22 
On Residual Phenomena. 
[ J anuary 
another phosphate of soda agreeing with the arseniate both 
in composition and in crystalline form. Such explanations 
of apparent exceptions may, perhaps, to some minds seem to 
“ prove” or at least confirm the “ rule,” on the same prin- 
ciple that lawyers consider the title to an estate strengthened 
if it has been unsuccessfully called in question. Yet what 
truly confirms the rule is not the exception, real or apparent, 
but the detection of the fadt that it is no exception at all. 
Before anyone can venture to pronounce the imagined “law” 
that all hybrids are necessarily barren to be “ proved ” by 
instances to the contrary, he should have prepared himself 
to show that such instances are contrary in seeming only, 
and not in reality. 
We are thus brought to the threshold of the third possible 
case : — The theory or the law may agree with or account for 
the fadts to a certain extent, leaving, however, a margin un- 
explained. Such margins are the “ residual phenomena ” 
which we are about to consider : their value as a clue to fur- 
ther discovery will best appear from example. We may 
begin with one of the simplest cases, which, nevertheless, 
led to the discovery of the alkaline metal lithium. One of 
the foremost consequences of the chemical law of combination 
in definite proportions is that all samples of any compound, 
supposing them pure, must contain the same ingredients in 
the same relative quantities. A specimen of sulphate of 
magnesia, whether natural or artificial, whether prepared 
recently or a century ago, contains exadtly the same propor- 
tions of magnesia and of sulphuric acid. On a certain 
occasion the Swedish chemist Arfwedson, having been en- 
gaged with the analysis of a mineral petalite, obtained what 
he at first took to be sulphate of magnesia ; but on closer 
examination he was led to doubt this conclusion, not by any 
striking discrepancy in colour, solubility, or taste, but by an 
excess of weight. His suspicions being thus aroused, he 
soon found that he had in his hands an element as yet 
unknown. 
Turning from chemistry to physics, we find other examples 
equally striking and simple. Thus the speed with which 
sound travels through the air had been deduced with great 
precision, from its known cause and mode of propagation. 
But when the conclusions thus reached were brought to the 
test of adtual experiment, the agreement, though approxi- 
mate, was not complete. The result was sufficiently close 
to show that the calculation had been, in the main, based 
upon corredt principles, but there was still a margin, a 
“ residuum ” of velocity unaccounted for. Whilst examining. 
