pointers, of the seconds recorded by the clock, and, by the other, of the exact times at which a star has been seen at each wire of the transit instrument. Now of course what is essential in this method is that there shall be a power of determining not only the precise second or tenth of a second of time, but also the minute at which contact takes place, otherwise there would be a number of seconds dots without knowing to what minute they corresponded; it would be like having a clock with only a second-hand and no minute-hand. The brass vertical sliding piece shown at the lower left-hand side in Fig. 96, carries at its upper end two brass bars, each of which has, at its right-hand extremity, between the jaws, a slender steel spring for galvanic contact; the lower spring carries a semicircular piece projecting downwards, which a pin on the crutch rod lifts in passing, bringing the springs in contact at each vibration: the contact takes place when the pendulum is vertical, and the acting surfaces of the springs are, one platinum, the other gold; an arrangement that has been supposed to be preferable to making both surfaces of platinum. By means of the screws _n_ and _o_, which both act on sliders, the contact springs can be adjusted in the vertical and horizontal directions respectively. Other contact springs in connection with the brass bars _p_ and _q_, on the other side of the back plate, are ordinarily in contact, but the contact is broken at one second in each minute by an arm on the escape-wheel spindle. The combination of these contacts permits the clock to complete a galvanic circuit at fifty-nine of the seconds in each minute, and omit the sixtieth.[13] In this way we may suppress the sixtieth second, thus leaving a blank that marks the minute; and all that the observer has to do after he has made a record of the transit, is to go quietly to the barrel, and mark the hour and minute in the vacant space. A barrel of this size will contain the observations which would be made in some hours; so that at the end of that time it may be taken off, and it will give, with the least possible chance of error, a permanent record of the work of the astronomer.