Practical limitation of the recent decline (this is retained in 1920) in infant mortality to the second half of the calendar years appears, therefore, to be a definite fact even when the cause of diarrhoea is fully excluded. . . . It must be noted, however, that the large decline in infant mortality during the first ten years of the century was by no means limited to the first six months of the calendar year, but appears to have been fairly well distributed over the whole of it. In this case the crude rates may be taken as a sufficiently reliable guide to the course of events."

It is convenient here to quote his remarks on the effect of temperature on infant mortality:-"The winter quarter now yields the highest mortality and the summer the lowest without a single exception during the 10 years (1911-20), the autumn (fourth quarter) rate being higher than that of the spring (second quarter) in 7 out of the 10 years. The winter maximum and summer minimum apply moreover at all periods of the first year of life after the first day with very few exceptions, most of which may be explained by the epidemics of diarrhoea and of influenza. But as between the second and fourth quarters the order of mortality varies with age, the second furnishing the higher rate of the two at ages 6-9 and 9-12 months in all years except 1912 and 1918, and the fourth with still fewer exceptions at earlier ages (0-4 weeks, 4 weeks to 3 months and 3-6 months).

"The fact that a well-marked correspondence of mortality with temperature exists so early as the first week of life is of interest as showing how early post-natal causes come into operation. It would seem that, while it may be possible to lessen the mortality of the first day only by measures before and during birth, that of the remainder of the first week as well as of later ages, being much influenced by environment, should be susceptible to improvement by increased post-natal care. Whatever tends to protect the new-born as well as the older infant against the added risks of winter will evidently tend to reduce mortality." (83rd Annual Report of English Registrar-General, 1920, page xxxvi.) It is very important to remember this statement when the death-rates in the colder northern latitudes have to be compared with the warmer southern latitudes. It is only reasonable to assume that, if the seasonal variations in temperature so obviously and seriously affect the infant mortality curve, the country of severer weather and greater variation is likely to suffer most. This is too well known to need emphasising. It is right, however, to say that the climate of every country generates its own methods of special protection, and, although the variation of the infantile death-rate closely follows the variations of the seasons, rising and falling inversely as the temperature (within limits), special precautions in the severer climates may, to a certain extent, counteract the deadly effects. This subject will emerge later.

The relation of temperature to infant mortality is not a simple relation. The English Registrar-General finds that, among children of the earlier months, the second quarter of the year (a warmer quarter) is more deadly than the fourth quarter (one of the colder quarters). "For the first four months of life the mortality of the fourth quarter is slightly above the yearly average and that of the second well below it; but after this the curve for the fourth quarter

falls and that for the second rises, so that after six months their relation to each other is reversed, and for the last five months of the first year of life the second quarter deaths are above the yearly average and those of the fourth well below it. After six months of age, therefore, the colder fourth quarter would seem to be more favourable to life than the warmer second quarter; but it seems possible that the excess mortality at this age may be due to illness contracted in the first quarter, and that similarly some of the fatal effects of illness contracted in the fourth quarter may be deferred till the first of the succeeding year. If this be so, the figures furnish no real exception to the inverse rate of mortality to temperature, it being only necessary to suppose that the older infant fights a longer battle than the younger before succumbing to the effects of cold. And again it is not necessary to assume that temperature alone makes all the difference. Other climatic conditions accompanying low temperature may contribute to the season variation of mortality; and the effect of season may even be to a greater or less extent indirect if the care which infants receive is on the whole less adapted to winter than to summer conditions."—(Ibid., page xxxviii.) This is further qualified by the statement made later that in cold countries, like Norway and Sweden, the infant mortality is low. Cold is undoubtedly a dangerous factor in early infancy, but the capacity to resist cold can be safely cultivated. Further, there is overwhelming evidence to show that colder weather means denser overcrowding of the houses with all the unhealthy results flowing from that factor. Warm weather is relatively the period of "open-air life" both for adults and for infants. These results of cold climates can all be counteracted by better education in the management of infancy.

In Scotland the decline in infant mortality, although marked in recent years, has not proceeded rapidly. In 1855 it stood at 125; in 1861, 110; in 1871, 130; in 1881, 112; in 1891, 127; and in 1900 at 128. From 1855-1900 the variations ran from a minimum of 108 to a maximum of 138. But since 1901 the Scottish rate has never again exceeded 129, and it has fallen as low as 90-3 in 1921, the lowest figure on record up to that date since 1855. The decline from the beginning of the century has not shown the range of variation of the previous 45 years, but the range has been considerable. Since 1901 the highest rate recorded has been 123 in 1904. For the period from 1908 onwards the rate has, with variations, declined. In the year 1915 the rate was 126.5; but for the reasons already given this may be regarded as partly an inflated rate, but the actual number of deaths in 1915 was 14,441 (against 13,710 in 1914), i.e. the highest absolute number since 1909. There was, therefore, a real rise of death-rate, but the figure 126.5 is probably too high.

