262 



NATURE 



[July 14, 1898 



four- or six-tube Bunsen burner, and the basin and reservoir 

 bottle are supported above On a wooden frame. The basin is 

 of zinc, and may be raised or lowered so that the mouth of the 

 geyser tube is flush with the bottom of the basin or raised 

 above it as shown. The supply tube is recurved slightly at the 

 bottom of the flask, so that the cold jets which siphon in from 

 the reservoir will not be directed against the glass wall of the 

 flask and break it. The reservoir bottle is connected by rubber 

 tubing with the supply tube, so that the bottle may be freely 

 raised or lowered to various levels indicated by the dotted 

 lines a, b and c. 



Experiment l.—''Old FaithfuP' Type. 

 When heat is applied below, the reservoir level being at a, 



fter about 14 minutes an eruption takes place, characterised 

 by violent ebullition in the flask below, ejection of the water- 

 column to a height of about 4 feet and of a mixture of steam 

 and water for a few seconds longer ; then the water-level in 

 the reservoir is seen to fall suddenly, a stream is seen to be 

 flowing into the lower flask from the curved tip of the supply- 

 4;ube, the cooling of the base of the column is accompanied by 



ondensation of steam and downward suction, the water rises 



to level a again and a period of repose fellows. It should be 

 noted that if the level of th? cooler water in the reservoir is at 

 a, the expanded warmer water in the geyser tube is somewhat 

 above a. The process described is repeated at regular inter- 

 vals of about li minutes, the duration of each eruption being 

 about 20 seconds. If the water in the reservoir be not ren ewed, 

 it gradually becomes warmer and the intervals are of s horter 

 duration. In this case, or with the reservoir level som ewhat 

 higher, as at b, and the geyser mouth raised above the basin, as 

 shown in the figure, we have in miniature the conditions of 

 "Old Faithful." li 



Experiment 2. — ^'Excelsior" Type. . 

 The conditions are altered if we raise the reservoir level to 

 the point shown in the figure, namely, just above the hei ght of 

 the geyser mouth {c). In such case there is continuous over- 

 flow of the hot water, and if the outlet tube t be left open, 

 this will continually fl^ow off; this over-flow must be constantly 

 compensated at the supply tube by cooler water from the 

 reservoir, so that the water in the. flask never reaches the boiling 

 point. If the water-level of. the reservoir be maintained con- 

 stant, this circulation will continue indefinitely, and in such 



NO. 1498, VOL, 58] 



case there will be a dome-shaped mass of hot water continually 

 boiling up and overflowing at the geyser's mouth, as in the case 

 of the Excelsior Geyser. Now at this stage, if the water-level 

 in the reservoir be allowed to sink under the drain upon it, it 

 may fall to a level six inches below r^ without interrupting the 

 continuous overflow ; in other words, it may fall back to the 

 b level, and yet the geyser will continue to act as a boiling 

 spring, without entering into an eruptive phase. The cause of 

 this is to be found in the differential expansion of the water 

 noted above, and a convectional upflow which acts as a driving- 

 power even against a reversed head, after overflow has once 

 been established.^ The overflow tube t may at this stage be led 

 into the reservoir at the b level ; this establishes a permanent 

 circulation, the only loss being by evaporation. With the 

 diminution in pressure if the level of the reservoir sinks, there is 

 a tendency towards diminished inflow of cooler water at the 

 supply-tube ; this implies rise in temperature of the water at the 

 base of the geyser-column, which tends to augment both 

 volumetric expansion and convectional velocity. Hence there 

 is here a critical point where the hydrothermal and hydrostatic 

 forces are in very delicate equilibrium ; if the reservoir is lowered 

 an inch, the overflow decreases, ebullition takes place below, 

 and an eruption of extraordinary violence takes place. The 

 same effect is at once produced by placing the glass stopper in 

 the reservoir bottle, and so checking the atmospheric pressure. 

 When the mouth of the geyser-tube is flush with the bottom of 

 the basin, an eruption may be induced by stopping the over- 

 flow tube t and permitting the water-level to rise in the basin, 

 thus augmenting the pressure on the geyser-column. Eruptions 

 once started will continue intermittently, if the hydrostatic con- 

 ditions are maintained constant ; if, however, the water-level of 

 the reservoir again rises to a point where continuous overflow is 

 possible at the geyser's mouth, the eruptions will cease and a 

 hot-spring phase will follow. 



Field Application of the Results of Experiment. 



The two simple experiments described, when compared with 

 the facts of nature, account for the most essential variations 

 observed in the phenomena of geyser eruption. Both are methods 

 of draining the reservoir — the one continuous, the other 

 spasmodic. In the same way the geyser-springs drain off the 

 superficial waters that accumulate from the abundant rainfall of 

 the Yellowstone Plateau. The "Excelsior " cauldron is stated 

 by Hague ("Geol. History of the Yellowstone National Park, 

 Transactions Am. Inst, of Min. Eng., vol. xvi., 1888) to dis- 

 charge constantly into the Firehole River 44CX5 gallons of boiling 

 water per minute, " and there is no evidence that this amount 

 has varied within the last two or three years (1887)." Weed 

 {/.c.) has estimated, on the moderate assumption that one- 

 third of the eruption-column of Old Faithful is water, that 

 3000 barrels are thrown off at each eruption. Here we have 

 examples of continuous and spasmodic drainage methods, both 

 sending their waters eventually to the Madison River, and re- 

 supplied from a local source. 



The geyser basins are topographic hollows, which supply 

 vents for the meteoric waters accumulated in fissures of the 

 decomposed rhyolite. These waters are heated by vapours 

 escaping from the only partially cooled deeper lavas, and are 

 escaping in the form of springs and geysers. In the springs the 

 overflow is occasioned by hydrostatic pressure ; in the geysers it 

 is permitted by occasional violent discharge. The transition 

 from one phase to the other may readily be induced, as shown 

 in Experiment 2, by very slight changes in the hydrostatic 

 pressure, i.e. variations in the mean level of ground- water 

 (Grundwasserspiegel), or in the local head for any specific case. 

 The head of water may be modified at either the source (supply 

 reservoir) or the orifice of exit ; head is diminished by lowering 

 the reservoir through formation of new outlets or through de- 

 creased supply, or by building up a cone around the geyser 

 tube. Conversely the head of water may be increased by 

 excessive supply (rainfall) at the reservoir, by clogging of out- 

 lets, or by the water finding a new vent at a lower level. 



Soaping Geysers. 

 It has long been known that by artificially confining the st«im 

 in small-mouthed geysers of high surface temperature, eruption 

 may be brought about prematurely. In Iceland the Strokr is 

 1 Such convection currents gain no mo;nentum without overflow, hence 

 at the a level convection played no essential part in the phen omena 

 observed. 



