A STUDY IN VITAL STATISTICS. 
VITAL STATISTICS may be defined as that branch of 
statistics which studies the population of a country, as 
distinct from the acts or products of the population. The 
term corresponds almost exactly to what the Germans call 
statistics of population (Bevolkerungsstatistik) and to what 
some French writers call demography (demographic). There 
are two great divisions of vital statistics: one investigates the 
condition of the population at a certain date, and thus disregards 
the element of time and change; the other investigates the 
changes in the population between two points of time. The 
statistical material for the study of the first branch is con- 
tained in our censuses, state and national, which seek to give 
an instantaneous photograph of the population at a certain 
moment. Practical difficulties may keep this from being more 
than a purely ideal goal, like a limit in mathematics; but it is a 
goal that may be approximated indefinitely. Our censuses 
also give much aid in studying the second division of vital 
statistics. From a comparison of two censuses we may draw 
certain conclusions as to the changes that have occurred in 
the interim. Thus if the population has increased, we may 
infer with certainty that the increase from births and im- 
migration has exceeded the decrease from deaths and emigra- 
tion by the amount of the total increase. But censuses do not 
of themselves furnish any analysis of the changes of population, 
any determination as to the relative importance of the various 
causes at work. Hence the need for continuous records of 
certain important and frequently recurring changes. Such 
records now commonly include the number of births, im- 
migrants, cases of sickness, deaths, emigrants, marriages and 
divorces. The phrase vital statistics is often used in a limited 
sense to denote a comparison of these changes with the total 
69 70 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII- 
population and with one another; it is even sometimes nar- 
rowed to a study of births, marriages and deaths. In the 
foregoing analysis I have sought to justify its use in the wider 
sense and to show that vital statistics is a distinct and clearly 
marked branch of statistics, if indeed it be not, as is claimed 
by several French writers, a separate science. 
The materials for the study of vital statistics are contained 
in the various censuses and in periodical registration reports 
of some form. Our national censuses are tolerably satis- 
factory, but the continuous registration of important changes, 
except the immigration over sea, is left to the separate states 
and is in the main very imperfect or entirely wanting. As 
there is no record of overland immigration and no record 
of the counter-current of emigration, it is obvious that 
the sources for the study of vital statistics in the United 
States are extremely incomplete. With the exception of 
Spain, all the great states of Europe publish their yearly 
record of marriages, births and deaths. Even the imperfect 
statistics of Russia are in this respect superior to those of the 
United States. Six years ago the national government, to 
meet a pressing practical need for information, entered this 
field so far as to collect and publish the existing statistics of 
marriages and divorces for the preceding score of years.1 The 
result of the investigation of marriages, the compiler himself 
declared, 
is thoroughly incomplete and unsatisfactory. Very few states have 
any registration system by which marriages are recorded. ... In 
some the work of compilation at the central office is so carelessly 
and inaccurately done as to detract greatly from their value. . . . 
Twenty-one states provide for returns to some state officer of 
marriages celebrated. These returns as a rule, however, give but 
few facts relative to the persons married, and the facts that are 
given are not identical and are compiled so carelessly as to be 
nearly worthless.2 
1 A Report on Marriage and Divorce in the United States, 1867 to 1886; by 
Carroll D. Wright, Commissioner of Labor. Washington, 1889. 
2 Report, pp. 18, 19. No. 1.] 
A STUDY IN VITAL STATISTICS. 
71 
In a recent examination of the volume 1 I sought to analyze 
and interpret its figures so far as they relate to divorce. The 
object of the present paper is to set forth some results of 
my further study in the field of vital statistics as above defined. 
The sources are so fragmentary as to prevent any attempt at 
presentation of the conclusions in systematic form. Since 
our judicial statistics of divorce are much better taken and 
preserved than our civil statistics of births, marriages and 
deaths, for a long time to come the former must occupy a dis- 
proportionately large place in any treatment of the vital statis- 
tics of the United States. 
For any detailed study of vital statistics it is important to 
know the population for each year, in order to compare the 
number of births, deaths, marriages or divorces with the whole 
number of people. As this population is ascertained only 
once in ten years, or in some few states once in five years, the 
population for the intermediate years must be found indirectly. 
The method of assuming that in each year of the interval the 
population changed by the same amount, one-fifth or one-tenth 
of the change observed in the five or the ten years between 
censuses, is the most simple and obvious one, and while not 
entirely accurate, it is sufficiently so for the purposes of this 
article. By this method the population of the country or of 
any state, county or city may be found for any year between 
1870 and 1890. But we cannot suppose that this method 
will give even approximately accurate results between i860 
and 1870, because the rate of increase during the Civil War 
cannot have been as great as it was just after its close. Five 
states, Massachusetts, Rhode Island, New York, Iowa and 
Kansas, took a state census in 1865 ; and Michigan took one 
in 1864, from which the population of that state in 1865 may 
be estimated. These six states, embracing in i860 over one- 
fifth of the population of the country, increased in five years, 
1860-65, from 6,817,653 to 7,035,740, or only 3.2 per cent. 
At first thought it might seem fairest to apply this ratio to 
1 The Divorce Problem. Columbia College Studies in History, Economics and 
Public Law, No. i. New York, 1891. 72 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
the whole country and compute the population of the United 
States in 1865 by assuming it to have been 3.2 per cent 
greater than the population in i860. But it is to be observed 
that the last three of these states were growing with great 
rapidity through the decade, so that the total population of 
the six states increased, 1860-70, 29.1 per cent, while the in- 
crease of the whole country was only 22.7 per cent. On the 
other hand the population of the three eastern states slightly 
decreased, 1860-65, from 5,286,421 to 5,279,814. Hence 
it seems fairest to assume that the growth in the western 
states was neutralized by the loss of population in the southern 
and border states, and that the population of the country as a 
whole was the same in 1865 as in i860. On this assumption 
the population for each year 1865-70 may be estimated in 
the same manner as for subsequent years. 