Obviously, therefore, the general conditions in England cannot be compared with the general conditions in Scotland. The factors operating in England differ from the factors operating in Scotland. Without a very extended analysis it would be difficult to set forth the main factors in each country; but, as the figures given below will show, when a comparison is made of approximately comparable areas, the difference between Scotland and England is found to be not greater than the differences between the North of England and the South of England. The larger country has the

advantage of the dilution resulting from the lower rates in the warmer latitudes of the South. The fact, therefore, that the general infant mortality of England should, taken as a whole, be distinctly lower than the general infant mortality of Scotland is precisely what we should expect, even if in both countries there is a progressive decrease. Let us, therefore, for the moment consider only the years from 1911 onwards. For these years it is possible to obtain comparable data. In England the rates were in sequence as follows:-1301, 948, 1084, 104-6, 109-7, 91-2, 96.5, 97-2, 891, 79-9; average for 1911-20, 100'4.

In Scotland the corresponding figures:-1125, 1055, 109-6, 1106, 1265, 971, 107-5, 99-8, 101-6, 920; average for the years. 1911-20, 106.3.

For both countries the rates for 1922 and 1923 were :-England, 77 and 69.2; Scotland, 1014 and 78.9.

On the whole, for the last decade, Scotland compares somewhat unfavourably with England.

Analysis of the Infantile Death-Rate into periods.

To get to closer quarters in the comparison, it is important to discover just at what periods of the infant life the comparison becomes unfavourable to Scotland.

(a) Children under one week.-The deaths of children under one week run at much the same level in both countries. In Scotland the deaths vary a little above or below 25 per 1000 births. In England they vary a little above or below 23 per 1000 births. In neither country within the last 10 years is there any indication of a serious decline in the death-rate of the first week after birth. As already pointed out, the infant of the first week is seriously affected by temperature and, as precautions against cold can be taken, even in a cold climate, it is to be expected that better provision for maternity will affect for the better even the death-rate for the first week. It is certain that, in neither country, do the precautions, as yet applied, tend to produce a decline in the death-rate.

(b) One week and under four weeks.-Much the same statement can be made about the death-rate from one week to four. In Scotland this death-rate varies a little above or below 14 per 1000 births. In England it varies a little above or below 13 or 13.5 per 1000 births. In both countries there is a slight tendency downward, but in neither country is this tendency well marked.

(c) Total under one month.-The death-rate of children under one month stood at about 40 in both countries in 1912; but in England the tendency is slightly downward, the rate reaching 40 again in 1919, but declining in 1920 to 35. In Scotland, from 1912 to 1915, the rate stood at a little over 40; but from 1916 onwards it has risen above 40 only once and in 1921 it stood about 38.5.

The late Dr. Ballantyne looked upon the first month of life as rather a continuation of the pre-natal life than a beginning of the post-natal life. He considered that in the first month of life the pre-natal influences were still at work, and that the effective individual life of the child began more properly with the second month.

In the comparison between Scotland and England it cannot

be said that either country has much advantage for the first month of the child's life. Temperature, of course, is always a potent factor, just as in the first week of life.

(d) From one month to three months.-In both countries the death-rate from one month to three months shows a very slight tendency downward. In Scotland since 1912 this rate has only once exceeded 20 per 1000, and this was in the year 1915, when possibly the fall in the birth-rate accounts, at least partially, for the rise of the death rate. From that year onwards this rate has kept well below the 20 line and in 1921 stood at 15 per 1000. In England from 1912 the rate has risen only once to 20 and in 1920 stood at 16 per 1000. In 1921 it stood at 14.84. In 1922 the rate for Scotland was 151, for England 12.68.

In the period from one month to three the effect of environment begins to tell. Children at that age do not take scarlet fever, which is now in any case a minor cause of infantile death; but they do take measles and whooping-cough and go under very readily, whoopingcough and bronchitis being more deadly than measles in the first three months of life. But, so far, in both countries, the tendency is downwards.

