Ъег, is 24 days. M. Denning's own observations, reduced from 

 2170 meteors in 1876 — 7, dive an average of 22 days. In 

 Dr. Schmidt's Catalogue tliere are 45 meteor showers with dura- 

 tion of about 30 days or more. In my own Catalogue, reduced 

 from 2000 meteors seen in England 1849 — 1867, the average 

 duration for 40 showers for the whole year is 33 days, — omit- 

 ting 12 showers over 54 days, some of which are doubtless not 

 realy one shower (ici de nouveau se pre'sente la question, — où 

 donc sont les limites de la dure'e admissible pour un seul courant?). 



I think we may therefore take it for granted, until at least 

 proved to be incorrect, that the average duration of a meteor 

 shower having a pretty constantly fixed radiant area, say of from 

 3" — 8" in diameter (et quand le courant se de'place, alors cette 

 aire embrasse pour toute la durée quelques dizaines de degrés), 

 is not less than tliree weeks. 



ÎÎ0W, as some of these meteor showers only endure for one 

 or two days, it is not unreasonable to assume for a maximum 

 a duration of even six weeks; that is to say, meteor showers, 

 so far as ascertained, endure from one to at last 40 days, gi- 

 ving in most cases a pretty fixed radiant in the heavens. There 

 are not an inconsiderable number of cases in which the duration 

 would appear to be even as great as 50 or 60 days, but this 

 would appear to be so surprising that further proof and long 

 continued nightly watching and recording is requisite to confirm 

 it. The Perseids belong to a shower having a strong maximum, 

 as well known, about August 10, but it is perhaps not equally 

 well known that this shower feebly commences about July 24 

 and continues until about August 17, ending much more suddenly 

 than it commences. 



The Leonids hast for a few days with a strong maximum of 

 only a few hours, the Adromedids for not more than half-a-day. 



When, as we have shown, it may appear highly probable that 

 the average duration of a meteor shower (of which at last 200 

 are now known whose orbits are intersected by the Earth) is 

 about tliree weeks, it must stand to reason that a considerable 

 number may be expected to show a duration of a least five or 

 six weks, if not verv occasionally more. 



Captain Tupman has clearly pointed out the special conditions 

 required to produce a nearly fixed meteor radiant for several 

 weeks, viz. „The meteor orbit must nearly coincide with the plane 

 of the ecliptic, the perihelion distance of the central position be 