After finding the total population for any year, the birth rate, 
death rate, marriage rate or divorce rate for that year may be 
computed in case the total number of births, deaths, marriages 
or divorces for that year is given by a registration report or 
otherwise. This is done by dividing the number of births, 
deaths, marriages or divorces for the year by the population for 
the year and multiplying the decimal thus obtained by 100, 1000 
or 100,000. The result obviously expresses the number of 
births, deaths, marriages or divorces in each 100, 1000 or 
100,000 of the population, provided that these changes are dis- 
tributed with entire uniformity, an assumption never more than 
approximately correct. The advantage derived from such a 
computation is that it makes unnecessary further consideration 
of differences in population. It thus renders easy a comparison 
of the amount of marriage or divorce in communities or states 
widely different in their populousness, as well as within the same 
community or state for different years. In the following pages 
the number 1000 has been uniformly employed as the basis of 
the rates except for divorce, where, in order to avoid decimals, 
100,000 has been substituted. In some cases the graphic 
method of presenting results or illustrating relations has been 
used, as easier of apprehension than columns of figures. No. 1.] 
A STUDY IN VITAL STATISTICS. 
73 
Marriage Statistics. 
Only in five states, Vermont, Massachusetts, Rhode Island, 
Connecticut and Ohio, and in the District of Columbia, can 
the number of marriages be obtained with approximate com- 
pleteness for each one of the twenty years.1 To these may 
be added Illinois, where the records since the Chicago fire 
are apparently complete. Any results that may be obtained 
from an examination of these states or parts of them are 
hardly to be regarded as applicable south of the Ohio or 
west of the Mississippi, where the conditions are widely 
different. 
How does the marriage rate in these states compare with the 
rate in European comitries ? The following table gives the 
number of persons marrying to every thousand of population 
in several European countries 2 and in these states for the year 
1876, the mean of the twenty-year period. 
Marriage Rates in 1876. 
Hungary .... 19.8 
German Empire . . 17.0 
England and Wales . 16.6 
Austria 16.2 
Italy . ’ 16.2 
France 15.8 
Scotland . . . . 15.0 
Ireland 10.0 
District of Columbia . 20.0 
Rhode Island . . . 17.8 
Ohio 17.5 
Illinois . . . • . 17.1 
Vermont 15.9 
Massachusetts . . . 15.4 
Connecticut . . . 13.2 
Although no trustworthy conclusions in detail can be drawn 
from a single year, and although the pressure of the hard times 
in 1876 was probably somewhat more influential in the United 
States than in Europe in lowering the marriage rate, yet the 
general conclusion I believe to be warranted, that in the north- 
eastern and north-central states the average marriage rate is 
only slightly higher than the average rate in Europe; and that 
if we except Ireland on the one side, and Connecticut on the 
1 Report, p. 129. 
2 The figures for the European states are derived from the appendix to von 
Oettingen’s Moralstatistik, third edition, pp. i-iv. 74 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
other, as either abnormal or imperfectly reported, the variations 
from this average rate are not very much wider on one side of 
the Atlantic than they are on the other. 
Does the marriage rate in this country show any tendency to 
change ? In most European countries the marriage rate has 
somewhat decreased in the past twenty years. A similar 
falling off is apparent in those states of this country about 
which the facts are obtainable. The following table gives the 
marriage rate for the first and the last year reported and the 
decrease. 
Decrease of the Marriage Rates of Certain States in Twenty Years. 
Vermont 
Marriage Rate 
in 1867. 
Marriage Rate 
in 1886. 
i5-6 
Decrease. 
2.2 
Massachusetts . . . 
. . . 21.6 
18.0 
3-6 
Rhode Island . . . 
. . . 24.4 
17.6 
6.8 
Connecticut . . . 
. . . 19.5 
i5-8 
3-7 
District of Columbia . 
• • • 33-3" 
20.7 
12.6 
Ohio 
. . . 23.8 
16.4 
7-4 
Illinois1 
. . . 23.6 
I7*7 
5-9 
As there has been some decrease in every case, it is fair to 
presume that the same is true of the other states adjacent to 
these. It might be objected that the rate in 1867 was 
probably increased by the celebration in that year of some 
marriages which would have occurred earlier had they not been 
postponed by the war. But a comparison of the rates in these 
states for each year of the twenty confirms the conclusion 
indicated by the table above, that there has been a general 
though irregular downward tendency through the twenty years. 
Is the decrease of marriage more marked in the cities or in 
the country ? It is difficult to find a clear and general answer 
in the figures. The state of Ohio makes no returns by coun- 
ties, and hence no comparison between city and country is 
there possible. In Massachusetts I have selected counties 
containing almost no large city, and have compared their 
1 Exclusive of Cook County, containing Chicago, in which the records were 
destroyed in 1871. No. 1.] 
A STUDY IN VITAL STATISTICS. 
75 
average marriage rate with that of Suffolk County (contain- 
ing Boston). In Connecticut I have compared the average 
marriage rate of the three counties which contained in 1890 the 
eight largest cities, with the average rate of the other five 
counties of the state. 
Decrease of Marriage Rates in Urban and Rural Districts Compared. 
Districts Mainly Rural. 
Marriage Rate 
in 1867. 
Marriage Rate 
in 1886. 
Decrease. 
Vermont 17.8 
15.6 
2.2 
Massachusetts (7 rural counties) 19.9 
16.1 
3-8 
Connecticut (5 rural counties) . 19.6 
4-5 
Illinois (outside Cook Co., 1872-86) 17.6 
17.7 
0 
Districts Mainly Urban. 
Marriage Rate 
in 1867. 
Marriage Rate 
in 1886. 
Decrease. 