(e) Total under three months.-The death-rate for children under three months shows a small but definite tendency downwards in both countries. In Scotland it stood at slightly below 60 per 1000 in 1912, and, with the usual variation in 1915, it dropped well below the 60 line in 1916, and in 1921 it stood at about a little over 53 per 1000. In England the same rate went up to 65 in 1911, then dropped to 56 in 1912, and since then has sunk down to 51 per 1000 in 1920, and in 1921 it fell to 50. The decrease in England is markedly steeper than the decrease in Scotland. England seems to have a slight advantage in the second and third months of life.

(f) Three months to six months.-In Scotland, since 1912, the death-rate from 3-6 months has only once exceeded 20 per 1000, and this, as in the other case, was in 1915. The tendency of this rate is quite definitely downwards, and in 1921 stood at about 146 per 1000. In England this rate has only once reached the 20 per 1000 line, and has shown since then the same downward tendency. In 1921 it stood at 14:02. In 1922 the Scottish rate was 157. The English rate was 11.03. In the last two years England thus shows a somewhat more rapid rate of decrease for the period three to six months.

(g) Six to twelve months.-In Scotland the death-rate for children of 6-12 months was about 28.5 per 1000 in 1912. It rose considerably in 1915, when, as in the other cases, the birth-rate was disturbed; but, since 1916, with one slight variation, it has never gone above the 30 per 1000 line and it sank to 20 per 1000 in 1920 and 21 per 1000 in 1921. In England this rate in 1911 was 39 per 1000. In 1912 it dropped to 24 per 1000, rising again in 1915 to 35 per 1000. In 1920 it stood at 17 per 1000 and in 1921 at 18.6 per 1000; in 1922 at 19-28. Thus once more Scotland compares unfavourably with England at the period of 6-12 months.

In this detailed comparison England has little advantage in the earlier periods, a recognisable advantage in the middle period and a

distinct advantage in the last period. The levels in England seem now to be shifting more rapidly than the levels in Scotland. Probably this means that the leading causes of death in England are more variable and depend upon more manageable factors.

Comparison of Scotland and North England.

For the purposes of the Registrar-General's Report, England and Wales are divided into the following:-North, Midlands, South, Wales and London. The North includes the counties of Cheshire, Lancashire, Yorkshire (the three Ridings), Durham, Northumberland, Cumberland and Westmorland. The population of this area is much greater than the population of Scotland; but, in many respects, the area is capable of fair comparison with Scotland. There are county boroughs, urban districts and rural districts. There are many varieties of industry: mining, textiles, iron works, shipping, and agriculture. The climatic conditions resemble Scotland more than they resemble the South of England or even the Midlands. The configuration is a combination of highlands, moorland, and seaboards of similar exposure. On the whole, the conditions with some closeness resemble the conditions of Scotland as a whole. It is possible, at least roughly, to compare the county boroughs, the urban districts and the rural districts of North England with the larger burghs, the smaller burghs, and the county districts of Scotland.

County boroughs and larger burghs.-In Scotland the infantile death-rate in the larger burghs stood at 130 per 1000 in 1911. It fell to less than 120 in 1912. It rose to 144 in the fallacious year 1915; but from 1916, when it fell to 107, it has, with the single exception of 1917, kept below the 110 rate, and in 1921 it stood at 101; 1922, at 104.

In North England the rate in the county boroughs stood at 152 per 1000 in 1911; it dropped to 113 in 1912; it rose to 131 in 1913 and maintained that level to 1915; it dropped in 1916 to 114; it rose to 120 in 1918, and since that time has rapidly fallen to 101 in 1921 and 96 in 1922. The county boroughs show a more rapid rate of decline than the larger burghs of Scotland, but the original level was higher. In Scotland the smaller burghs show a similar course, but they started in 1911 with a rate of 107 per 1000. With the exception of the years 1914 and 1915, this rate has kept well below the 100 line from 1916 onwards. In 1920 it fell as low as 82.5, rising to 85.5 in 1921 and 93.5 in 1922. The course of the curve in the smaller burghs is on the whole parallel to the course of the curve for the urban districts of North England. But in those districts the rate started from 147 in 1911; it dropped to 101 in 1912; it rose to 124 in 1913. In 1916 it dropped to 100, rising again in 1918 to 117. In 1920 it stood at 94 and in 1921 at 95, dropping to 88 in 1922. As in the county boroughs, so in the urban districts; the recent decrease is more rapid than the decrease in the smaller burghs of Scotland.

In Scotland the infantile rate for the county districts in 1911 was 91 per 1000. With the exception of the year 1915, this rate, except on two occasions, has kept below the 90 per 1000 limit. In 1920 it stood at 79, and in 1921 it stood at 755, but it has risen in 1922 to 84.