U.S. DEPARTMENT OF AGRICULTURE.

fokestry division.

Bulletin No. 7.

FOREST INFLUENCES.

1. INTRODUCTION AND SUMMARY OF CONCLUSIONS. ByB. E. Fernow.

2. REVIEW OF FOREST METEOROLOGICAL OBSERVATIONS: A STUDY

PRELIMINARY TO THE DISCUSSION OF THE RELATION OF FOR-
ESTS TO CLIMATE. By M. W. Haurington.

3. RELATION OF FORESTS TO WATER SUPPLIES. By B. E. Fernow.

4. NOTES ON THE SANITARY SIGNIFICANCE OF FORESTS. By B. E. Fer-

now.

APPENDICES.

1. DETERMINATION OF tHE TRUE AMOUNT OF PRECIPITATION AND ITS

BEARING ON THEORIES OF FOREST INFLUENCES. By Cleveland
Abbe.

2. ANALYSIS OF RAINFALL WITH RELATION TO SURFACE CONDITIONS.

By George E. Curtiss.

PUBLISHED BY AUTHORITY OF THE SECRETARY OF AGRICULTURE.

WASIimGTON:

GOVERNMENT PRINTING OFFICE.
1893.

LETTER OF SUBMITTAL.

U. S. Department of Agriculture,

Forestry Division,
Washington, D. C, November 1, 18V 3.

Sir: I have the honor to submit herewith for publication a review
of the meteorological observations which have been made, mostly in
foreign countries, for the purpose of determining whether and to what
extent forests influence climate, together with a discussion of tbe
manner in which forests affect the water conditions of the earth and
other matter elucidating the question of forest influences in general.
Kespectfully,

B. E. Fernow,
Chief of Division.
Hon. J. M. KusK,

Secretary of Agriculture.

TABLE OF CONTENTS.

PAGE.

I. Forest influences: Introduction and summaiy of conclusions, by B. E.

Fernow 9

Introduction 9

Suininary of conclusions 13

General considerations 13

Difi'erence of meteorological conditions within and without the forest. 14

Soil temperatures 14

Air temperatures under the crowns 14

Air temperatures within the crowns 15

Air temperature above the crowns 15

Air temperature in general 15

Humidity of air 16

Influence of forests upon the climate of the surrounding country 17

Influence of forests upon water and soil conditions 20

Sanitary influence of forests 21

II. Review of forest meteorological observations: A study preliminary to the

discussion of the relation of forests to climate. Bj^ M. W. Harrington. 23

The nature of the problem 23

~ Systematic observations in forest meteorology 26

France 27

Germany 30

Sweden 31

Austria 33

Objects and methods 35

Soil temperatures 40

Temperature of the air in the interior of forests 51

Annual range 51

Monthly range 53

Deciduous and evergreen forests 58

Temperature in the tree-crown and above trees 61

Temperature gradients 68

Temperature above forests 69

Tree temperatures 72

Vegetation and air temperatures 76

Theoretical calculation of heat absorption 80

Temperatures in woods, glades, and plains 83

Conclusions regarding temperatures within and without forests 88

Temperatures in wooded and treeless territories 89

Relation of evaporation to forests 96

Forests and humidity 1*^2

Rainfall in, above, and near forests ^*^

Precipitation over wooded and treeless districts HI

Forests, wind, and storms 118

Influence of forests on fogs and clouds 121

Influence of forests on hailstorms and blizzards 121

6

6 TABLE OF CONTENTS.

PAOB.

III. Relation of forests to water supplies. By B. E. Fernow 123

Tie total waf er supply 123

Influence of forests upon precipitation 125

Results of observations near Nancy, France 126

Results obtained in Bohemia 126

Results of observations in India and Brazil 128

Supposed influence of forests upon bailstonns 129

Influence of forests upon the disposition of the water supply 130

Elements of dissipation 130

Interception 130

Evaporation 132

Transpiration *. 136

Elements of conservation 137

Efl^ects of forests in case of snow 137

General effect of forests in reducing evaporation 138

Summation of the conservative and dissipative influences 140

Distribution of terrestrial waters 141

Analysis of determining influences 141

Water capacity and water conductivity of soils and soil-cover.i .. 144

Water conductivity of soils 148

Formation of springs and conditions attecting their flow 153

Influence of forests upon surfiice drainage 157

Retardation of the waterflow 158

Experience in the French Alps 159

Examples of the influence of forests on waterflow 162

IV. Notes on the sanitary significance of forests. By B. E. Fernow 171

APPENDICES.

I. Determination of the true amount of precipitation and its bearing on

theories of forest influences. By Cleveland Abbe 175

Relation of pattern and altitude of guages to accuracy of rainfall measure-
ments 1'76

Elimination of errors of the rain guage 180

Variations in geographical distribution of rainfall 182

Chronological variations of rainfall 184

Recommendations 185

A]»p]icatiou to forestry 1^5

II. Analysis of the causes of rainfall with special relation to surface condi-

tions. By George E. Curtiss 187

Climatic illustx'utions 189

LIST OF ILLUSTRATIONS.

FIC9. TACK.

1. Diflferencos of mean annual temperatures of soil (W—0) 42

2. DiBereuce of temperature (W— O) at the depth of 4 feet 44

3. Difterence of soil temperature (W— O), all stations — German observatories . 45

4. Dift'erences of soil temperature (woods and open fields). Comparison of

deciduous and evergreen trees (W— 0) 47

5. Differences of soil temperature (woods and open fields). Comparison of

elevations above sea level (W — O) 48

6. Differences of temperature for young trees, Lintzel Station, woods and

open fields (W-0) - 49

7. Effects of litter on soil temperature (littered surface— bare). Wollny's

observations ^9

8. Difference of soil temjjerature, under sod and bare surface (sod — bare).

Becquerel's observations 50

9. Forest air temperature differeuces, W— O. Geruian stations 52

10, II, 12, 13, 14, 15, 16, and 17. Forest air temperature differences, woods and

open fields (W — O), at Friedrichsrode, Hagenau, Sonnenberg, Ebers-

walde, Schoo, Marienthal, Hadersleben, and average 55, 57

18, 19, 20, 21, and 22. Forest air temperature, difference of woods and open
fields (W — O), deciduous trees, evergreen trees, elevated stations, near
sea level, and young forest (Lintzel) 59

23. Forest air temi>eratiire differences for the year at height of the tree top

(W— O) - <>2

24, 25,20, 27, 28, 29, 30, 31, and 32. Tree- top temperature, differences (W— O)

at Friedrichsrode, Kberswalde, St. Johann, Carlsberg, Sonnenberg,
Kurwien, Ilagenau, Schoo, Neumath 64, 65

33. Average differences of tree-top temperature, sixteen German stations

(W— O) 66

34. Tree-top temperature differences for four stations, with average height

of 24 feet 6<5

35. Tree-top temperature differeuces for four stations with average height of

46 feet *^7

36. Tree-top temperature diftorences, German stations (W— 0), deciduous

trees ^^

37. Tree-top temperature differences, German stations, evergreen trees 67

38. Vertical temperature, gradient in woods, degrees Fahrenheit for a hun-

dred feet Gti

39. Forest temperature differences at Ilalatte vinder deciduous forest 6J'

40. Forest temperature differences at Ermenonville under pines, soil, coarse

bare sand '^

41. Forest temperature differences above trees, from Fautrat'e observations. 70

42. Vertical temperature gradients from observations above trees 7C

43. Tree temperature, Geneva observations '^^

44. Temperature differeucea between tree and air '^

7

8 LIST OF ILLUSTRATIONS.

PAGE.

45. Differences of mean annual temperature between woods and glade, and

woods and plains, Swedish observations 85

46. Diftereuce of air temperature in forest, glade, and plain (Swedish) 86

47. Amplitude of W — G (lower curve) and W— O (upper curve) Swedish

observations 87

48. Differences of soil temperature at 20 inches depth 87

49. Evaporation and precipitation 96

50. Monthly evaporation in the fields (upper curve) and woods (lower curve),

in inches 97

51. Percentage of evaporation in the woods as compared with that in open

fields 97

52. Eatio of evaporation from water surface in fields (upper curve) and for-

est (lower curve) to precipitation (top line) 97

53. Percentage of evaporation in woods to that in the open air 98

54. Percentage of precipitation, received under trees (all kinds). German

observations 106

55. Percentage of precipitation, received under trees (different kinds) 107

.56. Px-ecipitation through trees at Cinq Franchees and Bellefontaine 108

57. Snow bank before and behind a fence 119

.58. Forest circulation 120

59. Day circulation 120

60. Night circulation 120

61. Standing or ground- water spring 154

62. "Fissure" spring 155

63. Cavern spring 155

FOREST INFLUElSrCES.

1— INTRODUCTION AND SUMMARY OF CONCLUSIONS.

By B. E. Fkrnow.

INTRODUCTION.

One of the arguments upon wliicli a change of policy in regard to
our forests, and especially on the i>art of the National Government, is
demanded, refers to the influence wliich it is claimed forest areas exert
upon climate and waterflow. It is argued that the wholesale removal
and devastation of forests affects climate and waterflow unfavorably.

Popular writers on forestry, friends of forestry reform, and the public
mind have readily taken hold of this proposition, enlarged upon it, and
generalized without sufficient and relevant premises, and before it was
possi1)le for science and systematic observations to furnish grounds or
sound deductions; hence we have had only presumptions supported by
superficial reasoning and occasional experiences. Even scientific writers
have discussed the question without proper basis, and have sought to
reason out the existence or absence of such an influence uj^on general
premises and such evidence as the history of the world seemed to fur-
nish, or else upon observations which were either of too short duration
to allow elimination of other disturbing factors or else were otherwise
unreliable.

From the complication of causes which produce climatic conditions
it has always been difficult to prove, when changes of these conditions
in a given region were observed, that they are i)ermanent and not due
merely to the general periodic variations which have been noted in all
climates of the earth, or that they are due to a change of forest condi-
tions and to no other causes; hence some climatologists have thought
])roper to deny such influences entirely. On the other hand, there are
as trustworthy and careful observers who maintain the existence of such
influences; but only of late has the question been removed from the
battle-field of opinions, scientific and unscientific, to the field of experi-
ment and scientific research, and from the field of mere speculation to
that of exact deduction. But the crop of incontrovertible facts is still

9

10 FOREST INFLUENCES.

scanty, and further cultivation will be necessary to gather a fuller
harvest and then to set clear the many comx^licated questions connected
with this inquiry.

Meanwhile a thorough beginning with a view to settle the question by
scientific methods and careful systematic measurements and observa-
tions has been made in Europe, wliere the existence of well-established
forest administrations, manned with trained observers, has rendered
practi<*able the institution of such work on an extensive scale — the only
one which can yield adequate results. Notably the observations made
at the duplicate stations in Prussia, and recorded for fourteen succes-
sive years, furnish reliable material for the discussion at least of the
relation of forest cover to meteorological j)henomena, and from these
in time the nature and extent of any influence upon the climate, if such
exists, may be determined.

Prof. M. W. Harrington, now Chief of the Weather Bureau, has, in
the following pages, compiled these and other observations, and in an
ingenious manner has compressed them iuto graphic illustrations, which
readily convey the results to the reader. While an attempt has been
made to discuss the problems and records in popular form, the student
of forest meteorological problems will find not less useful the clear and
unbiased statements of what these problems involve and what the
records do and do not show.

It may be proper to call attention to and accentuate the fact that the
question of practical importance is not so much as to the efl'ects upon the
general climate, but as to the local modification of climatic conditions
which a forest area may produce.

It can not be too strongly impressed upon those who disclaim any
influence of forest cover on climate, because the cosmic causes by
which this is produced are immeasurably greater, that there are two
classes of climate always to be considered separately, namely, the gen-
eral climate and the local climate. The latter is of most importance to
us, and alone can be modified by small causes. We modify it by building
a house around us, thus altering the temperature and moisture conditions
of the atmosphere so inclosed ; but the question is, whether we can alter
these conditions on a larger scale by such means as alternating forest
areas and fields or by large bodies of forest. We are not so much con-
cerned as to whether the total rainfall over the continent is increased,
but whether the distribution of precipitation in time and quantity over
and near a forest area is influenced by its existence; whether we or
our crops feel its absence or presence in our immediate neighborhood;
whether the protection it seems to afford and the changes it seems to
produce in the meteorological phenomena are or are not real and of
sufficient magnitude to influence our forest policy.

We would here call special attention to the memoir of Prof. Cleve-
land Abbe, contained in this bulletin, from which the difficulty of ob-
taining accurate records of rainfall with the gauges iu general use is

FOREST INFLUENCES. 11

apparent. Wlieu we find our means of measurement so deficient we
must be careful how far we base conclusions on their records.

In this connection the very suggestive paper by Mr. George E.
Curtiss, contained in this bulletin, shonld receive attention. His clas-
sification of rains into convective, orographic, and cyclonic will cer
tainly assist in developing- true conceptions as to cause and effect and
possible relation of surface conditions to rainfall.

While, then, conclusive deductions may perhaps not be as yet ad-
missible, we can not refrain from pointing out the results obtained in
the forest station at Lintzel (recorded on page 113), which seem to
show that forest planting did, under the conditions there prevailing,
produce a considerable change in meteorological conditions.

We can understand readily that if any influence exists it must be
due, in the first place, to the mechanical obstruction which the forest
cover presents to the passage of air currents and to the action of the
sun's rays upon the soil— it must result from a difterence in insolation
and consequent differences in temperature and evaporation over forest
and field. It is also readily uuderstood that the influence can become
appreciable only when large enough areas exhibiting such dift'erences
are opposed to each other, capable of producing local currents of air
which may intercommunicate the characteristics of the one area to the
other. The size and character of the forest growth, its density,
height, situation, and composition, are, therefore, nuich more impor-
tant in determining- its influeuce than has been hitherto supposed.
It is not trees, but masses of foliage, which may be effective. A large
sheet covering- an extended area from the influence of the sun would
produce almost the same differences in meteorological conditions that a
forest cover is expected to produce.

While, then, we may admit a priori that extent or area and condi-
tion of the forest cover are important, we have as yet no data from
which to calculate any proper size or proportion, and the attempts to
fix a certain percentage of forest cover needed for favorable climatic
conditions of a country are devoid of all rational basis.

Leaving the question of forest influences upon climate as still await-
ing final solution, we may speak with much more confidence of the
ettect which forest cover exerts upon the disposal of water supplies.
This effect can be much more readily studied and shows itself nuich
more conspicuously. It is perhaps also much more important to human
economy, for it is becoming more and more apparent that our agricul-
tural production is dependent not so much upon the amount of rainfall
as upon the proper disposal of the waters that fall.

Eecognizing this truth, the American Association for the Advance-
ment of Science, iu 1891, sent the following resolution to the Secretary
of Agriculture:

The American Association for the Advancement of Science respectfully submits for
the cuusideratiou of the Secretary of Agriculture that the future of successful and
more productive agriculture depends very largely upon a rational water manage-

12 FOREST INFLUENCES.

nieiit, meaning thereby not only the use of water for irrigation in the arid and sub-
arid regions, but the rational distribution and use in the humid regions of available
water supplies by means of -horizontal ditches and irrigation systems, combined
with proper mechanical jireparation of the soil, and with drainage systems, with
the object of fully utilizing the water for plant production and providing for the
safe and harmless removal of the surplus.

The i>resent policy of forest production and of allowing our waters to run to
waste not only entails the loss of their beneficial influence upon plant production,
but permits them to injure crops, to wash the fertile mold from the soil, and even to
erase and carry away the soil itself.

It is upon these considerations that the association respectfully suggests to the
honorable Secretary the desirability of utilizing the Weather Bureau, the various
agricultural experiment stations, and other forces, in forming a systematic service
of water statistics, aud in making a careful survey of the condition of water supplies,
which may serve as a basis for the application of rational principles of water man-
agement.

How poorly we understand the use of these supplies is evidenced
yearly by destructive freshets and floods, with the accompanying wash-
ing- of soil, followed by droughts, low water, and deterioration of agri-
cultural lands.

