216 GENERAL GEOLOGY 



observed that the explosions did not assume a perfectly vertical direction, but 

 deviated a few degrees from the vertical towards the south, as shown on the accom- 

 panying photograph, Plate LXVII. Fig. L 



We uiferred from this that the funnel of the active crater of Erebus dips 

 slightly to the north, that is, towards the centre of the second crater. It is obvious 

 that in quite recent time the centre of activity of Erebus has travelled from north 

 to south. As the eruption of June 14, 1908, progressed, the steam cloud rapidly 

 became asymmetrical, while the main steam column was bent over towards the E.N.E. 

 by the wind blowing from off the high plateau of Victoria Land from a W.S.W. 

 direction. We observed that the higher and stUl ascendmg portion when at about 

 2500 to 3000 feet above the top of the crater was swayed over by the cyclonic high- 

 level air-current uito a S.E. or S.S.E. direction. At about 2.30 p.m. there was 

 a specially grand outburst of steam ; it rushed upwards, still slightly bent away 

 towards the south, and dashed its head with great violence through the mushroom- 

 shaped cloud canopy, which it swiftly penetrated, emerging on the upper side, and then 

 rushing upwards once more to a height of at least 5000 feet above the summit of 

 the mountaui. At 3.15 p.m. the steam cloud was lit up with a bright glow, forming 

 a magnificent spectacle in the blackness of the Antarctic night during midwinter. 

 At 3.25 P.M. another remarkable bright glow lit up the whole of the lower part of the 

 steam column above the crater. By 3.45 p.m. the steam cloud had become much more 

 extensive, and had also risen higher, the earlier formed portions of the cloud being 

 now trailed away in long streamers towards the south-east, while lower branches bent 

 towards the E.N.E. and were dragged out in that direction by the plateau wind. 

 At 3.50 p.m. there was another bright glow. No doubt each of these bright glows 

 indicated a separate eruption, and was caused by the breaking open of the cooled 

 scum of the lava within the crater by fresh masses of steam and mcandescent 

 lava arising from beneath. We estimated that the uprushes of steam took place 

 at intervals of between four and five minutes, though this was not absolutely 

 constant. Towards 6 p.m. the steam column gradually swayed around towards 

 the N.N.W., and finally to the N.W. By this time the earliest-formed part of 

 the steam cloud had formed a surface of thin cirrus cloud about 20,000 to 30,000 feet 

 above sea-level. At 6.40 p.m. there was another very bright glow. At 8 p.m. 

 the eruption appeared to be subsiding, and the steam cloud meanwhile stretched 

 across the sky in a direction from E.S.E. to W.N.W. (the surface wind at the 

 time was about N.N.E.). It now resembled the sinuous thin folds of some 

 delicate drapery. Evidently the wind du-ection had by this time changed above 

 the summit of Erebus, so as to blow from E.S.E. to W.N.W. 



During the whole of this eruption we were unable to observe, although we con- 

 stantly watched with good night-glasses, any trace of lava actually flowing down the 

 sides of the cone. Apparently on this occasion the crater Avas only partly filled with 

 molten lava. • 



