"By this it seems that the most advantageous circumstances for the accused are, possessing a superior education, which implies also, in some measure, the means of making a defence. The most advantageous state of all is, being upwards of thirty years of age, being a female, having received a superior education, appearing under an accusation of crime against persons, and being willing to take one's trial:—while the most disadvantageous is, being under the age of thirty, being a man, ignorant of reading and writing, accused of a crime against property, and being contumacious, unable to produce his means of defence. The causes which modify the probability of condemnation, according to the state of the accused, seem to be sufficiently evident, without their being further insisted upon. But this may not be the case with regard to the degree of influence of those causes; this estimation is fraught with some difficulties, and, on reflection, it has appeared to me that for all the numerical elements subject to variation, we may very well estimate the importance of the difference between the extremes and the mean, or the importance of the causes which produce it, by comparing the extremes to the amount of the mean. It is nearly thus that geometers, who have applied themselves to the theory of probabilities dependent on facts relative to man, and especially Buffon, estimate the importance of a sum to an individual, by comparing it to the amount of his property.

"For this estimation we must consider the extreme differences above and below 614, which is the mean amount of the repression of crime in France, when we do not take into calculation the effect of any modifying cause; then the respective differences would give the measure of their importance, and consequently that of the causes which produce them, looking on the effects as proportional to the causes. For example, let us suppose that the desire is to estimate the respective influence exercised by the advantage of having a superior education, or of being a female, on the repression of crime in France; the repression for persons having a superior education is 0'400, and for females 0.576, the first of which is 214 below the mean, and the second '038: the value of these differences then is =348, and 9062. Superior education therefore appears to be five times as efficacious as being a female, in diminishing the repression of crime in the tribunals.

"The following table shows the comparative influence of the various modifying causes in the repression of crime, computed on the same basis, so that there does not exist, as I have already shown, any more influential cause in varying the repression of crime, than the contumaciousness of the accused. It has not only the advantage of proving this result, but it also shows the degree of influence possessed by this cause.

"There now remains a different species of inquiry for us, that is, to examine how far the enumerated causes can be looked upon as constant, since, to be truly so, the annual result should continue invariable; whereas,

0.067

0.563

on the contrary, the results show a small annual modification, which we have attributed to variable causes: these modifications are generally trifling if we confine our observations to a few years, but still they must be attended to. For instance, the repression of crime, in general, has not been 614 throughout the six years which have afforded the elements of our calculations; small annual differences have been remarked, and repression in its greatest differences from the mean has increased to 635, and decreased to 593; which gives 021 both ways, and, consequently, the mean degree of their importance is, or 034; thus the variable causes which affect repression have had, in their maximum and minimum of energy, a degree of influence which has equalled, and even surpassed, that of some of the causes which we have regarded as constant. To form a more just idea of the variable causes, we must examine annual effects on each of the elements which we have before considered. The following Tables will represent this :

:

"These tables teach us that the greatest variations in the constant causes which modify the repression have never exceeded the intensity of the causes themselves; though, in other words, it will be observed that, even in the most irregular years, the effect of the constant causes has never been effaced by

that of the variable or accidental. This will be better seen in the following Table, which shows the value of the variations in all the cases before enumerated, occasioned by the causes which modify repression:

"I have reasoned on the hypothesis that the results were founded on so many observations, as to leave no contingency in the value of the means; but this is not the case here. Some results are deduced from observations which are not yet numerous, and we know that, cæteris paribus, the precision of results increases according to the square root of the observations. This applies particularly to repression in the instance of the accused having received a superior education. The values deduced are obtained from a small number of observations, and the extremes have therefore differed so much the more from the means; but recurring to the system of the least squares, I have found that the precision of the numbers 0-400, and 0.6137, before obtained as the general repression, and as the particular one exerted against the well educated, are in the proportion of 0·0870 to 0.0075, or as 11 to 1.

"Separating, according to the preceding observations, whatever might be contingent in the difference of the extremes from the means, in order to consider only those causes which have had a more or less regular influence on repression, I think their influence may be represented as 0:034. These differences are such as evidently to show a gradual diminution in the repression of crime. Now this progressive diminution must have had its causes; and one of these, no doubt the most influential, is indicated in the general account of the administration of criminal justice in France during the year 1830. Six years,' it is there said, have elapsed since the general accounts of the administration of criminal justice have been published. During the first half of that period (1825, 1826, and 1827), the lists of the juries were formed according to the rules laid down in the code of criminal instruction; during the second half (1828, 1829, and 1830), these lists have been formed according to the law of the 2nd May, 1827, which altered the basis of the jury, and has called upon a greater number of citizens to perform this duty. Taking the sum of the results of the accusations during the entire period of six years, as well as during each of the two halves, and comparing the respective results, we find that the only difference between the jury formed on the code of criminal instruction, and that prescribed by the subsequent legislation, consists in the latter having a slight tendency to regard accusations less severely. The proof of this result will be seen in the following table.