It may be thought heterodox, but it is nevertheless true that the
manner in which most of the water of the atmosphere becomes avail-
able for human use (namely, in the form of rain) is by no means the
most satisfactory, not only on account of its irregularity in time aud
quantity, but also on account of its detrimental mechanical action in
falling; for in its fall it compacts the ground, impeding percolation.
A large amount of what would be carried off by underground drainage
is thus changed into surface-drainage waters. At the same time by
this compacting of the soil capillary action is increased and evaporation
thereby accelerated. These surface waters also loosen rocks and soil,
carrying these in their descent into the river courses and valleys, thus
increasing dangers of high floods and destroying favorable cultural
conditions.

Here it is that water management and, in connection with it or as a
part of it, forest management should be studied; for without forest
management no rational water management is -possible. The forest floor
reduces or prevents the injurious mechanical action of the rain and
acts as a regulator of water flow. Hitherto water management in rainy
districts has mainly concerned itself with getting rid of the water as
fast as possible, instead of making it do service during its temporary
availability by means of proper soil management, horizontal ditches
and reservoirs — drainage and irrigation systems combined. It seems
to have been entirely overlooked that irrigation, which has been con-
sidered only fen- arid and subarid regions, is to be applied for plant pro-
duction in well- watered regions with equal benefit and profit, if combined
with proper drainage systems and forest management. A discussion
of the manner in which the forest influences the disposal of water sup-
plies has been made a special part of this publication.

SUMMARY OF CONCLUSIONS. 13

It will be observed in this bulletin tliat the historical method of dis-
cussing the subject of forest influences, which consists in adducing gen-
eral observations throughout the world, has given place to the scientific
method, which relies upon si»ecil3C observations and experiments and
upon the ai)plication of well-established physical laws to the espUma-
tion of the facts observed.

SUMMARY OF CONCLUSIONS.

For those who wish to know only what the present state of the
question of forest influences is, we have summarized what conclusions
may be drawn from the facts presented in this bulletin, referring them
to the body of the report for the basis of these conclusions and the dis-
cussion in extenso. This arrangement, we are aware, is not customary
and logical, but since the object of this bulletin, to some extent, is to
tiimiliarize not only the student but the general public with the subject,
it seemed expedient to meet the convenience of the general reader in
this manner. For easy reference the pages of this bulletin containing
the data upon which each conclusion is reached are given in paren-
theses at the end of each paragraph.

GENERAL, CONSIDERATIONS.

(1) We must keep separate two main questions, namely. What is the
difference of conditions within and without the forest? and How far is
the difference of conditions within the forest communicated to the out-
side, i. e., how far does the forest influence the conditions outside?
(Pp. 23-40.)

(2) The general climatic conditions in which the forest is situated as
well as its situation with reference to elevation and exposure, further-
more its composition, whether evergreen or deciduous, its density, its
height and extent, the character of the forest floor, and other features
which determine its quality, must all combine in producing variety, at
least quantitatively, both as to difference of conditions within and
w ithout the forest and as to possible exchange of the same, and lience
the question of forest influences can be properly discussed only with
reference to these other conditions. We must refrain from generaliz-
ing too readily from one set of conditions to another set of conditions.
(Pp. 40-121.)

(3) In the matter of forest influence upon waterflow, besides the
above mentioned, other conditions, the topography and geology or
stratification of soil, must also be taken into account and generaliza-
tions without regard to these must be avoided. (Pp. 123-157.)

(4) No influence upon the general climate which depends upon cos-
mic causes can in reason be expected from a forest cover. Only local
modifications of climatic conditions may be anticipated; but these
modifications if they exist are of gre t practical value, for upon them

14 FOREST INFLUENCES.

rest success or failure in agricultural pursuits and comfort or discom-
fort of life within the given cosmic climate. The same condition must
be insisted upon with reference to forest influences upon waterflow,
which can exist only as local modifications of water conditions, which
are due in the first place to climatic, geologic, and topographic coudi-
tions. (Pp. 157-170.)

DIFFEREN(!E OF METEOROLOGICAL CONDITIONS WITHIN AND WITHOUT

THE FOREST.

(1) Soil temperatures. — The general influence of the forest on soil
temjjeratures is a cooling one, due to the shade and to the longer re-
tention of moisture in the forest floor as well as in the air, which must
be evaporated before the ground can be warmed. As a consequence
the extremes of high and low temperature within the forest-soil occur
much later than in the open, and both extremes are reduced, but the
extreme summer temperatures much more than the winter tempera-
tures.

The difference between evergreen and deciduous forests, which al-
most vanishes in the winter time, is in favor of the deciduous as a
cooling element in summer and autumn, while during sj^ring the soil is
cooler under evergreens. The effect increases naturally with the age
and height of the trees.

(2) Air temperatures under the crowns. — The annual range of air tem-
perature is smaller in the forest than in the open ; the eff'ect upon the
minimum temperature {i. e., the effect in winter) is less than on the
maximum temperature (r. e., the effect in summer). The combined ef-
fect is a cooling one. The range of temperature is more affected than
the average absolute temperature, or, in other words, the moderating
influence is greater than the cooling effect.

The monthly minima for middle latitudes are uniformly reduced dur-
ing the year, and the monthly maxima are much more reduced during
the summer than during the winter. On the average the forest is
cooler than the open country in summer, but about the same in winter,
with a slight warming eftect in sj^ring.

The difterence between the mean monthly air temperatures in the
woods and in the open varies with the kind of forest much more than
is the case for soil temperatures. The evergreen forest shows a sym-
metrical increase and decrease througiiout the year. The deciduous
forest shows a variable influence which diminishes from the midwinter
to springtime, ])ut increases rapidly as the leaves appear and giow,
becoming a maximum in June and July and then diminishing rapidly
until November. The annual average eflcct is about the same both
for evergreens and deciduous forests.

Forests situated at a considerable elevation above the sea have sen-
sibly the same influence on the reduction of the mean temperature as
do forests that are at a low level.

AIR TEMPERATURES IN AND OUTSIDE OF FORESTS. 15

Young forests affect the air temperature very differently from mature
forests; in the former tbe minimum temperatures are always reduced,
but the maxima are exaggerated. The observations on which this
conclusion is based ought, perhaps, to be considered as pertaining rather
to the case of temperatures in the tree-tops.

(3) Air temperatures icithin the croions. — The mean temperature of
the air in the tree- tops, after correcting for elevation above ground, is
rather higher than over open fields. The effect of tree-tops does not
appreciably depend upon the height of the station above ground. Tbe
effect upon the minima is generally greater than on the maxima, the
total effect being a warming one. A tree-top station is in general inter-
mediate, as to temperature, between a station near the ground in the
forest and one in the open field.

Evergreen forests show less difference between the temperature in
the crown and below, and altogether more uniformity in temperature
changes throughout the year, than deciduous growth.

The vertical gradient for temperature within the forest on the aver-
age of all stations and all kinds of forest trees is large, varying from
0.61° F. per 100 feet in April to 2.50° F. in July.

A reversal of the vertical gradient, namely a higher temperature
above than below, occurs in the woods, especially in the summer time.
It also occurs in the open air regularly at night, and may be three or
four times as large as that just mentioned. In general the action of
the forest tends to produce a vertical distribution of temperature like
that over snow or level fields on clear nights.

(4) Air temperature above the crowns. — The tem])erature, at consider-
able heights above the forest, appears to be slightly affected by the
forest and more so with evergreens than with deciduous growth. The
vertical gradients of temijerature within 30 feet above the tops of the
jrees are all reversed throughout the leafy season; the gradients are
also greater above the tree crown than below, at least during the clear
sky and calm air. The wind affects the temperature under and witliiii
the crowns, but makes little difference above them. The surface of the
forest crown appears meteorologically much like the surface of the
meadow or cornfield. It is as if the soil surface has been raised to the
height of the trees.

(5) Air temperature in general. — From the x^recediiig generalizations
it appears that the forest affects the temperature just as any collection
of inorganic obstacles to sunshine and wind; but as an organic being
the forest may be also an independent source of heat. Careful obser-
vations of the temperature within the trunk of the tree and of the leaves
of the tree show that the tree tem])erature is affected somewhat by the
fact that the rising sap brings up the temi)erature of the roots while
the return sap from the leaves brings their temperature down, and the
tree temperature considered as the resiilt of the complex adjustment is
not appreciably affected by any heat that may be evolved by the chemi-

16 FOREST INFLUENCES.

cal processes on wliicli its growth depends. It is not yet clear as to
A\^hetlier the chemical changes that take place at the surface of the leaves
should give out any heat ; it is more likely that heat is absorbed, namely
rendered latent, especially in the formation of the seed; the process of
germination usually evolves this latent heat; the immense quantity of
water transpired and evaporated by the forests tends to keep the leaves
at the same temperature as that of the surface of water or moist soil.

(6) Humidity of air. — The annual evaporation within the forests is
about one-half of that in the open field; not only is the evaporation
within a forest greatest in May and June, but the diiference between
this and the evaporation in the open field is also then a maximum,
which is the saving due to the presence of the woods. The average
annual evaporation within the Avoods is about 44 per cent of that in the
field. Fully half of the field evai^oration is. saved by the presence of
the forest.

The quantity of moisture thrown into the air by transpiration from
the leaves in the forest is sometimes three times that from a horizontal
water surface of the same extent, and at other times it is less than that
of the water. The transpiration from leaves in full sunshine is deci-
dedly greater than from leaves in the diffused daylight or darkness.
The absolute amount of annual transpiration as observed in forests
of mature oaks and beeches in central Europe is about one- quarter of
the total annual precipitation.

The percentage of rainfall, evaporated at the surface of the ground,
is about 40 per cent for the whole year in the open field and about 12
per cent for the forest, and is greater under deciduous than under
evergreen forests.

The evaporation from a saturated bare soil in the forest is about the
same as that irom a water surface in the forest, other conditions being
the same.

The presence of forest litter like that lying naturally in undisturbed
forests hinders the evaporation from the soil to a remarkable extent,
since it saves seven- eights of what would otherwise be lost.

The total quantity of moisture returned into the atmosphere from a
forest by transpiration and evaporation from the trees and the soil is
about 75 per cent of the precipitation. For other forms of vegetation it
is about the same or sometimes larger, varying between 70 per cent
and 90 per cent; in this respect the forest is surpassed by the cereals
and grasses, while, on the other hand, the evaporation from a bare
soil is scarcely 30 per cent of the precipitation.

The absolute humidity within a forest exceeds that of the glades
and the plains by a small quantity. The relative humidity in the
forest is also larger than in the glades or plains by 2 per cent to 4 per
cent. Forests of evergreens have from two to four times the influence
in increasing relative humidity than, do forests of deciduous trees.

The gauges in European forest stations catch from 75 to 85 per cent

INFLUENCE UPON CLIMATE OF SURROUNDING COUNTRY. 17

»

wlien placed under the trees, the bahince rep resen tinker that which
passes through the foliage and drii»s to the ground or runs down along
the trunks of trees, or else is intercepted and evaporated. The x)er-
centage withheld by the trees and which either evaporates from their
surfiice or trickles along the trunk to the ground is somewhat greater
in the leafy season, though the difference is not great. Deciduous and
evergreen trees show hut slight differences in this respect. More rain
is usually caught by gauges at a given height above the forest crown
than at the same height in open lields, but it still remains doubtful
whether the rainfall itself is really larger over the forests, since the
recorded catch of the rain-gauge still requires a correction for the in-
fluence of the force of the wind at the gauge.

In such cases where over a large area deforestation and reforestation
have seemingly gone hand in hand with decrease and increase of rain-
fall, the possible secular change in rainfall must also be considered.
Yet the experience of increased rainfall over the station at Lintzel
(p. 113), with increase of forest area, points strongly toward a i)ossible
interdependence under given conditions.

By condensing dew, hoar frost, and ice on their branches, trees add
thereby a little to the precipitation which reaches the ground, and by
preventing the rai>id melting of snow more water remains available
under forest cover.

The question as to the march of destructive hailstorms with reference
to forest areas, which seems settled for some regions in France, re-
mains in doubt for other, especially mountain, regions.

From these statements we would expect as a con secpicnce of defor-
estation an effect on the climate of the deforested area in three direc-
tions, namely : («.) extremes of temi)erature of air as well as soil are
aggravated; (b) the average humidity of the air is lessened, and possi-
bly (c) the distribution of precipitation throughout the year, if not its
quantity, is changed.

INFLUENCE OF FORESTS UPON THE CLIMATE OF THE SURROUND-
ING COUNTRY.

(1) An influence of the forest upon the climate of its surroundings
can only take place by means of diffusion of the vapor which is tran-
spired and evaporated by tlic crowns and by means of air currents pass-
ing through and above the Ibrests being modified in temperature and
moisture conditions; the mechanical effect upon such air currents by
which they are retarded in their progress may also be effective in
changing their climatic value.

(2) Local air currents are set up by the diflxjrence in temperature of
the air within and without the forest, analogously to tliose of a lake or
pond, cooler currents coming from the forest during the day in the
lower strata and warmer currents during the night in tlic upper strata.

12444— Xo. 7 2

18 FOREST INFLUENCES.

The latter currents, being warmer and raoister, can be of influence on
the temperature and moisture conditions of a neighboring flekl by mod-
erating temperature extremes and increasing the humidity of the air.
This local circulation is the one most important difference between
forest and other vegetation. How far away from the forest this circu-
lation becomes sensible is not ascertained. In winter time, when the
temperature differences become small, no such circulation is noticeable.

(3) The general air currents in their lower portions are cut off en-
tirely by the forest, which acts as a windbreak. This influence can
of course be experienced only on the leeward side. How far this pro-
tection reaches it is difticult to estimate, but it certainly reaches far-
ther than that of a mere windbreak, since by the friction of the air
moving over the crowns a retardation must be experienced that would
be noticeable for a considerable distance beyond the mere "windbreak
effect. Deforestation on a large scale would permit uninterrupted
sweep of the winds, a change more detrimental where the configura-
tion of the ground does not fulfil a similar function — in large plains
more than in hilly and mountainous regions, and at the seashore more
than in the interior.

The upper air strata can' be modified only by the conditions existing
near and above the crowns. At the same time they jnust carry away
the cooler and moister air there and create an upward movement of
the forest air, and thereby in part the conditions of this become also
active in modifying air currents. The greater humidity immediately
above the crowns is imparted to the air cnrrents, if warm and dry, and
becomes visible at night in the form of mists resting above and near
forest areas. These strata protect the open at least against insolation
and loss of water by evaporation, and have also a greater tendency to
coudensation as dew or light rain, if conditions for such condensation
exist. This influence can be felt only to the leeward in summer time
and with dry Avarm winds, whde the cooling winter effect upon compar-
atively warmer moist winds is not noticeable. Theoretical considera-
tions lead to the conclusion that in monntain regions only the forest
on the leeward slope can possil^ly add moisture to a Avind coming over
the mountain, but this does not necessarily increase the i)recipitation
on the field beyond. Altogether the theoretical considerations are as
yet neither proved nor disproved by actual observations, and as to
rainfall the question of influence on the neighborhood is still less set-
tled than that of precipitation upon forest areas themselves. Wherever
moisture-laden winds ])ass over extensive forest areas the cooler and
moister condition of the atmosphere may at least not reduce the i)ossi-
bility of condensation, which a heated plain Avould do; but observa-
tions so far give no conclusive evidence that neighboring fields receive
more rain than they otherwise would.

(4) With regard to comparative temperatures in forest stations and
open stations that are situated not far apart from each other, it would

TEMPERATURES IN AND OUTSIDE OF FORESTS, 19

appear that the forest exerts a cooling iDfliieiice, but that more de-
tailed conclusions are hindered by the consideration that the ordinary
meteorological station itself is somewhat affected by neighboring trees.

