290 PROFESSOR PIAZZI SMYTH ON 
Step. 3. Lamp-flame temperatures; these are taken mainly from M. 
Lrecog DE BoIsBAUDRAN’S admirable observations, discussed and arranged for 
this purpose by myself, and offer 370 data. 
Step. 4. Electric sparks 1 inch long; for this at present I have only 
one series of measures ; but that is of a rather collective order; being my own _ 
observations of ai in an end-on gas-vacuum tube ; its data being 221 in number, 
as published by the Royal Scottish Society of Arts in their Transactions for 
1878-79. 
Step 5. Electric sparks 2 inches long ; these are taken also from M. DE 
BoIsBAUDRAN’S admirable series of apparently all the simple substances known 
in chemistry as being easily amenable to the heating effect of the above spark. 
The data are 1048 in number. 
Step 6. Electric sparks of 6 inches and more long ; and very much further 
raised in temperature by the use of condenser and Leyden-jar apparatus. This 
is an extensive series of all the known chemical elements, as experimented on 
by THaLen, ANGsTROM, Bunsen, KirncHorr, PLucKER, HucGerns, and others, and 
probably represents the highest temperature yet attained artificially. The 
data here are 2685 in number. 
These several collections are exhibited in their numerical arrangement 
through the spectrum in Part IV., and graphically in the plate appended, with 
the effect of showing that for every increase of temperature, the maximum of 
spectroscopic phenomena is found farther and farther removed from the red, 
and towards the violet ; so that— 
Step 1 has its maximum in about W.N. 39,000 
Step 2 -f f 41,000 
Step 3 i es 47,000 
Step 4 ss ue 49,000 
Step 5 Ps 5 49,000, and 
Step 6 » 51,000. 
Now therefore comes the question, as to what is the similar spectrum place 
of the maximum of Solar Activity as exhibited by the number, size, and position 
of its lines 4 
If we take ANncsrrom’s Normal Solar Spectrum, then according to the 
numerical columns in Part IV., the maximum of its 1410 lines (treated as the 
other steps are, so as to give double weight to thick and intense lines over thin 
and faint ones), occurs at 55,000 of the Solar Spectrum scale. But as I have 
already indicated that the violet end of the Natural Solar Spectrum is cribbed 
and confined most lamentably in that map, I take my own Lisbon solar spectrum 
with 2016 lines, and can state according to the numbers also in Part IV., and 
the last figure in the plate, that its maximum is rather in 61,000 of W. N. 
spectrum place ; and that it ought even to have, though unseen by man through 