Rhode Island 
. 24.4 
17.6 
6.8 
Massachusetts (Suffolk Co.) . 
. 26.7 
20.6 
6.1 
Connecticut (3 urban counties) 
• 19-4 
16.1 
3-3 
Illinois (Cook Co., 1872-86) . 
• 31-9 
23.6 
8-3 
District of Columbia . . . 
• 33-3 
20.7 
12.6 
As Vermont contains no city with a population of 15,000 or 
over, the whole state has been treated as a rural district. 
In Rhode Island 77 per cent, and in the District of Columbia 
83 per cent, of the population were classed as urban by the 
census of 1880; accordingly both are treated as urban in the 
table. The general result of this comparison is to indicate 
that the decrease in the marriage rate has been somewhat 
greater in cities than in the country. The exception in the 
case of Connecticut may be due to the imperfect registration 
of marriages in that state, and to the fact that omissions have 
been more numerous in the urban counties. The rural rate is 
about normal, while the urban rate is very low. 
Is any change occurring in the average age at which marriage 
is contracted? No suggestion of an answer to this question 
has been found except in Massachusetts, where the average 
age of the parties is reported. As the main significance 76 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
attaches to the age at which people enter upon marriage, I 
give here only the average age of those who married for the 
first time. 
Average Age of Marriage in Massachusetts. 
1871. 
1890. 
Increase. 
Bachelor grooms . . 
. 26.3 
27.2 
•9 
Maiden brides . . . 
• • • 23.5 
24-3 
.8 
Does a study of the marriage rates reveal any causes of the 
changes that have occurred? Such a study is more readily 
made by the aid of a diagram than from a column of figures. 
As we have reports of the marriages in Massachusetts for a 
long period of time and as that state is fairly typical of the 
part of the country to which our study is confined, I present 
here a diagram of the marriage rates in that state for the 
past forty years.1 
Marriage Rates 
This diagram shows clearly the influence of the Civil War 
and of the three depressions of trade and industry in dis- 
couraging marriage. At the same time it illustrates the general 
downward tendency of the rate. Somewhat similar but less 
marked fluctuations are shown in Rhode Island and Connec- 
ticut for the period since the war. On the other hand, in the 
1 Compiled from Mass. Registration Report, 1889, p. 272 ; 1890, p. 245. No. 1.] 
A STUDY IN VITAL STATISTICS. 
77 
predominantly agricultural state of Vermont there is almost no 
trace of any influence of the depression of trade. This sug- 
gests that the hard times discouraged marriage more powerfully 
in commercial centers than in rural districts; and the inference 
is confirmed by a comparison of city and country rates in 
Massachusetts and Illinois. In the three years, 1873-76, the 
Suffolk County rate fell off 7.8, while the average rural county 
rate fell only 3.1. The Cook County rate fell, 1873-76, no less 
than 14.7, while in the rest of Illinois the fall in those years 
was only .7. In all these cases the influence of the hard times 
would be shown even more clearly if only first marriages were 
included. The marriage of bachelors is a much more sensitive 
barometer of the economic condition and hopes of a community 
than the whole number of marriages. With economic distress 
the proportion of marriages by bachelors sinks and that of 
in Massachusetts. 
marriages by widowers rises. Thus in Massachusetts, in the 
three years, 1873-76, the number of bachelors marrying fell off 
24 per cent, but that of widowers remarrying decreased only 
11 per cent. 
Divorce Statistics. 
As before stated, the divorce rates given in this article 
represent the average number of persons divorced (not the 78 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
number of divorces) to every 100,000 of the population. 
Since the marriage rates give the average number of persons 
marrying, and not the number of marriages, to every 1000 of 
the population, it is obvious that the frequency of marriage 
and of divorce may be readily compared by moving the 
decimal point in the divorce rate two places to the left. 
How does the divorce rate in the United States as a whole 
compare with the rate in other countries ? The following table1 
will give an approximate answer to the question. As the num- 
ber of divorces is quite generally increasing in Europe and 
this country, I have chosen the year 1886 instead of 1876 for 
comparison. Where no date is given, 1886 should be under- 
stood. 
Divorce Rates (including Separation) in Various Countries. 
Ireland 28 
Italy (1885) .... 3.75 
England and Wales . . 3.79 
Canada 4.81 
Australia (including New 
Zealand and Tasmania) 11.14 
German Empire . . 25.97 
France ..... 32.51 
Switzerland .... 64.49 
United States . . . 88.71 
Japan2 608.45 
The divorce rate of Japan is especially interesting, since I 
believe Japan is the only non-Christian country that collects 
and publishes its vital statistics. Whether it is even approxi- 
mately a fair type of other non-Christian countries or whether 
its high rate is due to local and exceptional conditions, I 
am unable to state. In the year 1886 there were in Japan 
315,311 marriages and 117,964 divorces, more than one di- 
vorce to every three marriages and more than four and a half 
times as many divorces as there were in the United States, 
although the population of Japan was only about two-thirds 
as great. But among the professedly Christian countries the 
United States has the largest divorce rate. 
1 The data from which these rates have been computed for European countries 
are mainly found in the valuable appendix to the Report on Marriage and 
Divorce. Those for Australia have been kindly furnished me by T. A. Coglan, 
government statistician of New South Wales. 
2 Resume Statistique de l’Empire du Japon (Tokio, 1891), pp. 10, 17. No. 1.] 
A STUDY IN VITAL STATISTICS. 
79 
What change, if any, is taking place in our divorce rate ? 
This question may be most intelligibly answered by the 
following diagram expressing the divorce rate of the United 
States for each of the twenty years, 1867-86, in the form of a 
curve, or more accurately, a broken line. The amount "of 
divorce to every 100,000 of the population for each year is 
indicated by the distance from the bottom of the diagram to 
the broken line, and the prevailingly upward tendency of the 
line expresses the gradual rise in the divorce rate. In only 
five of the nineteen cases does the line descend. This shows 
that in only five years was the rate lower than it had been the 
year before, while in fourteen cases there was an increase. 