The study of the stations in Asiatic and European Eussia seems to
show that in the western part of the Old World the presence of large
forests has a very sensible influence on the temperature. Similar
studies for stations in the United States seem to show that our thin for-
ests have a slight effect in December but a more decided one in June.
It appears also that our wooded regions are warmer than the open
plains, but there is no positive evidence that this difference of temper-
ature is dependent upon the quantity or distribution of forests, or that
changes in temperature have occurred from this cause.

(5) When a forest incloses a small area of land, forming a glade, its
inclosed position brings about special phenomena of reflection of heat,
local winds, and a large amount of shade. For such forests it is found
that the mean range of temperature is larger in the woods than on the
open plain; the glade climate is more rigorous than the climate of open
plains; the glade is cooler and its diurnal range larger during the
spring, summer, and autumn.

Favorable influences upon moisture conditions of the air are most
noticeable in localities where much water is stored in underground with
overlying strata which are apt to dry when our summer drought pre-
vails. Here the forest growth is able to draw water from greater depths
and by transpiration return it to the atmosphere, thereby reducing
the dryness and possibly inducing precipitation. In most climates this
action Avould be less eflective or of no use. Hence in regions with
oceanic climate Avitli moist sea winds like England and the west coasts
of Europe or of the northern United States, deforestation from a cli-
nuitic point of view m;iy make no ap[>reciable difference, such as it
would make in continental clinmtes like the interior of our country, the
llocky Mountains, and southern California.

Whether large or small areas of forest and open fields alternating,
or what percentage of forest is most favorable can not as yet be dis-
cussed, since we are not clearly informed even as to the manner and the
amount of influence which forest cover exercises. In general we may
expect that an alteriiation of large forested and unforested areas, in
regions which on account of their geographic situation have a dry an«l
rigorous climate, is more beneficial than large uninterrupted forest
areas, which would fail to set up that local circulation whicli is brought
about by differences in temperature and permits an exchange of the
forest climate to the neighboring field.

20 FOREST INFLUENCES.

INFLUENCE OF FORESTS UPON WATER AND SOIL CONDITIONS.

(1) In consequence of deforestation evaporation from the soil is aug-
mented and accelerated, resulting in unfavorable conditions of soil
liumidity and affecting unfavorably the size and continuity of springs.
The influence of forest cover upon the flow of springs is due to this
reduced evaporation as well as to the fact that by the protecting forest
cover the soil is kept granular and allows more water to penetrate and
percolate than would otherwise. In this connection, however, it is the
condition of the forest floor that is of greatest importance. Where the
litter and humus mold is burned up, as in many if not most of our
mountain forests, this favorable influence is largely destroyed although
the trees are still standing.

(2) Snow is held longer in the forest and its melting is retarded, giv-
ing longer time for filtration into the ground, which also being frozen
to lesser depth is more apt to be open for subterranean drainage. Al-
together forest conditions favor in general larger subterranean and less
surface drainage, yet the moss or litter of the forest floor retains a
large part of the precipitation and prevents its filtration to the soil, and
thus may diminish the supi)ly to springs. This is especially possible
with small precipitations. Of copious rains and large amounts of snow
water, quantities, greater or less, penetrate the soil, and according
to its nature into lower strata and to springs. This drainage is facili-
tated not only by the numerous channels furnished by dead and living-
roots, but also by the influence of the forest cover in preserving the
loose and porous structure of the soil.

Although the quantity of water offered for drainage on naked soil is
larger, and although a large quantity is utilized by the trees in the
process of growth, yet the influence of the soil cover in retarding evap-
oration is liable to offset this loss as the soil cover is not itself dried out.

The forest, then, even if under unfavorable topographical and soil
conditions (steep slopes and impermeable soils) it may not permit
larger quantities of Avater to drain off underground and in springs, can
yet influence their constancy and equable flow by preventing loss from
evaporation.

(3) The surface drainage is retarded by the uneven forest floor more
than by any other kind of soil cover. Small precipitations are apt to
be prevented from running off superficially through absori^tion by thr
forest floor. In case of heavy rainfalls this mechanical retardation in
(connection with greater subterranean drainage may reduce the danger
from freshets by preventing the rapid collection into runs. Yet in
regions with steep declivities and impermeable soil such rains may
be shed superficially and i^roduce freshets in sjiite of the forest floor,
and an effect upon water conditions can exist only from the following
consideration.

SANITARY INFLUENCE OF FORESTS. 21

(4) The well-kept forest floor, better tbaii even tlie close sod of a
meadow, prevents erosion and abrasion of the soil and the washing of
soil and detritus into brooks and rivers.

This erosion is especially detrimental to agricultnral interests as
well as water flow in regions with this surface and impenetrable sub-
soils, and where rains are apt to be explosive in their occnrrence, as in
our western and southern country. The best soil of tlie farms is often
washed into the rivers, and the water stages of the latter by the accu-
mulations of this soil are influenced unfavorably.

(5) Water stages in rivers and streams which move outside the
uionntain valleys are dependent upon such a complication of climatic,
topographic, geological, and geographical conditions at the headwaters
of their afHuents that they withdraw themselves from a direct correla-
tion to surface conditions alone. Yet it stands to reason that the con-
ditions at the headwaters of each afHuent must ultimately be reflected
in the flow of the main river. The temporary retention of large amounts
of water and eventual change into subterranean drainage which the
well-kept forest floor produces, the consequent lengthening in the time
of flow, and esijecially the prevention of accumulation and carrying of
soil and detritus which are deposited in the river and change its bed,
would at least tend to alleviate the dangers from abnormal floods and
reduce the number and height of regular floods.

SANITARY INFLUENCE OF FORESTS.

(1) The claimed influence of greater purity of the air due to greater
oxygen and ozone production does not seem to be significant.

(2) The protection against sun and wind and consequent absence of
extreme conditions may be considered favorable.

(3) The soil conditions of the forest are unfavorable to the produc-
tion and existence of pathogenic microbes, especially those of the
cholera and yellow fever, and the comparative absence of wind and
dust, in which such microbes are carried into the air, maybe considered
as the principal claim for the hygienic significance of the forest.

We may summarize that the position (»f the forest as a climatic factor
is still unceitain, at least as to its practical and quantitative imi)or-
tance, but that its relation to water and soil conditions is well estab-
lished. As a climatic fact«u' it would appear that the forest of the
plain is of more importance than that of the mountains, where the more
potent influence of elevation obscures and reduces in significance the
influence of their cover; as a regulator of water conditions the fcnest
of the mountains is the inqK)rtant factor; and since this inlluence
makes itself felt far distant from the location of the forest, the claim for
attention of Government activity and for statesmanlike jiolicy m ith
reference to this factor (»f national welfare may be considered as well
founded. Every civilized government must in time own or control the

22 FOREST INFLUENCES.

forest cover of tlie mountains in order to secure desirable water condi-
tions.

In conclusion I may urge that systematic observations bearing' on
tlie subject of foreign influences should be instituted in tliis country
by a Government agency, perhaps under the authority of the Weather
Bureau and with the cooperation of the agricultural experiment sta-
tions. No other country is so well adapted for the study of this ques-
tion as the United States, ottering all the varying climatic conditions
of a whole continent under one government, with changes in forest con-
ditions constantly progressing.

IT -REVIEW OF FOUEST METEOROLOGICAL OBSERVATIONS: A STUDY
PRELLMINARY TO THE DISCUSSION OF THE RELATION OF FORESTS
TO CLIMATE.

By M. W. Harrington.

THE NATURE OF THE PROBLEM.

Tlie covering of the earth's surface deteriuiues many of the minor
features of climate, and especially those features which are of import-
ance for agriculture. The insular climate, the coast climate, the desert
and prairie climates, the differences between steady winters with con-
tinuous snow-covering and variable winters with little or no snow, all
have peculiarities that are due to the character of the covering of
the earth's surface at or near the station which has the climate in ques-
tion.

A striking illustration of the effect of the surface-covering on climate
and weather is seen in the case of a snow- covering. This has been
studied recently by an eminent Russian meteorologist, Dr. VVoeikofl',
who finds the influence of the snow to be surprisingly far-reaching. It
chills and equalizes the temperature; it promotes the passage of bliz-
zards and otlier winter storms; it retards the spring; it feeds the water
more slowly to the streams; it promotes the continuance of clear, dry
weather with high barometric pressure; and it has mauy other effects
which are of hardly less importance.

The forest is to be considered, in its effects on climate and weather,
as a special form of surftice-covering. Its effects are of the same order
as those produced by a covering of sand, or sod, or water, but the
forest effect has some peculiar features which are due to the fact that
the covering is elevated to some extent above the soil. This imparts
to the soil in some degree the climatic characteristics due to a topo-
graphical elevation, and also causes a series of wind-break effects which
are not found with the other forms of surface-covering. On account of
this distinctive feature, the problem of forest clinuitology separates
into two problems, which must be considered each by itself. The one
relates to the climate of the interior of the forest, the other to the
effects of the forest on the (ilimate of the open country around it. The
two are quite different; the first is of relatively little imiiortance ex-
cept as it relates to the second. It is the second which is of interest

23

24 FOREST INFLUENCES.

and imi^ortance so far as relates to the suitability of a climate for resi-
dence and agriculture.

The solution of these two problems is very much complicated by the
variety of conditions of the forest itself. The forest, woods, or wood-
lands may vary in extent all the way from the Amazon forest, which
covers a large part of South America, or the corresponding forests of
tropical Africa, to the woodlot of the farm, the grove, or the i)ark. In
density of tree-growth they vary quite as widely. The openings in
forests are esjiecially subject to forest climatic action, as any one knows
who has found himself in a tropical forest glade at the hottest hour of
the day; and these openings vary in form and in size all the way from
the forest gallery or swale through the oak oi)enings and little prairies
to the regions where the prairies x^revail and the forest is reduced to
islands of woods or to fringes along the streams. In height the trees
range from upwards of 300 feet to the low coppice or brush of a few
feet elevation; some forests are of fairly uniform height, like the north-
ern forests of firs and spruces, while others are composed of plants of
all heights from the herb to the forest giant. This is mostly the case
with the tropical forests where the giants are scattered between high
and low trees and open glades, so arranged that the sunbeams, not-
withstanding the density of the vegetation, sift through to the soil
itself. In undergrowth, in i^ersistence of leaves, and in quantity of
shade, there is also every possible variation.

The eifect of a mixed forest of deciduous trees must differ from that
of an evergreen forest with needle-shaped leaves, and this again from
forests of such trees as flourish in Australia, New Holland, and else-
where, whose leaves stand more vertically so that the shade bears
little relation to the leaflness. Such trees are found in numy ])arts
of the world; our own Kentucky coffee tree, is of this sort, and the
lack of shade under a tree of this kind in full leaf in July is always a
surprise.

Again, the litter accumulating in the forest has an important influ-
ence on its action, and the effects of the forest must be also much modi-
tied by many factors which determine its condition but are not strictly
a part of it, such as the soil on Avhich it grows, the abundance or
scarcity of water, the slope of the land, the altitude above .sea level,
the latitude, and the prevailing degree of cloudiness. Any of these
peculiarities may sensibly modify not only the clinuite of the interior
of the forest but also the influence which the forest may exert on the
climate of adjacent territory.

The literature (»f forest meteorology is already a large one. Liiffel-
holz-Colberg published in 1872 a catalogue of the publications on the
significance and importance of the forest, and since the issue of that
catalogue the literature of the subject has grown at a rapid rate.
Nearly all the publications to which he refers bear on the meteorological
aspects of the forest. The longer bibliographical references given by

METHODS OF INVESTIGATION. 25

liim number 190, a,ud there are briefer references to very many more.
Loflfelliolz-Colberg's list begins with Fernando Columbus, the son of
Christopher Columbus, who attributes the heavy rainfall on Jamaica
to its wealth of forests, and the decrease of rain on the Azores and
Canaries to the removal of their forests. In the sixteenth and seven-
teenth centuries the subject was already attracting the attention of
the French Government, and in fact governmental interest in the sub-
ject goes back to the time of the immediate successors of Charlemagne.
It is interesting to read over the abstracts of opinions which are recorded
in Lott'elholz-Colberg's smallbook. Every variety of opinion can be found
there, from those which attribute to the forest about everything which is
desirable in climate and even endow it with a powerful influence on mor-
als, to those who believe it is entirely without influence ; and from those
who think that its influence does not extend bej^ond its own margin,
to those who would attribute the deterioration of the climate of the
Old World to the removal of the forests of the New. The reasons for
this extraordinary variation in opinion are to be found in the method
employed for solving the problems of forest meteorology. Leaving out
of account the solutions which were purely sentimental or purely tour-
istic, the conclusions were generally founded on what may be called
the historical method. This consists in finding a country which has
been once wooded, but from which the forests have been removed,
or one which was once open, but later became wooded. The clinuite at
the beginning and end of the time involved is then ascertained or
assumed, and the changes in the climate are attributed to the change
in the forest cover. The uncertainties of this method are so great as
to make it generally useless. It is seldom possible to be sure of the
early forest condition of the country in question. For this purpose
reliance must generally be placed on incidental references by the earl-
ier writers, and these are usually ambiguous and uncertain. Even
where the change in forest conditions can be proved bejond the possi-
bility of doubt, the character of the early climate can not be ascer-
tained with a sufficient degree of accuracy. If the period in question
preceded the introduction of meteorological instruments, then the char-
acter of the cliuuite must be judged by the nature of the crops raised or
from facts of a similar nature given by earlier writers. If observa-
tions were taken they were generally rude and by means of imperfect
instruments: their errors would probably approach in magnitude the
quantity of forest influence which is to be determined. Besides, the
fixing of the data of climate with our modern exact instruments and
better methods requires observations for a long series of years. The
variations of the elements of climate (the temperature, humidity, rain-
fall, winds, and cloudiness) are great from day to day and from year
to year, and it takes numerous observations, scattered through many
years and taken on a uniform system, to give fixed values for the cli-
mate. The rainfall tor two successive years on the same spot might

26 FOREST INFLUENCES.

differ by several times as luucli as the differeuce wliicTi could be reason-
ably expected between that on prairie and on forest. ■ In addition to
this, there is no certainty, generally, that any change of climate which
is satisfactorily proven is necessarily due to a change in the forest.
Changes are constantly going on in the climate of all stations and en-
tirely without obvious connection with surrounding forest conditions.
In fact, their causes are beyond our present knowledge.

These are the secular changes, so called; they are probably periodic,
and their periods may be tens, scores, or hundreds of years. It is hardly
possible to find a long series of observations in which traces of these
secular changes are not evident.

The historical metliod has really given but meagre results of a reliable
character. On the other hand, the extravagant results to which it has
sometimes led have cast discredit on the whole subject of the influence
of forests on climate. It will not be used in what follows except in a
single case, in which all the conditions which can reasonably be required
seem to be fairly fulfilled.

There are also recognized results of forest action which are entirely
distinct from the meteorological influences with which we are here con-
cerned and which refer to soil and water conditions. Such is the in-
fluence of forests on the flow of surface water and the occurrence of
floods, on the amount of ground- water and the preservation of springs,
on the holding of movable soils or reclaiming of swampy ones. These
are all of the highest importance, but they will not be considered here.
The questions of temperature and of the changes of the air are the only
ones belonging to forest meteorology proper.

SYSTEMATIC OBSERVATTOISrS IIS^ FOREST METEOP.OEOGY.

There are several series of sporadic observations on forest meteor-
ology made before the beginning of this century, as, for instance.
Hunter's observations on the temperature of trees in 1775 and 1778,
and Schaeplf's observations on the same subject in New York in 1783.
The first systematic observations seem to be those taken at Geneva from
1 790 to 1800 by Pictet and Maurice. They referred to tree temperatures
and were made ou a horse-chestnut tree two feet in diameter.