"Six years hence, we may compare these results with those which will be produced by the declarations of the actual jury, the constituent elements of which have been still further increased by lowering the elective requisites, and which does not condemn if the majority be under seven.

66

Thus, the preceding table shows us not only that the number of acquittals has not* diminished, but that even the punishments pronounced have been less rigorous; there having been fewer degrading punishments, and more correctional ones. And this observation on the tendency to regard accusations less severely, appears still more probable when we examine the crimes in detail; it is there especially that we may see whether the infliction of punishment has been abstained from on account of its severity. In effect, we find that the condemnations to death have decreased in a very sensible degree. The same observations may also be made, if we distinguish crimes against persons from those against property; a proof of which may be seen in the following table.

"On both sides we find fewer condemnations, and the condemnations themselves are less severe. It appears, then, that there are causes, whatever be their nature, which have had an influence in France, in slightly diminishing the repression of crime; the sequel will better show us whether one of these causes is to be sought for in the introduction of the law which has changed the basis of the jury, and whether this be the only cause. However it be, it is evident that the causes which have annually modified the repression in general have had less influence than the constant causes which modify the repression according to the nature of the crimes; for, in adhering to the two established periods, the first have only, on an average, occasioned two or three more acquittals on one hundred general accusations, whilst the second have almost invariably occasioned eighteen more acquittals in accusations of crime against persons than in those of crime against property. But it is only what may have been already seen by comparing the two preceding tables.'

Translator's Note. The omission of this negative in the original must be a typographical error, since not only the sense requires it, but the table alluded to increases the acquittals, and the subsequent one does it in a still greater ratio,

THE HEAVY CAVALRY AT SALAMANCA.

To those whose youth has been spent in eventful scenes, there is a natural pleasure in recalling such passages of their lives as have been most interesting, and upon which, from present retirement, they look back with quiet contemplation; but no one, perhaps, is likely to find more satisfaction in this sort of retrospect than the British officer who shared in the triumphs of the Peninsular war under the Duke of Wellington; and it is not, therefore, matter of surprise that scarcely a number of the United Service Journal appears, without animated discussions upon the events of the Peninsular war; generally with the view of elucidating its details, but occasionally with the object of drawing inferences and deductions, such as may serve for future principles and instructions in the art of war.

The discussions between the officer signing himself J. M., and those who have written in controversion of the argument he is desirous of founding upon various remarkable facts and occurrences which he has with some ability collected, in reference to the actual collision of cavalry and infantry, have attracted much attention and remark from most military readers. It is maintained sturdily by J. M., that the latter are at all times liable to be overthrown by the attack of the former, provided it be executed with determined vigour and resolution. Now it is certain that there is no officer of our cavalry, nor indeed of any other branch of our service, who, during the late war, had opportunities of seeing the charges of British cavalry, but will bear candid and satisfactory witness to the impetuous courage with which their attacks were almost invariably made. That their ardour not unfrequently carried them too far, and led to unfortunate disasters, will not be denied; but it is rather too hard that our cavalry should not be allowed even that merit, from the excess of which their failures, on the very rare occasions when they did fail, may be said to have solely arisen.

Among the many instances of their brilliant attacks, enough stress certainly has never been laid on the conduct of the brigade of heavy cavalry at the battle of Salamanca, on which memorable day they with equal steadiness and impetuosity burst upon the infantry of the French left, and literally swept them from the ground, overthrowing column after column, and sending to the rear above three times their own number as prisoners. The battle of Salamanca formed the closing scene of such a series of able and well-combined manoeuvres, and the turning of the French left by the rapid and admirably concerted march and attack of the Third Division was so prominent a feature, that it is only upon these considerations we can reasonably account for the slight mention made in the Duke of Wellington's despatch of the consecutive attacks of the heavy brigade of cavalry, which are merely noticed as "a successful charge under the orders of Sir S. Cotton (Lord Combermere)." It may be well to give a short and explicit statement of what actually took place as regards the attacks in question, leaving the impartial reader to judge for himself how far the heavy brigade may lay claim to having contributed materially to the successful execution of the bold and splendid manœuvre by which the left of Marshal Marmont's position was turned, the whole plan of his operations completely checked and reversed, and his army subjected to one of the most signal discom