And the average annual decrease in these five years was 
markedly less than the average annual increase in the fourteen 
years. 
Divorce Rates in the United States. 
May any causes influencing the divorce rate be determined 
from the figures? A study of the rates for the years of 
commercial depression will show that the increase of divorce 
in those years was checked or turned into a decrease. To 
show how closely the divorce curve of the United States 
and the marriage curve of Massachusetts correspond, they 
should be placed side by side. But even then the similarity 80 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
of the two curves in details is obscured by the fundamental 
difference that the general trend of the marriage curve is 
downward and that of the divorce curve upward. For purposes 
of comparison this difference may be eliminated on the 
assumption that the complex forces tending to decrease 
marriage and the complex forces tending to increase divorce 
were uniform in their action throughout the twenty years. 
The effect of applying such an assumption is practically so to 
change the curves that the ends of each shall lie in the same 
Immigration, Divorces and Marriages. 
horizontal line, while the variations from year to year are but 
little affected. The above diagram presents to the eye the 
curves thus altered. For confirmation the curve of immigra- No. 1.] 
A STUDY IN VITAL STATISTICS. 
81 
tion to the United States, also changed from an ascending to 
a horizontal curve, has been added to the others. As the year 
for which the immigration is reported ends June 30th, the 
angles in the immigration curve have been placed midway 
between the years. 
In the divorce curve it will be seen that two rates are given 
for each of the four years 1875-78. An explanation of this 
will be offered later. Suffice it to say at present that between 
1874 and 1876 the decrease of the divorce rate would probably 
have been greater than is indicated by the unbroken line, had 
it not been for the number of divorces granted in Utah in 
those years under a lax local law, and that the limit of de- 
crease without that law has been found and is indicated by the 
dotted line. But for the Utah divorces it is practically certain 
that the divorce curve of the country for those years would 
have lain somewhere between the two limits, one expressing 
the divorce rate actually found and the other what that rate 
would have been had no divorces been granted in Utah to 
parties residing at a distance, i. e. in a non-adjacent state or 
territory. 
The general similarity of the three curves is so marked that 
one is almost compelled to admit that in large measure they 
are influenced or controlled by the same forces. Especially 
clear is the great influence of the depression in trade from 
1873 to 1879 and the slighter depression from 1882 to 1886. 
The relation between the United States divorce curve and 
the Massachusetts marriage curve is quite remarkable. From 
1870 to 1871 the adjusted marriage curve rose while the 
divorce curve fell slightly. The influence of the panic of 1873 
was felt the following year in diminishing the number of mar- 
riages in Massachusetts, but not till 1875 was it shown in the 
divorce rate. It should be remembered, however, that a divorce 
speaks from the date of the decree and not from the date the 
suit is begun. The decrease of decrees entered in 1875, there- 
fore, would very probably correspond to a decrease of suits 
begun in 1874. The adjusted divorce curve fell between 1877 
and 1878 while the marriage curve rose. This may be explained 82 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
by the repeal of the lax Utah law in February, 1878. With 
these three exceptions, two of which, I believe, are only 
apparent, the two curves uniformly ascend and descend to- 
gether and reach their maxima and minima in the same years. 
The inference is irresistible. Depressions in trade have had a 
tendency to decrease divorces as well as marriages. In the 
great mass of the population they have discouraged change, 
have compelled men and women “ in whatsoever state they were, 
therewith to be content,” or at least to abandon or postpone 
the idea of change. 
This influence of commercial depressions on the divorce 
rate has never before been suggested, so far as I am aware. 
It is, indeed, somewhat doubtful whether it exists in European 
countries. In England, at least, in the years in which the 
number of marriages has been diminished by hard times, 
the number of divorces has not fallen off, but rather in- 
creased. In 1879 aRd 1885 the divorce rate in England and 
Wales was higher than for any one of the other eighteen years, 
while in these two years and in 1886 the three lowest marriage 
rates of the period were presented. The expense and delay 
involved in procuring a divorce there are so great that only 
somewhat wealthy persons can go into court, and they do not 
feel so severely the burden of a financial crisis. This con- 
jectural explanation derives some support from the fact which 
a French statistician of eminence 1 claims to have proved, that 
such periods of distress in Great Britain, while checking mar- 
riage among the poor, are attended by an increase of marriage 
among the rich. This difference between the effect of hard 
times in England and in the United States, together with the 
very rapid increase of divorce among the Southern negroes, and 
the fact that only about one wife in six of those obtaining 
divorces receives any alimony, are among the indications that 
divorce has become very frequent and perhaps most frequent 
among our lower middle classes, and has reached for weal 
or woe a lower stratum than perhaps anywhere in Europe. 
1 See Dictionnaire Encyclopedique des Sciences Medicales, article “ Grande 
Bretagne,” by Dr. L. A. Bertillon, § 32. No. 1.] 
A STUDY IN VITAL STATISTICS. 
83 
Is divorce more common in cities or in the country ? To 
this question an unhesitating answer may be made. Almost 
everywhere in this country the divorce rate of a large city is 
higher than the rate in the neighboring rural districts. The 
evidence of this statement may be most conveniently pre- 
sented in connection with the discussion of the following 
topic. 