The thermometer was placed in a hole in the north side of the trunk
six inches deep, and the space around the stem was closed up with wax,
so that the water and air from the outside would not influence the tem-
perature shown by the instrument. The observations were made with-
out interruption at sunrise, 2 o'clock p. m., and sunset. Comparative
observations were made at the same time on a thermometer to the
north, in the shade, and on four ground thermometers at depths of 3
inches, 1 foot 5 inches, 4 feet 3 inches, and 12 feet 8 inches. The ob-
servations were made with precautions similar to those now considered
necessary, but not generally practiced at that time.

FRENCH OBSERVATIONS. 27

Kaeintz carried on a good many scattered but instructive forest and
agricultural observations before the publication of bis meteorology in
1831-'34. In 1839-'40, Bravais at Bossekop, and Thomas at Kaaford,
botli in Lapland, took a long series of winter tree temperatures. Bra-
vais put the bulb of a mercurial thermometer in the heart of a liviug
pine wbicli was G inches in diameter. Thomas compared the temi)era-
tures of two large pines close together, one dead and the other living.
The latter showed a slightly higher temperature than the former.

In the winter of lS57-'r)8, Bourgeau carried on a series of tree tem-
Xjerature observations at Fort Carlton, on the Saskatchewan Eiver,
latitude about 52°.

FRANCE.

The beginning, however, of systematic observations on the short-
range forest meteorological problem of the relative condition of the air
in and about forests and the incitement to the modern activity in this
direction seem to have come from the observations of M. A. C. Bec-
querel and his son Edmund, wliich began at Chatdlou-sur-Loing, about
70 miles south of Paris, on July 30, ISoS. These observations, now re-
lating to forest meteorology, now to some other branch of the science,
now carried on at Chatillon or its vicinity, now at Paris in the Jardin
des Plantes, and generally with the aid of the French Academy of Sci-
ences, have been conducted by MM. Becquerel and his son from that
time to a quite recent date, and are probably still continued.

The memoirs and briefer communications which have been made to
the academy, based on these observations, form a large and highly in-
structive addition to knowledge in this science, and they have been
abundantly drawn on by subsequent writers. It is in some respects an
advantage, in others a disadvantage, that the Messrs. Becquerel have
used a special form of instrument in their observations and have fol-
lowed independent methods. The disadvantage lies in the fact that
this makes the observations not strictly comparable with those taken
at regular forest meteorological stations, and when the question is one
of relatively small dififerences, as that of temperatures and rainfall
within and without forests, this lackof comparability is a serious draw-
back. On tiie other hand, there is a great advantage in getting results
by difterent methods and with different instruments, for it serves to get
rid of errors that dejiend on the method. Conclusions drawn from two
entirely different sorts of observations are worthy of more confidence
than those drawn from two sets taken in the same way.

The observations of the Messrs. Becquerel were taken with electric
thermometers. They depend on thermo-electric principles, and the
elder Becquerel had published studies on them as early as 182G, and
afterwards used them in determining the temperature of the different
envelopes of tlames. The instrument as used was composed of a closed
circuit a part of which was of copper and a part of iron. The two are

28 FOREST INFLUENCES.

soldered at tlie ends, and a galvanometer is introdncod into the circuit.
By the i)rinciples of thenno-electricity whenever the two soldered ends
are at the same temperature, there is no current and the galvanometer
is not affected. If one of the junctions is warmer than the other, theu
a current is produceol in the one or other direction, and the needle of
the galvanometer swings to the corresponding side. If now one junc-
tion is i^laced in the medium the heat of which is to be measured, and
the other is left accessible and is warmed or cooled until the needle
stands at zero, the temperature of the latter junction is the same as that
of the former. One junction can be placed in a flame, or in an aninml
or ])lant, or at the top of a tree, or at a depth in the ground, and its
temperature can be ascertained by mani])ulating the other.

This method requires many precautions, and can not be conveniently
used for maximum and minimum temperatures, but incompetenthands
it is very serviceable and is capable of minute accuracy.

In the observations of the Messrs. Becquerel it was so arranged that
a change of one degree of temperature would cause a deflection of ten
degrees in the needle. The needle was brought to zero with the aid of
a, lens, and an accuracy of reading to a tenth of a degree centigrade in
the galvanometer was abundantly assured. Every precaution was
taken. For instance, a junction exposed to the rays of the sun was
covered with a triple reflector which screened from the sun's rays, but
permitted free ventilation. The results have every guarantee of ac-
curacy and are worthy of unusual confldence.

M. Becquerel's first memoir relating to the forest meteorological
problems was not published until 1864. Meantime other students of
the subject were showing signs of activity. The French forest inspect-
ors Contegril and Bellot took many observations in 1859 and 18G0 on
the disposal of rainfall in woods as compared with that in open fields,
and similar observations were taken in southern France. Hofmannat
Giessen, and Baur at Hohenheim had been taking forest observations
for some time. In 1802 and 18G3, Krutzsch, in Saxony, established
nine forest stations; several were established in Bavaria, and Count
Berchem-Hainhausen, in his private capacity as a large landowner,
sup])orted two in Bohemia. In 180G-G8, in Posen, Rivoli, also a private
landowner, made many simultaneous comparative observations in for-
ests and fields, including temperature, humidity, and evaporation. His
observations on individual features of forest action were conflned to
the season at which they were greatest. The fertile idea in modern
forest meteorological observations consists in having comparative sta-
tions at which observations are taken simultaneously. These stations
must be as nearly as ])ossible alike, except that one is to be in the
woods, the other outside. This happy idea, already employed by Bec-
querel and Rivoli, and donbtless by others, but not by all, was first
put to systematic use under governmental auspices at stations near
Nancy. The French forest administration undertook to set at rest the

FRENCH INVESTIGATIONS. 29

questions of forest influence on climate, and to this end instructed the
school of forestry at Nancy to undertake the necessary studies. In
April, 1886, observations were commenced by Prof. Mathieu, of Nancy,
with the view of determining the inttnence which a wooded or unwooded
country has upon the moisture received from the air. Three stations
were selected ; they were described by Dr. Hough in his report for
1877,* page 202. The most interesting feature about them is the forest
rain gauge at Cinq-Franchees. It was made to cover exactly the sur-
face covered by the tree-top and was jilaced with the tree in the center,
careful arrangements having been made to direct into the gauge all
water Avhich ran down the tree. In 1870 a fourth station was added,
that of the agricultural station of the school. The altitude of this sta-
tion is 712 feet, somewhat less than the others.

In February, 1874, forest observations were undertaken by MM.
Fautrat and Sartiaux, and contiiuied by the first. The stations were
in the department of Oise, a few miles north of Paris. This is a rolling
territory of no great elevations or irregularities, and the southern part,
containing the better-known forest of Compiegne, is well wooded and
somewhat swampy. One pair of stations was in the state forest of
Halatte, about a cluster of oaks and hornbeams of twenty or thirty
years. The forest itself is large, containing about twelve thousand
acres. M. Fautrat had the happy idea of trying some observations
above the forest. His stations were therefore duplicated, one being
below at the height of 4J feet (1.37 m.), the other above the trees
about 20 feet (6.09 m.), and 46 feet (14.02 m.) above the ground. The
forest station was reproduced 1,000 feet (304.8 m.) away in the open
fields. Each station had thermometers, psychrometers, rain gauges,
and evaporometer. Under the trees tiie number of rain gauges wiis in-
creased to six, distrilnited in such a Avay as to give a good indi(;ation
of the amount of rain passing through the foliage. The stations were
on a soil of fine sand, cemented by clay, and were at an elevation of 354
feet (107.9 m.) (base) above the sea and 400 feet (121.9 m.) taken above
the trees. About 5 miles (8.04674 kil.) away was a second pair of double
stations; they were in and near the state forest of Ermononville. The
trees were twenty-five year pines, about 40 feet (12.19 m.) high. The
stations at Halatte were duplicated here, including the elevation of the
station above the trees. This brought the higher instruments to within
a very short distance of the foliage — 5 or 6 feet. The soil here was a
coarse white sand, and the open-field station was on a sau<ly plain.
The elevations above the sea were 302 (92.05 ni.) and 341 (103.9 m.) feet.
There were six ordinary gauges under the trees and an additional large
one, over 5 feet (1.52 m.) in diameter, which surj'ouuded the trunk of a
large pine tree. The printed observations (in the Coniptes roidm) are
somewhat fragmentary and extend over about four years. There has

•Keport on Forestry, vol. i, Uuited States Departmout of Agriculture.

30 FOREST INFLUENCES.

been printed a separate publication on this subject, but I have not seen
it; its data are said by Dr. Lorenz-Liburnau to differ somewhat liom
those in the Com])tes rendus.

GERMANY.

In 1808, witli the aid of the Bavarian Government, Dr. Ebermayer
began his well-known series of observatious. The results for the first
year were made the subject of an independent publication. They relate
exclusively to the problem of the interior forest climate, and the con-
clusions reached by Dr. Ebermayer have been generally confirmed by
later observations. Dr. Hough has devoted twenty pages of his report
for 1877 (pp. 230-251) to an account of this service and an abstract
of its results, so that they do not need to be detailed here. Dr. Eber-
mayer, then in the forestry school at Aschaffenburg, is now professor
at Muuich, and his publications on forest meteorological topics still con-
tinue. In the beginning of 18G9 three pairs of parallel stations were
established in Switzerland. They are also described by Dr. Hough in
his report (p. 255) cited above.

In 1875 the German meteorological forestry service was established.
The stations are here in pairs, one in the forest, the other outside, and
the observatious are made in succession at the two stations of each
pair. The rule was adopted to have each of the stations at least 200
yards (182.8 meters) from the margin of the forest, but this was not
found to be always practicable, and the distance in one place sinks to
less than ninety yards (82.29 meters). The stations were selected to rep-
resent every variety of position possible as to distance from the coast,
topography, and kind of trees, and each station of the pair was put
under very similar conditions of soil and elevation. Seventeen such
stations have been established; ten in Prussia, three on the Imi^erial
lands, and one each in Wiirtembcrg, in Brunswick, in Thuriugia, and
in Hanover. The observ-ations at Wiirtemberg station continued only
from the beginning of 1880 to the end of 1885. The other stations are
still active; they are conducted on a uniform jdan and arc under the
direction of Dr. Miittrich, of the academy of Eberswalde. The instru-
ments are made by Fuess; they are carefully compared when put up
and are regularly inspected once in five years. Instruments are kept
in reserve at each station so that the defective ones can be at once re-
X)laced without breakiug the continuity of the observatious.

The Hanover station is Lintzel, and this is exceptional in that it is
among the young plantations of trees now formed on the great Luue-
berger heath. Field and forest stations both exist, but tree-top obser-
vations are not made. The in^ncipal features of the stations are shown
in the accompanying table, in which they are arranged in order from
northeast to southwest. By the elevation is meant that indicated by
the barometer, which is so nearly the same for each station of the pair
that it may be taken from either. The readings are exact, except for

SWEDISH OBSERVATIONS.

31

the third, fitth, sixth, seventh, eleventh, twellth, and fourteenth stations,
where they are approxinaate.

German stations for forest meteorology.

Station.

Latitude.

Longi-
tude east
of Ferro.

Eleva-
tion.

Eancl of trees and

age at founding of

station.

Distance to margin
of forest.

Beginning

Forest
station.

Field
station.

tions.

Fritzen

O (/

54 50
53 34
50 28
52 50

50 36

51 52

51 45

52 16

52 59

55 16

53 30
50 53
50 27
48 29
48 50
48 5!)
48 25

O '/

38 13
59 9
34 0

31 29
28 28
28 14
28 10
28 38
27 55
27 9
25 14

25 54

24 3

26 59

25 28
24 57
24 57

Feet.

' 128

423

2,484

79

2,349

1,296

2,549

420

325

125

10

1,998

2,024

2,403

4!.9

1,158

3,064

45-year spruce

80-140-year pmes. .

45-year spruce

45-year pine.s

60-70-year spruce . .
65-85-year beeclies.

45-year spruce

60-year beeches

Feet.
262
679
591
410
984
307
328
984

Feet.
459
433
869
591
492
1. 138
G50
656

1873 X i

1873 xii i

( /jirlsbor*^ .......

1874 xi i

Ebersnalde

SchmieiletVld

rriedrichsroile

Sonneuborg

Marienthal

Lintzel

1875. xii, i.
1881. X, i.
187+. x.i.

1877, vi, i.

1878, V, i.
1881 iii, i.

Haderslebeu

70-80-year beeches .

20. year pines

70-year beeches

45-year spruces

50-year spruces

55-65-year pines...

45-year beeclies

60-80-year beeches .

410

656
2, 461

361
1,040
4,167

820
3,937

394

1, 640

640

328

650

2,192

820

5,249

1875. x,i.
1876 \ i

Laliuhof

HoUerath

1874 X i

St. Johann

1875 V i

^euniath ...... .

187.5 V i

Melkerei

1875 V i

The field station is generally on cultivated soil, in a few cases on
meadows or marshy land.

A monthly and an animal report are published from these stations
and the results have been made the subject of several special studies —
notably the temperature observations have been discussed quite re-
cently (1890) by Dr. Miittrich.

The observations are taken on the same plan as those in Bavaria, but
observations of the tree temperatures and the measures of ground water
are omitted.

SWEDEN.

In 1876 the Swedish observations were begun under the direction of
the meteorological central office, and chiefly at the expense of the State.
The stations are here selected in a dilfereut manner from those in the
other services. There are few that are under trees; there are many
more in small glades or openings in the forest, and the comparative
lield stations are usually quite distant and in a freely open country.
This offers some advantages, for it permits the discussion of glade
climate or the climate in the vicinity of forests, as distinguished from
that under trees or that in open regions away from trees. A disad-
vantage is found in the fact that the results are not strictly coni]>aral)lo
with the results of observations taken elsewhere. Tlie observations
have, for several years, been under the direction of Dr. H. E. Hamberg,
who has published a discussion of the results in temperature and
humidity, the first in 1885, the second in 1889. The i)ublicatious made
up to this time relate to temperature and humidity only.

32

FOREST INFLUENCES.

^1

^ o

re '-a

p. :^ ft

n c3

('.

o
o

>v

--d

o o o o
o o o o

i-H O O »

o o o o •=>
o o o o o

00 O CO CJ CO

s

00

s

o

m

3

a

c3

3
O

iS ^ >

<« 03 g O =S

S O W O ^

3

2 F^ o

•"^ Ct Cj

be B o

J- — o

M r-l i--

O
O

C a? c3

H O

i ^ ^ ,-H 1-t r-H l>J CI

■^ Tj< lO w

ts

tjr. cr

o

COO-COCO«MOO"*b-t-eOWC5-tOOCOO<MTH
- (M ^ CO re CI CO « r-( r-l CI rH r-l CI -^ -^ CI CO Irt »0

o o cr- r- I- 'C

CO CO Cl Cl

I- Tl --H .— < r-( Oi 00 CC

oooocsCic:oc5Ci*r;<:n

tr: o o -^ »n

CO' CO CO CO i-H

OD -X 00 OO o '-0

CD CO i-O tO O lO lO m i.O lO »0 lO UO to lO lO »o

C O Ph C

« iJ -- "i^ '^

P^ o o o

_ 'X 4^ ^ -, 7?

a u a, o '^ -^

C o ■-< p, • c

C fH O O . hH

o

.2 -5

ft

•a-

<1

0) S

1 gP

:ce

[3 K ^

5 =1

3 tt .5
:<1 cc ft

P4

3 tB

CQ

AUSTRIAN OBSERVATIONS. 33

The stations mentioned so far were generally parallel stations, tliat is
double, one set of instruments being in the woods, the other outside.

AUSTRIA.

The radial stations proposed by Dr. Loreuz Liburnan were estab-
lished in Austria in 18S4 and following years. The outside stations
in each radial s} stem were to be in various directions and at varying
distances from the stations in the forest.

Dr. Liburnan foun<l great difficulty in arranging them, since some of
the stations should be at considerable distances, and all of the same
group should l)e at the same sea level. I^aturally the arrangement
would alsf» require a separate and distinct forest of no great extent.
These conditions were hard to fill and but few suitable places were
found in the whole Austrian Empire. Up to the i)resent time only
three such groups of stations have been established:

(1) In northeastern Austria, the group about Karlslust on an ex-
tended plateau, between Ketz and Znaim, with an average elevation of
1,300 feet (396 m.).