Is the difference between city and country divorce rates increas- 
ing or decreasing ? In the study of this question I have 
selected the following states, either as containing large cities 
or as fairly typical of the part of the country in which they lie: 
Massachusetts, New York, Pennsylvania, Georgia, Ohio, Illinois, 
Missouri and California. As the Cincinnati records have been 
lost, Cleveland, the second city of Ohio, has been compared 
with the rest of the state. It would be natural to choose New 
Orleans for the typical Southern city, but the Louisiana 
records are very imperfect, and hence I have selected Georgia, 
where the records seem to be excellent. In default of any 
single large city in that state, the average rate for the five 
counties in which are situated Atlanta, Augusta, Columbus, 
Macon and Savannah, its five largest cities, is compared with 
the average rate of the other counties. In New York State, 
the counties of New York and King’s are treated as a single 
population center, whose average divorce rate is compared 
with the rate outside. So in Pennsylvania, the average rate 
of Philadelphia and Allegheny Counties, the latter including 
the cities of Pittsburg and Allegheny, has been compared with 
that of the rest of the state, and in Massachusetts the rate of 
Suffolk County, containing Boston, has been compared with 
the average rate of eight counties containing almost no large 
city. The results of the comparison are stated in tabular form 
on the next page. A minus sign prefixed to a number in the 
column headed “ increase,” indicates that the number should 
be read as a decrease. The Massachusetts rates cannot be 
compared with those of the other states, because in the former 
the base is different and more accurate, viz., 10,000 married 
couples rather than 100,000 people. 84 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
Increase of Divorce Rates in City and Country in Twenty Years. 
State. 
City Divorce 
Rates. 
Country Divorce 
Rates. 
Increase in 
Cities 
in Twenty 
Years. 
Increase 
in the 
Country 
in Twenty 
Y ears. 
1867. 
1886. 
1867. 
1886. 
Massachusetts 
34-4 
42.3 
16.7 
29.6 
7-9 
12.9 
New York . 
59-8 
35-6 
29-3 
35-8 
— 24.2 
6-5 
Pennsylvania . 
29.0 
57-4 
39-1 
43-3 
28.4 
4.2 
Ohio . . . 
132.5 
164a2 
78.2 
99-1 
3i-6 
20.9 
Illinois . . . 
2S5-61 
182.8 
io9-7 
*34-7 
1 
bo 
25.0 
Georgia. . . 
29-S 
64.1 
22.2 
34-i 
34-6 
11-9 
Missouri . . 
00 
vb 
134-7 
45-° 
91.1 
46.8 
46.1 
California . . 
112.9 
288.8 
79-9 
I54-5 
17 5-9 
74.6 
Of course no conclusions can be drawn from only two years 
unless they are fair examples of the general trend. So in 
each case the rate has been computed for every year of the 
twenty, and it is believed that the foregoing are trustworthy 
indications of the variations that have occurred in the interim. 
The table shows that in Pennsylvania, Ohio, Georgia and 
California the increase of divorce has been more rapid in city 
than in country; that in Missouri the rate of increase has been 
practically the same in the two regions; while in Massachusetts, 
New York and Illinois the increase has been more rapid in the 
country. It is noteworthy that New York, Brooklyn and Chi- 
cago have shown a marked decrease of divorce rate in twenty 
years, the rate in Chicago in 1886 being not very much higher 
than in Cleveland and much lower than in San Francisco. 
The table also shows how generally the city rate is higher 
than the country rate. In only two of the sixteen instances 
does the city rate fall below that of the rural districts, and 
if we extend our observation over the twenty years for each 
of these eight states, it appears that in only nine of the one 
1 This is for 1872, as no earlier record is extant. 
2 This is for 1885, since the rate in 1886 was abnormally and inexplicably low, 
lower than for any other year of the twenty, so that I am forced to doubt the 
completeness of the record. No. 1.] 
A STUDY IN VITAL STATISTICS. 
85 
hundred and sixty instances does the city rate fall below the 
country rate. If these be accepted as fair types of the country 
at large, we may say that in about ninety-five per cent of the 
cases the divorce rate of a large city is greater than that in 
the other counties of the state. 
To what extent can legislation be shown to have influenced 
the divorce rate ? In the recent study of the divorce question 
already cited I examined the evidence offered by Mr. Wright1 
to show that legislation had influenced the divorce rate in 
fourteen states and territories and came to the conclusion 
that the evidence was insufficient to prove a marked influence 
of legislation in the majority of these cases. In five the 
evidence failed when tested; in five others the influence 
appeared on scrutiny to be slight or temporary; in two the 
evidence could not be analyzed in detail; so that in only two 
was the influence of law clear and considerable. In the light 
of some press criticisms and of newly discovered evidence, this 
conclusion requires some modification in respect to Utah and 
Massachusetts. 
Utah. This territory formerly had a very lax divorce law, 
allowing decrees to issue to persons who merely declared an 
intention to become residents and alleged incompatibility of 
temper as a ground. The rapidity, secrecy and cheapness 
with which a divorce decree of seeming validity could be 
obtained by non-residents led to the flooding of the Utah 
courts in some five counties with cases sent thither by 
lawyers doing business in Eastern cities. In 1878 the law 
was amended and the number of divorces in the territory fell 
from 914 in 1877 to 122 in 1879. This seemed to indicate 
that the amended legislation had reduced the annual number 
for the whole United States by about 800. My reason for 
questioning that conclusion, however, was that coincident 
with the increase of divorce in Utah in 1875, there was a 
decrease in Chicago and New York, from which cities most 
of the Utah cases were sent ; and coincident with the repeal 
of the Utah law there was an increase in the number of 
1 Report, pp. 150-157. 86 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
divorces in those cities. Add to this the fact that both New 
York State and Illinois, outside the largest cities, showed a 
steady increase during the years of numerous Utah divorces, 
and the evidence seemed convincing that we had to do with 
a deflection of the divorce current rather than with a decrease 
of its volume.1 
At that time I was ignorant of the influence of hard times 
in checking divorce — a fact discovered by subsequent study 
— and ignorant also, of course, that this influence is more 
potent in large centers than in rural districts. These two 
discoveries throw new light on the question. The decrease 
of the divorce rate in New York City and Chicago after 
1874 may be due either to the hard times or to the send- 
ing of divorce cases to Utah ; the increase of the divorce 
rate after 1878 may be due either to the repeal of the Utah 
law or to the revival of business. I am disposed to believe 
that both influences were at work, but that the depression of 
business and the reduction of wages were far more influential 
than the deflection of the divorce current to the Western 
territory. 
Mr. Wright has kindly furnished me with an unpublished 
statement of the places in which the parties divorced in Utah 
were married. Twice as many were married in New York as 
1 It should be noticed that these Utah decrees have in several instances come 
before the courts outside that territory for construction, and in accordance with 
a well-settled principle of law, have been uniformly held null and void for want of 
jurisdiction over the parties. For example, a physician living in Minnesota 
obtained a divorce in Utah in 1876, and in the belief that he was thereby released 
from the ties of his former marriage, a belief confirmed, as he averred, by com- 
petent legal advice, he speedily married again. A year later he was indicted for 
polygamy. On the trial the Utah decree was refused admission as evidence, the 
court holding that he must be presumed to have known the law and to have been 
aware that a decree issuing from a court without jurisdiction was invalid, and he 
was found guilty. He was sentenced to the state prison for two years, and on 
appeal to the supreme court of Minnesota the judgment and sentence were 
unanimously affirmed. State vs. Armington, 25 Minn. 29. Similar decisions in- 
volving these Utah decrees have been reached on criminal trials in New York, 
Indiana and Iowa, and in civil suits in Massachusetts, Kansas and Tennessee. 
The earliest, I believe, of these decisions was reached in 1877, so that they could 
hardly have exerted much influence in reducing the number of applications before 
the law was changed a year later. No. 1.] 
A STUDY IN VITAL STATISTICS. 
87 
in any other state of the Union. This may be plausibly ex- 
plained as due in great measure to the stringency of the New 
York law, which allows divorce only for one cause. Hence 
there would be in this state a great number of deserted or 
cruelly-treated husbands or wives, in addition to those merely 
discontented with their choice, and many of these might be 
allured by the advertisements of the divorce lawyers. The 
second state on the list was Illinois. As its divorce law is lax, 
a majority of those residing there and seeking divorce in Utah 
might have obtained a divorce under Illinois law, though less 
readily, secretly or cheaply. For the purposes of illustration 
assume that all the divorces of parties married in Illinois and 
City and County (Illinois) Divorce Rates. 
divorced in Utah were granted to residents of Chicago; or, what 
amounts to the same thing, assume that the number of such 
divorces to persons not residing in Chicago was equal to the 
number granted to residents of Chicago who had been married 
outside of Illinois. Assume also that if all these divorces had 
not been granted in Utah they would have been granted in the 88 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
local court. On these assumptions the divorce curves of Cook 
County and of the rest of the state may be constructed both as 
they actually were and as they would have been had the Utah 
divorces of parties married in Illinois all been granted in Cook 
County. These curves are shown by the preceding diagram. 
Note the very rapid decrease of the divorce rate in Chicago 
from 1872 to 1876, probably due in large measure to industrial 
depression. Note that this decrease is confined to the city and 
does not appear in the rest of the state. Note also the influence 
of the possible diversion of divorces to Utah in accentuating 
that decrease. It is probable that some but not all of the 
divorces of Illinois marriages granted in Utah would have been 
obtained in Cook County if the cases had not been sent west. 
If this is so, the true divorce curve of Cook County, as it would 
have been but for the advantage taken of the Utah law, would 
be somewhere between the unbroken line, indicating the curve 
of the actual divorce rate, and the dotted line, indicating the rate 
found by adding to the divorces granted in Chicago in each year 
the divorces granted that year in Utah to parties married in 
Illinois. Attention may also be called to another inference from 
the diagram, that the divorce rate in Chicago has decreased since 
the fire, while the rate in the rest of Illinois has somewhat risen. 
Massachusetts. In this state there was a noticeable increase 
of divorce between 1872 and 1874, which Mr. Wright stated1 
was probably due to a change of law in 1873, whereby the 
period of desertion necessary to give ground for divorce was 
reduced from five to three years. To this opinion I objected 
that such a change could have increased only the number of 
divorces for desertion, whereas in fact the increase of divorces 
granted on other grounds in the two years had been relatively 
much greater — 92 per cent as compared with 69 per cent. 
Hence some other reason for the increase must be sought. In 
an able and suggestive criticism of my discussion2 Dr. Dike says: 
1 Report, p. 154. 
2 Legislation and Divorce, N. Y. Evening Post, July 2, 1891. I am much in- 
debted to this letter; not so much, however, for any new facts it brought out as 
for the stimulus it gave me to reopen the question of the influence of legislation 
on divorce. No. 1.] 
A STUDY IN VITAL STATISTICS. 
89 
Mr. Wright apparently overlooked in his statements of the influence 
of legislation in Massachusetts — what he had shown in his state 
report of 1880 — the fact that four new grounds of divorce created 
in 1870 first affected the statistics in 1874 and made most of the 
increase of 92 per cent. 
Here is a new cause suggested for the increase — a change of 
law, but one of more sweeping character than that of 1873. 
Yet I must insist upon the position previously taken, that no 
causal relation can be established by observing an effect and 
merely pointing out an antecedent phenomenon that may 
account for it. Not only must the cause and effect occur 
together, but when the cause is absent the effect must like- 
wise disappear. Neglect of this principle seems to me a funda- 
mental error underlying much of Mr. Wright’s discussion of the 
influence of legislation on divorce, and the same oversight 
occurs in Dr. Dike’s reply. Dr. Dike gives a valid answer to 
my argument, but not, I think, to my conclusion. He shows that 
a sweeping change of law and an increase of divorces coincided 
in Massachusetts. But he does not consider the movement of 
the divorce rate at the same time in the adjacent states 
beyond the reach of this change of law. As a matter of fact, 
in every one of the eight other states from Maine to Pennsyl- 
vania inclusive, as well as in Massachusetts, the divorce rate 
increased from 1872 to 1874. Hence the increase in the latter 
state cannot be set down as due solely to the local change 
of law. The following table will present the facts concisely. 