(2) In eastern Galicia, near the Russian boundary, on an extensive
plateau, between Konstancyai and Skala, with an average elevation of
886 feet (270 m.).

(3) On the northern base of the Carpathians, in the undulating region
near Eachiu, at an elevation of 1,300 (396 m.) to 1,600 feet (487.6 m.).

01)servations of temperature and humidity were taken under the
trees, also in the tree crowns and at various elevations above the trees.
Corresponding observations at the same heights were made over open
land.

The results of the observations during the years 1885 to 1887 so far
published have not been before the writer.* From a review, however, it
is learned that the director. Dr. laburnau, a most competent authority,
draws the following conclusions:

(a) The ndrnial decrease of temperature with the elevation, observed in the open
country, is modified in the forest, during the day, by the warming of the crown
cover, especially of the upper portions. In the forest the crown cover has the same
relation to the temperature as have the soil and the smaller vegetation in the open
country. Thus the normal decrease of temperature with the elevation remains uni-
form in the open country, but is more modilied in the forest under the influence of
the warmed crown-cover thim in the open country under the influence of the soil.

(6) During tlie night the crown cover has also the same relationship to the soil
and vegetation of the o])en connlry, except tli.it temperature reductions have dif-
ferent values.

(c) As the temperature in the forest is higher in the upper regions of the crown
cover than in the lower and among the trunks and on the ground, especially during
the day, it follows that the shade contributes more to its reduction of temperature

* Investigations into the humidity and condition of temperature of the forest and
of the open country, and the results of the forest influence upon the climate in its
proximity.

12414-l!To. 7 3

34 FOREST INFLUENCES.

tliau transpiratiou; altogether, the cooliug effect of transpiration up to this time
seems to have been overestimated.

(d) Tlie warming effect of the upper crown cover upon the temperature during the
day in a beech growth extends only about 2 m. (6^ feet) above the crowns, and
indeed is most pronounced in the warmest mouths.

(e) The absolute values of the temperature differences effected by the forest are
not very important, but depend to a great exteut upou the species of the trees on the
one hand and the kind of soil cover of the open country on the other.

(/) It is remarkable that in the forest the relative humidity increases and decreases
with the absolute humidity, whereas it is known that in general, and also at the sta-
tion in the open country, these two climatic elements arc inverse. This is accounted
for by the fact that the forest is a source of atinosplieric ac^ueous vapor as well as

of cooling.

(r/) The results of the investigations at Ried show that the aqueous vapors from
the forest often affect the air to a great height. With this increase in the humidity
of the atmosphere the chances of preciiiitatiou are increased.

From tliese observations Dr. Liburiiau draws the following conclu-
sions regarding the influence of forests upon the general climate of the
region in its vicinity:

(h) Since the temperature in the forest, to above the crown-cover, is in most cases
lower than it is beloAV in the open country, it follows that modifying currents can
arise from the forest at any time of the day, especially in a generally still atmos-
phere. In summer, duriug the night and toward murniug. differences of tempera-
ture can produce a current of cool air from the forest and from the air strata imme-
diately above the forest; in the warmer part of the day, however, such currents can
be looked for only from below and in the crown cover.

((■) In the summer half of the year, winds coming from a distance and crossing a
forest can be cooled below their original temperature by the influence of the forest
considerably in the night and toward morning, but less duriug the day.

(j) The conclusions which Dr. Li burnau arrives at concerning atmosi)heric humidity
are based for the most part on the results of the observations at Ried. It was est.ab-
lished that Aviuds from the forest (at least J'rum a beech forest) convey a greater
water supply into the neighboring open country. It depends, however, upon the
original humidity of winds crossing a forest and upon the distribution of tempera-
ture whether or not their humidity will be increased and they will subsequently in-
crease the humidity of the adjacent open country.

Hence Dr. Liburnan concludes that the humidity of warm and dry
winds will be increased by the presence of a forest, but that such winds
as are already quite thoroughly saturated cau receive no essential
addition of aqueous vapor, but only an increase in i-elative iiumidity.
It was also observed at Ried that winds coming from the forest to the
station in tlie open country have a higher humidity, botli absolute and
relative.

(A) Finally, Dr. Liburnan remarks with reference to the distant influence (of the
forest) that this absolute humidity, as increased by the effect of the forest,* remains
in the atmosphere as long as it is not lost by condensation, and that, on the other hand,
the Increased relative humidity decreases as soon as the forest air reaches the warmer
open-country.

^Within the last two months a further contribution from these stations has been
published, discussing more in detail the influence of forest areas on the climatic con-
dition of the neighborhood, which publication, unfortunately, could not be obtained
in time for a review in this report. ^- E. F.

OBJECTS AND METHODS. 35

OBJECTS AND METHODS.

Ill 18G8, before the publication of Ebermayer's results or those of any
other systematic series of co-operative observ^ations in forest meteor-
ology, the Congress of German Foresters and Farmers at Vienna,
strnck with the paucity of results from the much talkedof forest mete-
orological observations, put the subject in the hands of a special com-
mittee composed of the most eminent German students and authorities
on the subject. The committee met at Regensburg in November. Its
attention was chiefly directed to other aspects of the forestry question,
but it proposed a sort of programme or series of problems as the projier
object of forest meteorological observations. These are:

(1) Determination of the temperature under woods and outside.

(2) Investigation of the temperature of trees as compared Avith that of the soil and
air.

(3) Investigation of the soil of forests, determination of its temperature at dif-
ferent depths, situations, exposure, and under differeut kinds of trees.

(4) Influence of the covering of the soil on damage by frost.

(5) Determination of rainfall under woods and outside.

(6) Influence of the covering on the humidity of the soil and on the evaporation
from it.

(7) Determination of the amount of evaporation in the forest and without.

(8) Determination of the ground water at diflerent de])ths, under woods and out-
side^ with various aspects and under various kinds of trees.

(9) Determination of the amount of carbonic acid and of the humidity of the for-
est during the season of vegetation, under a more or less compact cover of foliage,
under different trees and in varied situations.

This is au extensive ])rogramme, and is of interest as the first which
had been formulated under such auspices. It will bo noticed, however,
that it does not include observations at any elevation, and especially
observations above the woods. It does not seem to have occurred to
the committee that meteorologically the forest is a sort of surface cov-
ering to be compared Avith that of sod or crops, sand or water, and that
in other cases observations are taken as well above the surface cover-
ing as in it or below it. Observations above woods w^ere, however, not
inchuled by the committee, and since this programme served to give
object and aim to tlie new services of that day, as well as to those
established later, these important observations have been almost
entirely neglect43<l.

The Piiissian (afterward the German) forest meteorological service,
established six years later, adhered, in most points, to the Kogensburg
])r(»graiimie. In its iiistru<;tioiis to its observers it dotincs their task
as that of furnishing observations on —

(1) The temperature of the air in forests compared with that in open fields.

(2) The temperature of the air at 5 feet elevation, compared witli that in the tree
crown.

(3) The hygrometric condition of the air wiihin and wiMiont woods.

(4) The same at a height of 5 feet compared with that in the tree crown.

36 FOREST INFLUENCES.

(5) Evaporation within and without the woods.

(6) The rain and snow reaching the ground within and without the woods.

(7) The temperatures of forest soil at the surface and at the depths of 6 inches,
1,2,3, and 4 feet, compared with the temperatures at corresponding depths in the
open fields.

There were also the highly important conditions attached that the
instruments and their disposition should be alike, and that the hours
of observation and methods of reduction should be the s'ame.

In the second international meteorological congress, held at Rome
in April, 1879, it was resolved that —

In order to contribute to the progress of forest and agricultural meteorology, the
congress recommends as a programme of study :

(1) The influence of the meteorological elements upon vegetation.

(2) The inverse influence of vegetation upon the meteorological elements.

(3) Agricultural warnings.

The congress, finding the subject too important to treat hastily, con-
cluded to charge the international committee with the convening of a
special international conference to take into consideration the develop-
ment of agricultural and forest meteorology.

Such a conference convened at Vienna on the 6th of September,
1880. This was so representative a meeting and illustrated so well
what the most able and most experienced men thought on this sub-
ject that the reports, remarks, and resoluticms will be given at some
length so far as they relate to the action of forests on climate. Dr.
Lorenz-Liburnau, of the Austrian Department of Agriculture, and
president of the conference, said that when the influence of vegetation
ou climate was referred to, solitary plants could not be meant. It was
a question of masses of vegetation, such as meadows, moors, and forests.
There was also a distinction to be made between the climatic action
produced within the massed vegetation and that upon the vicinity. It
is only for a forest that the question oftlie interior climatic action can
arise, and on this question certain definite numerical results had been
obtained for central Europe, in certain localities. These show that
the inner temperatures are slightly lowered, the temperature extremes
are moderated, the relative humidity is raised, and the evaporation
diminished, while it remains doubtful whether the absolute humidity
or the frequency or quantity of precipitation are changed.

As to the influence of vegetation on its vicinity, exact data do not
exist, either for the forest or for other forms of vegetation in mass.
Yet, ou theoretical grounds, it can not be doubted that such an intlu
ence exists, within certain limits, and appropriate methods of observa-
tion must be able to show it. They should include observations on the
temperature, precipitation, cloudiness, insolation, and wind. The con-
ference adopted these views.

As to the method of observation, Dr. Liburnau ])ointed out that the
existing arrangement of one station within and a station for compari-
son outside the region covered by the vegetation to be investigated

OBJECTS AND METHODS. 37

might be called the method of parallel stations. This is a commendahle
arrangemeut, and promises the speedier results, the more the climate
at the station takes on a continental character. The distance of the
stations of each pair from each other is an important point. In each
pair the meteorological elements should be taken at the ordinary height
and also in the tree crown. The temperature of trees and soil should
be included. As to precipitation, it is of importance to get not only
the amount which passes through the foliage, but also that which runs
down the trunk. The penetration of precipitation into the soil, the
evaporation from the surface in different kinds of forests and with
different coverings of the soil are of importance because the state of
humidity and temperature deiiends on them. Observations of ozone at
different heights are contingently recommended. These observations
should be exactly duplicated, so far as possible, at the held stations.
The treetop observations should be dui)licated by field observations at
the same height. To get the influence of the forests on their vicinity,
a different system of stations is desirable; this should be a system of
radial stations as contradistinguished from the parallel system. This
consists of one station in the forest and several outside, arranged in
the directions of the four principal points of the compass. There should
be at least two in each direction and at different distances from the
forest, so that one set should include at least nine stations. All the
observations carried on at the parallel stations should also be taken at
the forest station of the radial system, but at the field stations the
number of instruments may be reduced to a psychrometer, evaporome-
ter (Piche's), and a rain gauge, and the observations at the height of
the tree crown may be omitted.

The question of exposure of thermometers aroused much discussion
and it was agreed that comparative observations should be under-
taken. As to hours of observation. Dr. Miittrich favored morning and
noon (2 p. m.), while M. Grandeau and Dr. Ebermayer favored morning
and evening hours. Dr. Eiegler thought the 2 ]). m. observation was
to be especially recommended because it was about the hour of maxi-
mum forest effect. The conference finally decided to recommend
morning and evening hours, with the reading of maximum and mini-
mum thermometers. Observations of insolation were discussed, but
MM. Mascart and Tacchini assured the conference that no suitable
actinometer exists. It Avas, therefore, decided that observations should
at first be undertaken only on the time, dnration, and relative inten-
sity of sunshine; the latter on sensitive paper, at stations of the first
order. The subject of an actinometer capable of giving insolation tem-
peratures with exactness was recommended as a subject worthy of the
especial attention of scientific institutions.

It w^as decided that soil temperatures should be taken at various
depths up to 0 inches (1 mm). The humidity should be obtained by
the means of August's psychrometer and a hair hygrometer, and tri-

38 FOREST INFLUENCES.

daily observati©iis sliould be taken. A long discussion followed on Dr.
Libnrnau's recomniendation of Piclie's evaporometer. It was stated
that tlie observations gave no means of comparison with actual natural
processes, that the readings of the instruments were untrustworthy
and the errors hard to correct, that it was desirable to follow more ex-
actly the processes of nature, that it is not possible to find an instru-
ment fulfilling all necessary conditions, and that on this account the
Italian service had rejected all evaporometers. The conference finally
decided that, while observations of evaporation were important, there
is no form of measuring instrument which can be especially recom-
mended. An instrument should be invented which would permit the
measurement of evaporation from a free water surface and also from
soils of different characters. Provisionally observations should be
carried on with the weighing instruments and with Cantoni's modifica-
tion of Piche's instrument.

Clouds should be observed; fog should be recorded when it covers
the station and its density should be given on a predetermined scale.
Herr Friesenhoff recommended a measurer of dew depending on the
power of rock salt to appropriate atmospheric water. Drs. Neumayer
and Ebermayer expressed the opinion that balance drosometers should
be tried. They should have larger dimensions (2 feet square or more)
which should be filled with different materials. The conference com-
mended the subject of a good measurer of dew as a jjroper subject of
study. The measurer of precijiitation should be on the model recom-
mended by the International Congress and Committee. The seepage
of ground water is important, but the so-called Ijsimeter is not to be
recommended. Wind observations should be taken, as at ordiuarj^ sta-
tions.

The system of parallel stations, already ado])ted in Bavaria, Switzer-
land, and at the other German stations, was especially commended,
lladial stations were commended tor the broader fields of observation,
and it was recommended that a series of such stations should be
established in eastern Germany and in Austria Hungary, and the
final recommendation was made that observations should be made at
the forest station of each radial system, directly over the crown of tlie
trees. Thus this happy method, the fundamental importance of which
makes it strange that it was not employed from the first, reached at
least the stage of recommendation for general use. The practical ditfi-
culty of its use is probably what has prevented its adoption. This could
be easily overcome by the use of registering instruments which are
made to run for fifteen days without attention, or by the use of the now
excellent instruments for electric registry at a distance.

To the proposal of the first meteorological congress for replies as to
what could best be done, Dr. Woeikoff made a suggestion which was
not taken into consideration by the conference. He proposed India as
the best field for such studies, especially with reference to rainfall.

GERMAN METHODS. 39

For this purpose observations could best be taken in tlie central prov-
inces and Berar, as tbey are hottest in April and May and the forests
are extensive. India also permits a study of the effects of irrigation,
on forests.

The German service supj)lies each of its stations with psychrome-
ter, maximum and minimum thermometers, six soil thermometers, ba-
rometer, evaporometer, and rain gauge. The soil thermometers for the
surface and the depth of 0 inches are supported by small tripods. The
four thermometers for greater depths are in wooden cases, open at the
bottom and extending downwartl only so far that the bulb of the ther-
mometer may remain in contact with the soil at the depth required.
The cases are placed together and are open at the top to permit of
ready access to the instruments. A cover over the open ends keeps
out the water and exterior air currents. The thermometers are taken
out for observation, read as quickly as possible, and replaced. The
evaporometer is a square receptacle of zinc with a surface of about 2
square feet and a depth of about 5 inches. A measured amount of
water is placed in it, and the loss, expressed in thickness of layer evap-
orated, is noted at the beginning of each month. This instrument is
protected from iirecii>itation by a wooden roof, but the air is permitted
to have free access. The force and direction of the wind are estimated,
the former on the 1 to 4 scale.

The rain gauge has an opening of about 2 square feet (18.58*^"') and
is placed at a height of 5 feet (1.52'"), like the other instruments. The
snow gauge has the same opening, but a larger body, to prevent the
snow from being blown out.

The observers were intrusted with other observations, esi)ecially
those relating to the leating, flowering, and other .stages of certain
plants, and the list of plants to be observed is given. It includes
26 wild plants, mostly trees and shrubs, and 7 cultivated ones.
Observations are required of the arrival and disappearance of a few
designat«Hl species of birds and insects, and the time of heat of the
deer and rabbit. The order in which observations were to be taken,
the mode of recording, the methods of reduction, and all other details
are minutely given in the instructions to observers, and j)ainstaking
care is required throughout.