Increase of Divorce in North Atlantic States, 1872-1874. 
Divorce Rate 
in 1872. 
Divorce Rate 
in 1874. 
Increase. 
Maine 
• J33 
147 
14 
New Hampshire 
. . 123 
170 
47 
Vermont .... 
. . 92 
no 
18 
Rhode Island . . 
. . 194 
199 
5 
Connecticut . . . 
. . 162 
186 
24 
Massachusetts . . 
. . 44 
76 
32 
New York . . . 
26 
3i 
5 
New Jersey . . . 
. . 18 
22 
4 
Pennsylvania . . 
• • 33 
37 
4 90 
POLITIC A L SCIENCE QUAR TERL Y. 
[ Vol. VIII. 
From this table it appears that the increase in Massachusetts, 
where there was a special local cause at work, is not conspicu- 
ously greater than the increase in three of the five adjacent 
states, New Hampshire, Vermont and Connecticut, where no 
local cause is alleged. On the basis of these figures it seems 
to me impossible to maintain that the increase in Massachusetts 
was certainly or even very probably due to the local cause, 
unless we are able to discover what general influence caused 
the increase in the other states, and to show that this was not 
present or worked less powerfully in Massachusetts. If any 
quantitative estimate of the influence of the change of law in 
Massachusetts is to be derived from the figures, it may perhaps 
be done most fairly by the following method. The other five 
states of New England have together a population about equal 
to that of Massachusetts, and exhibit a general similarity to 
that state in social and economic conditions. Comparing them 
as a unit with Massachusetts for the two years under consider- 
ation, we find that they showed an increase of 25 in their 
average divorce rate, while that of Massachusetts increased 32. 
This excess of seven may be a rough approximate indication 
of the increase in Massachusetts due to the local cause. 
May the amount of interstate migration for the purpose 
of securing divorce be approximately estimated ? One of the 
main services that statistics can render to social science is to 
make possible a quantitative measurement of facts that without 
statistics can be only qualitatively distinguished. This must 
be my excuse for suggesting for the quantitative determination 
of the current of interstate migration to secure divorce, a 
method which is admitted to be imperfect and to give only 
roughly approximate results, but which is believed to be correct 
in theory and the best method possible with our limited data. 
Only two illustrations will be given, since these will suffice to 
illustrate the method, and that is all that is here desired. 
It is often said that natives of New York, where the law 
is strict, resort to Rhode Island to obtain divorces.1 In 
1 “ Our statute ... is so much more favorable to easy divorce that instances 
are not rare of persons coming to the state and remaining the time required to No. 1.] 
A STUDY IN VITAL STATISTICS. 
91 
1870 there were 1,073,572 natives of New York State living 
in other parts of the Union, and of these 3932 were 
living in Rhode Island. That is, of all the natives of 
New York State who in 1870 were living beyond the limits 
of the state but within the limits of the country, .36 per 
cent resided in Rhode Island. The migration of the natives 
of New York to other parts of the Union may be con- 
ceived as a movement due to the attractive force exerted 
upon them by the advantages offered or believed to exist 
elsewhere; and the proportion of these persons going to 
any particular state or territory may be conceived to measure 
the attractive force of that state or territory in terms of 
the total force. The attractive force of Rhode Island for 
natives of New York in 1870 was, then, .36 per cent of the 
attractive force of the whole country outside of New York. 
In 1880 .54 per cent of the natives of New York residing 
without that state and within the country were living in 
Rhode Island, or in other words the attractive force of 
Rhode Island had risen to .54 per cent of that of the whole 
country. Now let us compare with these figures the number 
of persons going to Rhode Island from New York to obtain 
divorce. The number for each year is not given, but we are 
told1 that in the twenty years ninety-seven divorces were 
granted in Rhode Island to parties who had been married in 
New York. We know also 2 that in the whole country outside 
of New York in the twenty years 9,205 divorces were granted 
to parties who had been married in New York State. That is, 
of all the divorces of New York marriages granted outside of 
that state, 1.05 per cent were granted in Rhode Island. Now 
if what we may call the current of legitimate migration from 
New York to Rhode Island was somewhere between .36 per cent 
and .54 per cent of the current from New York to the whole 
obtain standing in our courts. —Governor s Message for 1883. Again: “I trust 
that our laws will be so amended that at least Rhode Island may no longer have 
the unenviable reputation it now has in respect to divorce and may no longer be 
a resort from neighboring states of parties who cannot be divorced by the laws of 
their own state.”—Message for 1884. 
1 Commissioner Wright’s Report, p. 194. 
2 Ibid., p. 196. 92 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
country, and if the current of divorce migration to Rhode 
Island was about 1.05 per cent, the difference between these 
two may be held to express the amount of migration guided 
by the desire to obtain divorce. In a general way this 
analysis would indicate that from thirty-five to fifty of the 
ninety-seven cases in which Rhode Island divorces were 
granted to parties married in New York were cases of bona fide 
migration, in which divorce merely followed change of resi- 
dence, without serving as a motive for it. While in any single 
instance like the foregoing such a course of reasoning may 
result in erroneous conclusions, I firmly believe that in the 
great majority of cases it would give an approximation to the 
truth and be far better than ignorance. It is an amplification 
and, as I think, an improvement of the method followed by 
Mr. Wright. 