From 1884 Dr. Lorenz-Liburnau put in operation some of the fertile
ideas which he had already suggested. The three systems of radial
stations were organized as already described. Dr. Liburnau thought
that the general cliaracter of the influence of the forest on climato
was already determined and that there was now required only the
elaboration of details. The radial stations were designed to fix the
relations of the forests to the climate of the vicinity with varying direc-
tion of wind, and a special series of observations were made to fix the
relations of forest humidities and temperatures at various elevations,
below, in, and above the tree tops, and to determine their modifications
with general changes of the weather.

40 FOREST INFLUENCES.

The most elaborate series of observations of tbis kind were tbose taken
near Ried, in lower Austria, in 1887 and 1888. They were especially
devoted to the problems of liumidity, and to obtain tUorougbly satis-
factory results tbe ordinary wet and dry bulb tbermometer was rejected
and an exact chemical method substituted. This consisted in passing
a known volume of air through calcium chloride and phosi)horic acid
and ascertaining by weight the water taken up by the latter. The case
containing the chemicals was elevated, by means of a pulley on a mast,
to the height at which the observation was to be taken, and the air was
drawn through it by means of an aspirator below, connected with it by
a rubber tube.

The observation occupied an honr and thirty minutes and the tem-
l^erature was taken at the beginning and end by means of an upsetting
thermometer, elevated by means of a pulley to the height required.

The results were very exact for the absolute humidity, less so for the
relative. The heights at which the observations were taken were 4
inches (101.0"""), 16 (4.87'"), 36 (10.97'"), and 51 feet (15.5'"). The
second was under the tree crown, the third in it, and the fourth was
18 or 20 inches above it. The trees were red beeches, 60 to 70 years old,
with a few white beeches and firs. The station itself was in and over
the beeches.

SOIL TEMPERATURES.

The soil affords an excellent means of measuring the average tem-
perature effects of the forest. This is due in part to the fact that the
temperature of the air is not derived so much from the direct rays of
the sun as it is from the heat reflected or radiated from terrestrial
objects. These absorb the heat from the sun's rays more readily than
does the air and yield it again to the air. The soil serves in another
way also to average the effects of forests. The soil is more slowly
warmed by the sun than is the air, and the deeper the stratum the more
slowly it is warmed. The result is that rapid and passing changes of
air temperature do not affect it, or affect it only superficially. It tends
to show only the great periodic changes in temperature, and from it we
are able to ascertain Avhat influence tlie forest has on these changes.

The results which follow are always given in terms of the difference
of temperature of the soil in woods and in open fields. If we put W for
woods or forests and 0 for open fields, the data given hereafter* for

*NOTR ON THE CONSTRUCTION AND READING OF THE DIAGKAMS IN THIS REVIEW. —

The horizontal Hues (ordiuates) above or below the zero line represent values or
amounts, degrees of temperature, inches of precipitation or evaporation, percent-
ages, etc. The vertical lines (abscissae) represent time, dividing the field into
twelve seasonal divisions corresponding to the twelve mouths of the year, the outer
lines both standing for the month of December or commencement of winter. The
curve lines are constructed by noting on each monthly line the values found for the
month, and then connecting these points by either straight or rounded-oft' lines.
Unless otherwise noted, the values so plotted are the differouces between the read-

SOIL TEMPERATURES. 41

temi)erature are always, except where otherwiKse stated, the vahies of
W—0. A positive sign ( + ) imlicates always that the woods are
warmer thau the fields; a negative sign ( — ), that the woods are cooler
than the iields, Tlie degrees are FahrtMiheit and the other units of
measure are the usual English ones, unless otherwise stated.

In Fig. 1 are given the mean annual differences of temperature for
fields and woods at the surface and at the depths of G inches and •! feet.
These are for the German stations, a brief description of which may
be found on page 30. They are arranged in the order of their differ-
ences at the surface, and at the right hand are given the average
values. It will be noted at once that the differences vary much for
the different stations. The mean annual difference for Hadersleben is
less than one degree at the surface, while that at Melkerei or Neumath
is nearly four degrees. It will also be noticed that the differences at
6 inches vary with respect to each other and also with respect to those
at the surface. They are generally less than the surface differences at
the same stations, but HoUerath furnishes a notable exception. On
the whole they run with the surface differences, but are somewhat less.
Again the differences at the depth of 4 feet differ among themselves,
but not so much as do the otliers. Their relative uniformity is due to the
depth at which the temperatures are taken, the smaller surface changes,
in temperature and in water contents, being little felt. These dififer-
ereuces are usually greater than those at the deptii of 6 inches and
less than those at the surface. Bat there are exceptions to this, not-
ably in the cases of Hollerath and Bt. Johann. These differences in
the action of the forest are due to several possible causes, such as dif
ferences in soil as well in the field as in the forest station, in kind of
trees, in their density, in the exjiosure, in tlie character of the air
drainage.

iuo-s under two sets of conditions, namely, in most cases tlie values which were found
for the stations in the woods (W) diminished b\ the values found for the stations iu
the open iield (O), or IF — O.

The value of this difl'ereuce is positive, if the curve runsabove the zero line — that
is to say, the records for the woods ( W) showed higher values than that for the open
held ( 0) ; it is negative, /. e., the record for tlu'. woods was lower, if tiie curve line runs
below the zero line. The greater, therefore, the vertical distance of any point iu
the curve from zero line, the greater is the influence of the woods. In temperature
readings, for instauce, the curve above the zero line would dtMiote tiiat the woods
were warmer; below the zero Hue, that tlio woods were cooler than the open field
l»y as many degrees as the curve runs above or below the zero line, the latter repre-
senting that state of conditions when W := O, (. c, when there is no dilVcieuce in the
readings for the two sets of conditions.

Where values for each set of conditions are plotted separately, the area included
within the two curve lines (hatched) exhibits the diftereuce betAveen the woods and
opea field.

To exhibit more readily the amount of influence of the forest, the areas included
by the zero line and the curve for mean values is also hatched in most cases.

42

FOREST INFLUENCES.

Average,

Haderslebeii,
Lintzel,

Hdllerath,

Schoo, -

Lahnhof, -
Marienthal,

Fritzen, - -
Friedrichsrode,

St. Johann,

Kurwien,
Carlsberg,

Eberswalde,

Sclimiedefchl,
Hagenau, -

Sonneiihcrg,
Neiimath, -

Melkerei,

Scale of degrees.

—

■~~

1

1

I

■"■■•■"

_ _l 1

_ _ ^ — -^

—

■""

- —

-

• -

■

■—

—

1

~

-— — -

_^-i.

•~

_[_

rz

Scale of degrees. L

' At surface. Al- six inches.

(M

I

en

I

_1_

I

_J

-" At four feet.

Fig. 1.— Differences of nuaii annual temperatures of soil (TT— O).

DIFFERENCES OF SOIL TEMPERATURE. 43

Some of tliese will be discussed later; the material on hand does not
permit the discussion of the others. They are instructive, however,
in showing- how great is the difference in action of diiierent forests.
For instance, the effect on the temperature of the surface at INIelkerei
(_3".95) is nearly five times as great as at Hadersleben (— 0".S4). This
difference of action of the forest is fsir-reaching, for it extends through
a layer of 4 feet of badly conducting soil. The ditference at this depth
at St. Johann (— 3".03) is about tliree times that at Eberswaldo (— 1".0G)
or at Hadersleben (— 1".04), and is five times that in the young forest
at Lintzel (— O^oS). The average of the seventeen stations (represent-
ing about two hundred years of observation) should give us good and
significant results. It shows for the surface— 2".59, for a deptli of G
inches (152 mm.)— f.ST, and for a depth of 4 feet (1.22 m.)— 2".02. The
influence of the forest on the soil, then, is a cooling one, on tlie aver-
age, and for central Europe the cooling amounts to about two and a
half degrees for the surface. The cooling is due to several causes : The
first is the shade; the foliage, trunks, branches, and twigs cut off much
of the sun's heat, absorb and utiliz<' it in vegetative processes, or in
evaporation, or reflect it away into si)ace. Thus the surface soil in the
forest receives less heat than the surface of the fields. The same screen
acts, however, in the reverse direction by preventing radiation to the
sky, thus retaining more of the heat than do the open fields. The bal-
ance of these two processes, it seems from observation, is in faAT)r of
the first and tlie average result is a cooling one.

But the thatch of living foliage is not the oidy screen possessed by
the forest soil. It has in addition the screen of the forest litter, and.
this is in a condition to be even more effective than that of living
foliage. It lies in contact with the ground, preventing the dissipation
of the heat of the soil by moving air, and at the same time lying so
loosely as to tV»rm air spaces, which act as insulators in the way of
preventing the exchange of heat between the forest soil and air.
These cooling influences are enforced by the moist condition of the
forest soils. It does not warm so easily as the drier field soil, and more
of the heat which reaches it is used in evaporative processes than in
that of drier soil. The balance of all these processes is, it appears, in
the direction of cooling. And the cooler forest soil will cool, to some
degree, the air in contact with it, and this air, flowing off to some
other place, will take this cooler temperature with it, so that the cool
ness of the forest soil will make itself felt over the vicinity.

The differences of temperature at the depth of G inches (152 mm.)
are more tlian half a degn^e less than at the surface. In this is to be
seen the specific effect of the forest litter; it adds a covering to that
possessed by the surface, so that while the deeper layer is cooled as
much by the protection from the sun's rays as is the surface, it is not
cooled so much by radiation of heat to the sky. Its temperature is,
consequently, relatively higher, and it approximates somewhat more
the field temperatures.

44

FOREST INFLUENCES.

It is not SO easy to explain why the difference at 4 feet averages
a little less than at the surface and a little more than at 6 inches. It
may be due to retardation in the penetration of diurnal temperatures
into the soil. This may be slower in the cool, moist forest soil than in
the warmer and drier soil of open fields. If this is the case observa-
tions at the same instant at the two stations would determiue the
temperature at two different parts of their curves. The diurnal
changes probably extend to a greater dei)th than 4 feet, and if they

SCHOO.

^

HADERSLEBEN.

Trr

\,

HOLLERATH.

^^

/

y

y

'"

CARLSBERG.

"^

^

y^

^

NEUMATH.

,.-•

^— ;^

\

"\

^

.'-'

~~-

^^

^^V

*■■-

'

/'

^^

EBERSWALDE.

-\^

"'--.,

..--

'

/

1 AH N HOP.

-^

\

^C"^-

s

X<

'^^^^

MELKEREI,

^^^

^v^

s

^

KURWIEN.

-^'^^^

-T^

"x

^\'\

^^'-^

^■^

/

1

^''

FRIEDRICHSRODE ,

.

~~^-^

X^

V^N

\

/

f /

""" —

\V^

^,

"i

FRITZEN.

\

\

^^\

,y'

/

^ !

I'

SCALE

OF.

DEGREES

OF

FAHRENHEIT.

n°

\ '\

\

\

y

■■'/

/

1

1°

\

\

\\

\

-^^

c:!^

/ (

t

?°

\

V"'

•
•*

/

1
1

s°

.^..

—

/

1
1

4"

\

\

1
1

fi"

\

\

/

1

6°

\
\

/

7°

\

a°

s

y

WINTER.

SPRING.

SUMMER.

AUTUMN.

Fig. 2. — Difference of temiieraturo ( TT^O) nt the depth of 4 feet. (Tlio line running nnch>r the name
of a .station is ita zero line. The curve for tlie station i.s represented by tlie nean\st lirolvin, un-
lirolieii, oi' dotted line like that in the margin.)

have different times of nmxima and minima (later in the forest) the
increase of the difference at this depth could be explained. It may be
that a part (undoubtedly a small one) of the differences at the surface
and at 0 inches may be due to this. The observations were taken at 8
a. m. and 2 ]). m.

Turning now to the progress of the difference, " Woods" minus " Open
field" readings, through the seasons, or, as we shall generally express it,
W — 0, as shown by the differences of the monthly means, another series
of interesting facts appear.

DIFFERENCES OF SOIL TEMPERATURE.

45

Fig. 2 shows tlie dift'eieiices of temperature through the year at the
depth of 4 feet. The curves for all tlie stations for Avhich good monthly
means for ii 11 the months could be obtained are put on one sheet for
comparison. In order not to crowd them too much they are placed at
different levels, the zero level for each station being represented by
the line running under the name of the station, and the curves belong-
ing to each station either unbroken, broken, or dotted lines nearest to
its zero line. Seasonal lines are drawn for the monthly values, winter
beginning with December, spring with March, summer with June, and
autumn with September. The stations are arranged in the order of
the greatest summer differences.

The distances between the horizontal lines representing degrees
Fahrenheit, the absolute amount of W — 0 can be i-ead. Thus, for Ebers-
walde, during the month of December, W — 0 is between two and three
degrees, the soil in the forest being by that amount warmer than that
in the open at 4 feet (1.22m) depth, while in the middle of July it is
just five degrees lower.

+ 1

0

'^^

!:tr?<

^-^>.

>^^

'^^.

-r

S

/

'7-

*

-?°

\\

\
\

N

/ /

-?,"

\

*

/-

/ /

'

-4°

K\

V

/ J
/ /

*

-fi°

\^

'v^^

/

A,^

-R°

\

^

'/

-7"

\

/

WINTER.

SPRING,

SUMMER.

AUTUMN.

Fio. 3.— Difl'erence of soil temperature (TF— O), all stations— German observatories.

It Avill also readily be seen that the values for Hollerath, Carlsberg,
an<l Neunmth are never above their zero line; that is to say, are always
negative, the soil in the woods being always cooler.

This figure shows in more detail the differences between the stations
Avhich were so readily apparent in <he annual nutans. The time of
maximum is generally .lidy, bnt Sclioo, in a. mean often years, gives a
sec()ndary minimum at tliis time, and for Oarlsberg and INIelkeiei the
maximnm difference falls in August. Mostof the stations give a, posi-
tive sign Ibr l^ — O in winter, some of them showing a^ large excess
(Eberswalde+2^.41, Fritzen + 2o.20)5 but there arc several exceptions,
and all those which do not cross the zero line come very near to it.
The maximum monthly difference is — 8o.l7 for Fritzen in -Inly. The
least July value is — lo.93 (for S(dioo), less tiian a (juartcr of the pre-
ceding.

The sign of If— 0 generally changes from negative to positive in
Decembei', and the reverse in March. The progress of these curves is

46 FOREST INFLUENCES.

nearly alike, and on eunstructiiii;' the curve lor the average tliis appears
smooth aud symmetrical, as is seen in Fig. 3. The dotted line here
represents the progress of the differences at the depth of 4 feet. All
the irregularities have here disappeared. The temperature is seen to
be warmer in the woods than outside during the entire winter. It
then falls rapidly during the spring, reaches its lowest point about
July 1, and rises again rapidly during the autumn. The broken
lines represent the progress of the differences of temi)erature at the
depth of <> inches and the unbroken that at the surface. The first have
the smaller maximum values and also indicate .. longer time during
which it is warmer for woods. The maximum monthly differences are:
For 4 feet, — 5^.01 in July; for 0 inches, — 5^.48 in June; for the surface,
— T^.O") in June.

A glance at the curves shows more strikingly the retardation sug-
gested by these values. By the course of the curves it appears that
the maxima are successively later as we descend, being a full month
later at 4 feet than at the surtace. Probal)ly the same thing occurs
with diurnal changes; that is, they are retarded by some hours at the
lower deptli. It appears from these curves that, on the average, the
forest soil is warmer than that of tlie open tields in winter, but cooler
in the other seasons, and the total cooling effect is much greater tJian
the warming one. The numerical data show that while the maximum
negative value of W — O at the surface is 7'^.(>5, the maximum positive
value (January) is 0^.66, or only about one eleventh of the former. At
6 inches (152.4 mm) the numbers are 5^.48 and 0^.62, and the ratio one
to nine. At 4 feet (1.22 m) they are 5°.! and 0°.36, or a ratio of one to
fourteen. The forest, therefore, not only cools the soil, but also moder-
ates the extremes of its temperature. The amplitudes of the mean
monthly values are decreased by 0.6(3 + T'^.O.'j = To.Tl at the surface,
0^.(12 + 50.48 = (P.IO and (>o.;3G + 5o.(H = .jo.aj at (i inches ami 4 feet
depth, respectively, and through the soil this moderating influence
nuist be appreciable in the air.

The stations of the German service are so happily distributed that a
study can be made of the differences in the influence of deciduous and
evergreen trees upon soil temperature, and also those for elevation
above the sea. For the first, seven stations under deciduous trees and
seven under evergreens were taken, Fritzen and Eberswalde being
(unitted to get a symmetrical arrangement, and Lintzel, because the
trees were young. The reductions Avere made for the two sets of sta-
tions independently, and the results are shown in Fig. 4. The results
are here plotted separately for the surface, for G inches, and for 4 feet.
The dotted lines refer to the deciduous trees, the broken ones to the
evergreens. The differences are not great, but they are quite consistent
at the three depths. They may be expressed by the statement that the
differences vary a little more under deciduous trees. This is what
should be expected from the character of their foliage. The same

DIFFERENCES OF SOIL TEMPERATURE.

47

explanation would sutticr for the slight retardation of the fall of the
curve for deciduous trees. There is one feature, howeAcr, that is unex-
pected. It is, that the difference between the two sorts of forests is not
especially marked in winter. As a matter of fact, in the high latitudes
in which the observations are taken, the sun is so low in winter that the
bare trunks and branches are fairly efficient screens against his rays.

The darker color of these bodies also permits them to be more
warmed up when the sun does appear. Besides, the persistent cloudi-
ness of the winter and the covering of sikow in these latitudes tend to
equalize the effects of the soil of these two types of forests at this
season. The annual means for the two kinds of forest differ but little.

lij

u.

QC

o

u.

1-
<

+ 1°

g

-1^

■:^'-'

"

'~--

N.,

^^'

\

V

N

^-5^->

^

\

. -

- _ _

y/

-4°

s V

y

+ 1°

-

^■- .•

0

■

:_-^--

_

^^

-r---~

u5

UJ

I
o

z

X

<

- r

-2°

\ ^

\

/r'

-S°

k

/

V'

-4°

\

f

-5°

>T_

yy

+ 1"

r ^s-^

T" ^,

0

•"-

■--=r

-.^

,''

m
111

o
<

c

D
OT

1-
<

- f

■\

L

-2°

C\

t

-3"

\'-

^

—

-4°

\

.

/

-5°

\ \

-6°

,

.'-' /

-7°

%-

^

/

-8°

»,

^ f

'

WINTER

SPRING.

SUMMER.

AUTUMN

__ _ __ DE

CiDUOUS TREES. EVEaGREEN TREES J

Fir;. 4.— Diflference.s of soil temperatiur (wood.s and open fields). Compari.son ofdeciduon.s and ever-

For the surface they are —I'^Su for deciduous trees, — 2<^.0li for ever-
green trees; at 6 inches they are-io.89 and-io.82; at 4 feet— 2o.05 and
—20.19. They are represented in the diagram for the 4-feet depth by
the two straight lines, to show the slight difference between the two
types of forest cover. The difference in amplitudes is somewhat
greater. The amplitudes are:

Surface. Six iuclios. Four foet

For deciduous trees .
For cvfriireeii trees .

80.50

6°. 04
00.04

50.12
40.37

48

FOREST INFLUENCES.

As to the effect of elevation above sea-level, there are seven stations
over 1,900 feet (579 meters) and seven below 500 feet (152 meters). The
mean elevation of thetirst set was about 2,100 feet (732 meters), that of
the second about 240 feet (9.3 meters). Lintzel was omitted, and ever-
green and deciduous trees were about eiiually distributed in the two sets.
The reductions were performed for each of these sets separately, and the
results are shown in Fig. 5. The differences are a little more marked
than in the case of deciduous and evergreen trees. The range is

V.

o

<

+ r

0

^

^.

-^"^

- 1°

-.-^""^

>.^^.

^

/

..

-?°

X

-r

'^.^

s

■^ 7

f

-4°

VN^

- r

x"

r^-.

— -^'

^

0

"^

^~-->

y

to

Ul

I
o

z

X

1-
<

- 1°

/
/
^

^''

-''"

-?°

"A

^

•'

-3'

1

!^

/

/

.'-'

-4°

/

//

-5'

•^N,

4 r

<;

^ . ^

y

0

■^-,

V

^

y

•

111
o

tc

03

1-
<

-1°

/

/

.'''"

-2°

/

y

-3°

J

-4°

\

/

/

-GT

^^N

v.

y

'■'/

-7°

■

-r8°

— ^

WINTER

SPRING

SUMMER.

AUTUMN. 1

. 2400 FEET. 240 FEET, J

Fig. 5.— Difl'erentea of soil teiuiieratiire (woods and open fields). Coniparisou of elevations above

sea level (If— O).

noticeably smaller at the higher levels. The forest cover there has a
greater cooling effect in summer and a greater warming effect in winter.
Its influence in moderating a climate is, consequently, slightly greater.
The difference in inthience of young trees and old is also of interest,
but the material is scanty. Lintzel is a station among young trees, but
there are only seven years of observation (including 1888, the latest on
hand at this writing). The curves are given in Fig. 0, and their lack
of evenness betrays the brief series of observations from which they
were drawn. The surface summer reduction seems very large, and that
at 4 feet unusually small, but these results may be due rather to the
soil at the station (Lilneburg heath) than to the trees. Dr. Wollny has

DIFFERENCES OF SOIL TEMPERATURE.

49

made some studies of the ettects on soil temperatures of vegetation at
different heights and has found that on the whole they increase with
the height.

That the forest litter plays an important part in the matter of soil
temperature can not be doubted, but no details are given on this point
in the publisjied reports of the forest meteorological services. Numer-
ous experimental observations have, however, been taken by Dr.

■ At surface.

— AtG inches below surface.

■ At ifcct below surface.

Fig. 0. — Diflferencea of temperature for young trees, Liutzel Station, -woods and open fields ( W — O).

Wollny, the results of which are graphically exhibited in Fig. 7. The
unbroiven curve is here the result of several years' observation ; the
broken ones represent shorter series, made to test the eftects in indi-
vidual soils. The comparison is always made with bare soil, and the
positive sign here means a higher tem])erature in the covered soil.
The general trend of the curves is strikingly like that in forests.

+ 1°

^

^ -

0

"^

\

y

—

/

- 1°

\

V -

•..

>

'-"'

y

/

-2"

\

V

-''^'■<

-x^

^^

»»^

/

-5°

\

\

/

/

-A°

\

/

- S"

\

-^^

/

WINTER. ,

SPRING.

SUMMER.

AUTUMN.

Average results. . — — liesvlts on loamy soil. liexvlts on quartz soil.

l''l(i. 7. — Kftects of littor on soil teuiperatun' (littered .surface — hare). ( 11'— O).

The material at hand does not ])ermit the study of other modifying
influences — .soil, latitude, prevailing cloudiness, Avater in the soil, etc.
As a means of comparison, it may be of interest to give the progress
of soil temperatures under sod as compared with that under bare soil.
This is given in Fig. 8, which represents the means from an eleven-
year series of observations by the M^Sl. l>ec(pu'iel. The trend of the
curves is quite unlike that of tlu' forest.s. 'J'he .sod is n^latively warm-
est in autumn ami coldest in s])ring, and (he mean is in favor of warmer
12i4i— No. 7 i

50

FOREST INFLUENCES.

rather than cooler temperatures. The curve is cotistructed ou the .same
scale as all the other figures of this chapter. The small May indenta-
tion is very curious and is unexplained.

+ 3"

+ 2°

^*.

'~ ■^

^^^'

^ r

^

V>-

y

"'•■

0

-

^

^

"^•-^

^^X

- 1'

s

^ ""

^-~~_

^ _.

ri.

^

-2'

~'^—

-3*

WINTER.

SPRING.

SUMMER.

AUTUMN.

Depth of 0.8 inch. Depth of 4 .G inches. Depth of 9.3 inches.

Fig. 8.— DiflVrence of soil temperature, under sod, and bare surfaee (sod — bare). Becquerel's observa-

tion."?.

From the report of Mr. Fernow for 1880, the following' tables are here
reproduced, combining results of various sets of stations in different
parts of Germany; the temperature is given in centigrade, the minus
sign denoting lower temperature iu the forest, the plus sign higher
temperatures.

Differences of temperature of the soil inside and outside of a forest.

Pebrnary-
April.'

May- July.

Aujinsl^
October.

Noveiiiber-
Jauiuuy.

Tear.

Sur-
face.

0.9 ni.

Sur-
face.

0.9 lu.

Sur-
face.

0.9 m.

Sur-
face.

0.9 in.

-0.7
—0.1
-0.1
+ 0.9

Sur-
face.

0.9 m.

Alsatian Mountains

—1.0
—1.8
—1.3
—1.3

-1 0.5

—0.8

-0.6

0

—7.8
—4.5
—4.6
—4.4

—2.8
—3.9
-4.1
—3.6

—5.7
—2.6
—2.6
—2.3

—3.2
—3.0
—3.0
—2.2

10. 3

0

+ 0.3
+ 1.3

-3.5

2. 2

-2!l
—1.6

—1.5

9 2

Bavaria (other stations)

Eastern Prussia

—2.0
1 9

The mitigating influence ou the soil temperature ajipears still more clearly when
the maximum and minimum temperatures for the year or the range of temperature
is compared.

Range of temperature.

Without
the forest.

Within
the forest.

Bavaria

Degrees.
39.5
S5
41.8

Degrees.
29 5

Alsatian Mouutains

''I

Eastern Prussia

26 7

AIR TEMrERATURES IN FORESTS. 51

TEMPERATURES OF AIR IN THE INTERIOR OF FORESTS.

We pass uow from the soil, Avliicli is fixed and tlie teiiipeiatiue changes
of which are due to the changes of the siirroimdiug. medium, to the con-
stantly moving air, the temperature changes of which are to some
extent due to the motions of the medium itself. The results for the
air from this cause are less definite, more variable, and more cai»ricions
than those derived from observations on the soil. At the surface of
the soil there is only an imperfect relation between the temperatures of
the soil and air. Yet it is undoubtedly true that the air temperatures
uear the surface are due in the first place to heat derived from the soil.
It is to the motion of the air, carrying with it its temperature, that
this disparity between soil and air temperatures is due.

The observations used here are those of maximum and minimum
temperatures.* These afford satisfactory results, because what we are
concerned with is not so much the temperatures in woods as tlie differ-
ences between temperatures in woods and those outside. Besides, in
using the extreme teniperatures, all danger is avoided of being misled
by a difference in the diurnal progress of temperatures at the two sta-
tions. The extreme temperatures, also, are those which best show the
influence of the forest, for this is exercised chiefiy here, as in soil tem-
peratures, in reducing the extremes. The data used arc always, unless
otherwise stated, the values of W— 0 (woods minus open fields reading),
and it is to be remembered that a minus sign (or position below zero
line) indicates colder temperatures in woods, a i)lus sign (or position
above zero line) warmer. The observations are taken at the same
height from the ground in the woods and outside.

ANNUAL RANGE.

Fig. 9 shows at a glance the mean annual values of the temperature
difference of W—0 in the case of maxima, minima, and the mean de-
rived from them. The heavy line across the figure represents the zero
line (for W—0=0). The annual values for 11^— O for the maxima are
always negative and they are laid off to scale below the zero line. The
annual values for W— 0 for the minima are always positive aiul they are
laid off above the zero line to tlie same scale.

It appears at once that the action of the forest varies greatly in the
different stations. It always cuts down the annual mean of the monthly
maxima, but to a very variable extent. In the case of St. Johann (five
years of observation) it cuts oft" the maxima by an average of 4o.89,
but in tlie case of Schoo it is only 1°.44, or less than one-fourth, and in
the case of the young forest of Lintzel, it cuts down the nnixima by
only 0°.35. The average amount of reduction for all stations is 2o.85.

"Dr. Miittrich lias published ii eoinploto I'cduol ion dl" tin- air tciiiperatures for the
German service \\\\ to lSi)0. It m ly he fomid in Danckelniaun's ZritschriJ'ffiir I'orst
uiid Jaydwexeii, xxii, 1890, oWi aud 397.

52

FOREST INFLUENCES.

Fritzen, - -

Kurwien, - -

Carlsberg, - -
Eberswalde,

Schmiedefeld, -
Friedrichsrode,
Sonnenberg,
Man'enthal,

Lintzel, - -

Haderslehen, -

Schoo, - - -

Lahnhof, - -

HoUerath, - -
St. Joliann,

Hagenau, - -

Neumath, - -

Melkerei, - -

Average, - -

Minimum

Mn.rimvni

I

1

I

1

1

1

}

•

1

— 1

1

1

1

1

1

1

1

Scale.

+

+
_l_

+
_J_

o

I

I

n
I

_l_

I

J

Fig. 9.— rorestairteniiieraturodiffiTMices (TF— O). German statious. i^fean annual (iross-bar), max-
ima (below zero liue), miuima (above zero liue). ami range (leugtli ot lines).

AIR TEMPERATURES IN FORESTS AND OPEN FIELDS. 53

Very much tlie same is true of tlie muiiiua temperatures. They are
always cut off in the annual mean, but by a varyiog amount. The
largest reduction of the minima is for Sonnenberg {2°M) ; the smallest
is for Melkerei (0o.74), except for the young forest of Liutzel station,
where it is only 0^A3. The average amount is lo.33.

The length of the lines opposite the names of the stations in Fig.
9 indicates the reduction of the range in temperature in the woods as
compared with the outside. This is a reduction of range, because the
forest lowers the highest temperatures of the day and raises the lowest.
This reduction varies like the preceding, of which it is the sum for each
station. For St. Johann it is 00.25; for Sonnenberg 5^^.92, while it is
only 10.80 for Hadersleben and Qo.TS for Lintzel. This is a measure
of the degree by which the air temperature is moderated in the woods
in the average. The absolute moderation is often greater, as the woods
become more efficient in reducing greater extremes, espe(;ially in low-
ering the highest temperatures. This is seen by considering the abso-
lute range for the year— that is, the number of degrees between the
highest temperature for the year and the lowest. The mean of the
absolute annual ranges at all the German stations is for the open fields
920.90, for the woods 8I0.8G, which gives a reduction of llo.04. The
mean range for all stations is 40.IS, or but a little more than a third of
this. The value of W—0 for the mean annual temperatures is one-
half of the sum of the value of W—0 for the maxima and minima.
The reduction of the minima is always less (except for Lintzel, young
woods) than the reduction of the maxima, and the combined effect is
necessarily a cooling one. On Fig. 9 the position of the mean tem-
perature compared with the zero line is the point in the middle of the
line i-epresenting the range of W — 0 for each stati<m. This is marked
by a short cross line. It will be observed that in every case, except
the one mentioned, this mark falls below the zero line, or on the cool-
ing side.

The anumnt of reduction of the mean temperature varies with the
reduction of the extremes, and is very different for different stations.
For St. Johann it is I0.7O, for Melkerei it is I0.72, but for Knrwien it
is only 00.17, while for Lintzel the mean tem])eratnre is 0o.04 higher
in the woods than out. The average reduction of mean annual tem-
perature for all the Gernmn statitms is 0o.76.

It appears, therefore, that the forest moderates (by reducing the ex-
tremes) find cools (by reducing the maxima more than the minima) the
temperature of the air tcithin it. The moderating influence is decidedly
greater than the cooling effect.

MONTHLY RANGE.