The application of this method may be made to another 
instance in which the data have been published. New York 
citizens are said to move to Pennsylvania in order to obtain 
divorce. How largely is this true ? The attractive force of 
Pennsylvania on natives of New York, as measured by the 
number residing in the former state, was in 1870 8.2 per cent 
and in 1880 8.3 per cent of the force exerted by the whole 
country. On the average during the twenty years the attrac- 
tive force of Pennsylvania on persons married in New York 
and obtaining a divorce elsewhere was 8.6 per cent. If the 
difference roughly expressed the migration for the purpose of 
obtaining a divorce, we may say that probably at least 700 of 
the 765 persons who were married in New York and divorced 
in Pennsylvania went there with the motives of ordinary 
immigrants and not to secure divorce. 
Is it possible to trace a geographical distribution of divorce ? 
In the population bulletins of the United States census the 
country is divided into five groups. The lines of separation 
are partly geographical and partly social and economic. Two 
lines are geographical, following approximately the two great 
mountain systems and dividing the country into an eastern, a 
central and a western group of states. The third line is social No. i.] 
A STUDY IN VITAL STATISTICS. 
93 
and economic, marking the difference between the former 
slave states and the free states. Each of the five groups of 
states thus formed, the north Atlantic, the south Atlantic, the 
north central, the south central and the western, is treated 
for many purposes as a whole and as such compared with the 
others. This method may be profitably employed in studying 
the divorce statistics. Accordingly I have estimated the 
population of each group, found the number of divorces 
granted in it for each of the twenty years, and computed the 
average divorce rate. The result is most clearly set forth in 
the following diagram. 
Divorce Rates in Groups of States. 
From this diagram it will be seen that the states from 
Delaware to Florida have uniformly had the lowest divorce 
rate in the country. At the start the second group in freedom 94 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
from divorce was the southern states of the Mississippi valley. 
But the rate in that group has increased very rapidly, passing 
the north Atlantic group in 1876 and rapidly gaining on the 
north central, while for the last two years shown its rate was 
higher than that for the country as a whole.1 
The north Atlantic group of states, from Maine to Penn- 
sylvania inclusive, shows no increase of divorce rate in the 
twenty years; in other words the growth of divorce has just 
kept pace with the growth of population. If we divide this 
region into two groups, we find that in New England the 
divorce rate has slightly decreased in twenty years, while in 
New York, New Jersey and Pennsylvania as a whole it has 
slightly increased, the two offsetting each other. 
Perhaps the most striking fact illustrated by the diagram is 
the high divorce rate and its rapid increase in the western 
group of states. At the beginning of the period the two 
southern groups were most free from divorce, then came the 
two northern groups, while the western group had the highest 
rate ; but for the last eleven years the two eastern groups were 
most free from divorce, then came the two central groups 
and finally the western group. Disregarding for a moment 
the western group, we may say that for the period 1867-75 
inclusive the difference between south and north was greater 
than that between east and center ; but for the period 1876- 
86 inclusive the reverse was true, and the average differ- 
ence between east and center was greater than the average 
difference between south and north. In the single matter 
of divorce rate the differences between south and north 
have been dwindling since the war, while the differences 
between east, center and west have been very rapidly increas- 
ing. Perhaps this change may be most clearly brought before 
the eye by combining the four groups east of the Rocky 
Mountains, first into two groups of southern and northern 
states and then into two groups of eastern and western states, 
for the first and the last years of the period, marking the in- 
crease in each case and the differences between south and 
1 Compare diagram, p. 79. No. 1.] 
A STUDY IN VITAL STATISTICS. 
95 
north and between east and west at the beginning and end of 
the period. 
Comparison of Divorce Rates. 
Southern States. 
Northern States. 
Difference. 
1867 
71 
47 
1886 . . . 
... 68 
92 
24 
Increase . 
... 44 
21 
Eastern States. 
Central States. 
Difference. 
1867 . . . 
... 43 
65 
22 
1886 . . . 
... 49 
1 *3 
64 
Increase . 
. . . 6 
48 
It thus appears that the difference between south and north 
has decreased nearly one-half in twenty years, i.e. from 47 to 
24, while the difference between the eastern and the central 
states has nearly trebled, from 22 to 64. Or if we look at 
the changes in the rate rather than at the differences, the in- 
crease in the south has been more than twice as rapid as in the 
north, and that in the central states no less than eight times 
as rapid as in the eastern. 
A general conclusion to be drawn from this is that since the 
war the conditions of family life in southern and northern 
states, so far as they find expression in the divorce rate, have 
been growing more similar, while the conditions at the east and 
at the west have been growing more diverse ; in other words, 
that in this respect the ante bellum divergence between north 
and south has been changing into a divergence between east 
and west. 
Can any cause be suggested for this growing difference be- 
tween the divorce rates of east and west ? It may be worth 
while to state my own hypothesis. All over the civilized world 
we find two great currents of migration in progress, one from 
the rural districts to the cities and the other from long settled 
districts to newly opened territory. Now it is a general fact 
that the divorce rate is higher where either one of these cur- 
rents stops, than at its source. Thus in Europe the city 
divorce rate is from three to five times that of the surrounding 96 
POLITICAL SCIENCE QUARTERLY. 
[Vol. VIII. 
country, and in the United States about 95 per cent of the 
cases studied show the city rate to be likewise higher. So, 
too, in the Australian colonies, settled almost entirely from 
Great Britain and Ireland, and governed in such matters by 
English law, the divorce rate is several times as high as it is in 
the mother country. Each of these two streams of migration 
seems to involve a process of natural selection whereby the 
most energetic and self-reliant, and also the most discontented, 
dissatisfied and even criminal classes are sifted out and drawn 
off to the new homes. Among these the proportion of persons 
desiring divorce would be much greater than among those 
remaining behind. The prevalence of divorce in our far western 
states would thus be comparable to the frequent cases of lynch 
law in that region, both being natural though lamentable ex- 
pressions of the large proportion of lawless elements in those 
states. 
Walter F. Willcox. Cornell ilnioersitg. 
DEPARTMENT OP STATISTICS.