On taking up the consideration of the mean monthly values of W— 0
a very curious series of results appears. Fig. 1() illustrates the influ-
ence of the forest on the temperature of its interior air at Friedrichs-

54 FOREST INFLUENCES.

rode, but is a type of all the curves, differing- cliiefly iu the magnitude
of the changes it indicates. The line of zero differences is used as the
base line of the diagram. On it are laid off", at equal distances, sea-
sonal lines, on which the values of W—0 for the monthly miniiua are
measured off, above the zero line if plus, below if minus. The points
thus obtained are connected by the dotted line above and this, there-
fore, is the curve representing the reduction (W—O) of the mininni.
In exactly the same way are constructed the curves for the mean
monthly values for W—O for maxima, represented by the broken line,
which is, therefore, the curve for the reduction of the maxima for the
year. Between these two curves is drawn a curve with unbroken line,
which represents tlie reduction {W—O) for the mean temperatures.
This figure shows at a glance all the details concerning the influence of
the forest on its own interior air temperatures. The distance vertically
between the exterior curves is the reduction of the range. The space
between the middle curve and the zero line represents the change in ,
nu^nn temperature due to the influence of Avoods, cooling if the curve
lies below, warming if it lies above, the zero line. Where it lies below,
the space has been shaded to show phiinly tlie amount of cooling. In
this diagram, therefore, is shown the progress of all these elements
during the year. Analyzing the progress for Friedrichsrode (Fig. 10)
we see that the minima in the woods do not differ much from those out-
side between November and April. Then the forest begins to affect
them and the curve for W—O for minima rapidly rises above the zero
line. It reaches the greatest distance in August, ap]n'oaching the line
again from September to November, when the minima are once more
closely alike in the forest and outside. The forest has the greatest
effect on the minima in summer, somewhat less in May and September,
still less in April and Oc^tober. For the values of W—O for maxima
the progress is much tlie same, only in the reverse direction and two
or three times as great in magnitude. The greatest effect here falls in
July instead of August. It gives form to the W-0 values for mean
temperature, for which the greatest value falls in July. It appears also
that the temperature in the woods is almost always lower than in the
open fields. In March it is about the same; in winter it is a little cooler;
in summer it is de(;idedly cooler, especially in July.

These figures for IF— 0 are variously modilied in the various sta-
tions. Fig-. 11 (Uag-enau) shows a case where the reduction of the
minima, especially in summer, is relatively small, and the mean tem-
perature of the woods is less for the entire year. Fig. 12 (Sonnenberg)
represents a case where the red action of minima is almost as large as
that of the maxima and there is little reduction in mean tempera-
tures. In fact, the woods here are warmer in winter. In Fig. 13
(Eberswalde) the reduction of mean temperature is small and continues
during the year, like the preceding, but lor a different reason the
reduction of maxima and minima here run symmetrically, but shows no
great difference iu quantity.

AIR TEMPERATURES IN FORESTS AND OPEN FIELDS. 55

Forest air lempcraliire diffcrciian. tcooda and open fields ( /(' — <>).

+ 2°

+ 1°

--

~ "

-*"'

^^^

0

,-'-"'

^>

-■ -,^

- 1°

:^^^'"'

■^

^7^^^

^^

WW/

^^

P

^

"' ^^^^^^?

-2°

V

^^^

Kw/y/Z

^

^
^

^

-3«

\

^^

^^^

w^

/

.A —

-4°

\

/

-5°

\

/

•

-B°

\
\

/

/

-7°

\
\

X

-R*

■^^

/

-a°

WINTER.

SPRING.

SUMMER.

AUTUMN. 1

Fi,;. 10.— Fiit<lri(lisio(le.

+ 1°

-^--"■""""i

»'

0

•-*

— - •*

""

- 1"

-iu^m^

mm

'0m

^^P

^0^00^

m^^

w/m

'iMM

WM^A

'WM:

-2°

\ \ 1

t^l

^,__^

-3"

"~~

~r~~~

-"'^

~"~:--^

m/-

/

r^—

-4"

\

"<^

" ' " ' ' '

'""^^

/
/

-5°

\
\

/

-6"

^.

/
/ ■

-7°

---

-^^

/

-8"

WINTER.

SPRING.

SUMMER.

AUTUMN.

Fin. H. — Hagenau.

4-3°

■ * — ,_^

..^ —

:^ —

+ 1 ""

0

-=»55ii2^

W/m

vmM

''mm.

V/////^///,

^/m/>

'WM''^.

'/^m

y^^^^

-^

- 1°

y/>A2J^

-3°

'■•^-,,

.

--^

,^ _

-4°

^^

^

y^

-5°

^^,

—

■~-~

y

WINTER.

SPRING,

SUMMER.

AUTUMN.

Flti. 12.— Simueiiberir.

+ r

0

^ _ —

TTTTTTT,

t/777777.

77/Zl^

Ut r 1 1 J r

- 1"

- 2"

v//yyi^

''y/Z/^^^^

////////.

'ZZZZ^

^^v^

^2^:222

•vv//^

■ — — - J^

^

y

-3"

"^^

.^— T.

— -^-_.

■

-4°

-

WINTER

SPRING

SUMMER

AUTUMN. 1

.MINIMA.

MEAN.

MAXIMA.

FlU. 13.— Eberswalde.

56 FOREftT INFLUENCES.

Fig-. 14 (Sclioo) shows a small reduction of mean temperature limited
to late summer and early autumn. Fig. 1.") ( Marienthal) shows a curious
aberrant case. The woods in March and April are warmer than the
fields. But Fig. IG (Hadersleben) is perhaps the most curious of all.
Here the uiaxima of temperature are higher in the woods from Novem-
ber to Ai)ril than they are in the fields, aiul during the same time the
minima are lower. The temperature range through these months is,
therefore, greater in woods than outside. But the increase in range is
small and the mean temj^eratures are about the same in woods and
outside.

These figures, all of which, except the first, have been selected be-
cause they were peculiar, are instructive in showing how extremely
forest action varies while preserving consistently its characteristic
features. The variations are due to the causes mentioned under soil
temperatures, with the addition, in this case, of disturbances due to
prevailing winds. The average effects in which these disturbing ac-
tions are eliminated more or less completely are illustrated in Fig.
17. As this represents the result from sixteen stations, through about
two hundred years of observation, it may be considered fairly typical
for middle latitudes. From this diagram it appears that —

(1) As to miuinui, the forest reduces them with fair uniformity dur-
ing the entire year. For about half of the year they are reduced about
a degree Fahrenheit. In April this reduction begins to increase; it
grows gradually until it reaches a maximum of about 2 degrees in
Angust. From this it lessens until November, when its average value
is 1 degree.

(2) As to the maxinm, these are reduced about 1 degree in December.
This reduction grows slowly until it reaches about li degre(^s in April.
From this it grows very rapidly until it reaches abcmt 5 degrees. It
remains at about this value during the summer and decreases rapidly
during the autumn to its December value.

(3) The reduction of range is about two degrees in the winter, six or
seven degrees in the summer, and an intermediate amount in spring and
autumn. ^

(4) As to mean temperature, it is one or two degrees cooler in the
woods in summer, about the same in woods and open fields during the
winter and intermediate in si)riug and autumn.

(o) The moderating and cooling effects of woods already mentioned as
a phencmienon of mean annual temi)erature, extend, it appears, through
the entire year, but they are especially marked in summer. The cool-
ing effect tends to disappear in winter. The moderating effect is the
most important one and it is the most characteristic.

It has often been claimed that forests warm the air of their interior
in winter, but the (German observations do not show this. On the other
hand, there is an occasional warming effect in early spring, while in
winter the woods are usually slightly cooler than the open fields.

AIR TEMPERATURES IN FORESTS AND OPEN FIELDS. 57

Forest air temperature differenecc, woods and open fields ( W—O).

4- r

_ . _ ,.

L

0

^■^■.u:--.

.- — -

- 1*

— . — — -

'■^

^—

i^^^m^

mm

^^

^^^U^^

—

- 2'

'

.--"

- 3"

••^-^^— ,.

.-"^ i

WINTER.

SPRING.

SUMMER.

AUTUMN. 1

Fiii. 14.— Scboo.

+ 3'

-

..-'-'

+ r

X

^-fc^

0

rVT/ '////^

722:2!^

- r

r7Trrv7>m.

^^^■^

\

^■^^

/^

'/

- 2°

. — _

—

\

«i4^

mm^

- Z'

— \

.^

^^

- 4"

\

^

/"

- 5'

\

\

,-"

-6*

^

.^^

^^^

-—''

WINTER.

SPRING.

(^ SUMMER.

AUTUMN. 1

Fig, 15.— Maiieuthal.

+ r

^...

.__

0

^--;;^

^.^-

"-»-..

4?^

- !•

WM

^W

^^

^^

^-"^"3^

^

- Z"

\ '^

WM

^^^

^^*^

'

- 3"

\

- 4"

\

/"

- 5'

\

"^-^ __

^^^

WINTER.

SPRING.

SUMMER

AUTUMN. 1

Fig. IC— Hadeisleben.

+ ?"

+ r

"^"^-^^

,^

0

'

"■

__ — —

- r

IJWWWc

<wmYA.

//////M:,

WM^

^

^

^^P

»

^P

'/^^^^^w

- f "

~"~--

^^^

^^^^

^^

"

- 3"

N

/

/

- 4°

/

- 6°

\

^^^-^-

.y

- e°

N.

—

WINTER.

SPRING.

SUMMER.

AUTUMN. 1

MINIMA,.

MEAN.
Fig, 17. — Avcrase.

MAXIMA.

58 FOREST INFLUENCES.

M. Becquerel foiiiu] in liis observations that the woods were decidedly
cooler during persistent cold weutlier. For instance, at La Salvionniere
during nnusually cold weatlier, from the 24th to the 30th of December,
18G7, the temperature in the woods Avas 21.4° while that outside was
22.0°. From the Lst to the 12th of January, 180S, the temi)erature in
the woods was 14.5° while that outside was 15.8°. For the stations
near Chatillon-sur-Loing, at which Becquerel took his observations, the
forest temperature for the latter dates average 13.3o while the temper-
ature in open fields was 14.4°. He noted this in otlier cases and thought
that the lower the temperature in extreme cold weather the lower rel-
atively it was in the woods. This is, however, not always the case. In
Dr. Eberniayer's published series of observatioiis there is only one case
out of seven where the extreme teniperaturo was lowest in the woods
and this was only by a little over one degree. On the other hand, at
Seeshaupt, on January 25, 1808, the temperature in the fields was —20.4°
while in the woods it was (mly -f^.l'^, or a diftereuce of 22.5° in favor
of the woods. These are extreme and not mean temperatures, but the
publislied data do not permit the study of the daily mean temperatures.
Eivoli made a special study of this subject and <'amo to the conclusion
that the relative winter cold of forests was quite dependent on the direc-
tion of the wind. In Posen, 140 single observations from I^Tovember to
March gave a warmer temperature in the woods except lor south and
southwest winds, when the forest temperature was lower by a fraction
of a degree.

The ])ublished observations are not complete enough to give a good
idea of the diurnal imigress of temperatures under trees, but several
stndeuts of the subject (Becquerel, Kentzsch, Bergen) find that it is of
the same character as the annual progress. They generally represent
the daily minimum temperatures of woods as being higher tlian those
outside.

DECIDUOrS AND EVERGREEN FORESTS.

The German observations permit a study of the eifects of difi'erent
kinds of trees. The stations were classified for decidnous and ever-
green tbrests as in the section on soil temperatures, and the means of
tlie values of ir— 0 were calculated separately for each. Tlie results
are graphically represented in Figs. 18 and 19. The difference between
them is much more striking than was the case with soil temi)eratures.
The evergreen trees show a symmetrical in<'rease in the reductions of
maxima and minima until late summer, when they rapidly and sym-
metrically decrease. The greatest reduction of the maxima is 4.21°
(August), the least 1.20° (Decend)er). Tlie average reduction of max-
ima is 2.920. The gieatest reduction of minima is 2.20° for Septeml)er,
the least 1.29° in November, but it runs (capriciously from 1.29° to 1.08°
from November to May.

AIR TEMPERATURES IN FORESTS AND OPEN FIELDS. 59

Forest air tcmjurulitrc diffaencin (i/wvoils and open ftehh (W—O).

Fig. 18. — liccidiKiiis trees.

+ 2"

■^-.^

+ 1°

****■

r

- —

**,

-

0

-

- r

::::::i::iv

'm^

^M.

0m

y//'/'A'//A

WM.

^g^^^iii*^

-?."

■~-~-— -.

__

i,,uA^^

^^

^"^

-3°

"~-

—

^

^^

-4°

^

— —

—

/
^

-5°

"~*-.

— -— "'

WINTER.

SPRING.

SUMMER.

AUTUMN. 1

Fig. id. — EviTyrcfu trcos.

Fig. 20. — Elevated statiuns.

+ r

1 —

*^i

0

-'''

- 1"

iui2>^m.

■■/'///////A

wiiiiim.

iiiiiii^

^^

4y/

w

m.

mm

WM

t

'/^

y^ii^iu^

-?"

^**^

/. -

\—M

■■^y^f^U^j^

/^

-3"

\

k

/

-K''

"S; —

X

y-

y

-A"

^i

' "^ -» — -

, ___-

^^^

WINTER.

SPRING.

SUMMER.

AUTUMN. 1

i<

lO. 21.-

-Nc:ir St

a level.

0

.. ., ..

....

*^^r^

_-— '

^^..,

- 1°

r^~"

— -"

\v

. — -

"""^^

V.'/.'!M»^^

_-•■*''

- — '-^

-2°

■'

WINTER.

SPRING.

SUMMER.

AUTUMN 1

.« MINIMA.

MFAN.

. MAXIMA.

Fia. 22.— Vouii'' forest (F.iutzel).

60 FOREST INFLUENCES.

The mean animal reduction of the minima is I.660. The cooling of
the woods takes place thronghont the year, greatest (about one degree)
in summer, least (almost notliing) in December and January.

On the other hand, under the deciduous trees the time of appearance
and disai)pearance of the foliage is very plainly observable in the tem-
peratures. The year starts out with them as with the evergreen trees,
but in early si>ring theintlueuceof the forest shrinks and tends to dis-
appear. With the appearance of the leaves, in May and June, it grows
again and very rapidly, so that the greatest effect is brought on in two
months; through June and July it remains constant, but in August it
begins to shrink rapidly, and by November it has again approached the
degree of influence exhibited under the evergreen form. The greatest
effect on the maxima is observed in July, when the reduction amounts
to 6.50°, the least in April, when it is only 0.05^. The average is 2.92°,
exactly that for evergreens. The reduction of the minima is greatest
in September (2.03o), least in February (0.50°) and November (0.53°).
The average is 1.14° — half a degree less than for the evergreens. The
mean reduction of minima under deciduous trees is 0.89°, under ever-
greens 0.030. The mean reduction through the year is curiously alike
for evergreens and deciduous trees. The main difference is found in
the values of IV" — O in spring and summer. In early spring the de-
ciduous trees have a scarcely appreciable effect on the temperatures;
in summer their effect is to lower the temperature two degrees or more,
ab(mt twice that of evergreens.

As to the effect of elevation above the sea, the stations were classified
as in the section on soil temperatures and the results are represented
in Figs. 20 and 21. These figures do not differ essentially. The effects
on the elements are somewhat greater at the higher station and the
cooling caused by the forest extends through the year more evenly.
The diftereuce is, however, very slight when expressed in terms of the
reduction of the mean temperature. For elevated forests the reduction
is 0.84°, for those near sea level, O.OOo.

It will be of interest to see what is the effect of a young forest on the
temperature in it. Fig. 22 gives the results for Lintzel. The series of
observations was only seven years long and this is not long enough, it
appears, to give good even curves. It is curious to see that this curve
haidly suggests those found for mature forests. The minima are always
reduced, but the maxima are exaggerated in spring. Besides the reduc-
tion of mean temperatures has almost disappeared and is confined to
late summer and autnmn. As a matter of fact this is hardly a case of
temperatures under trees. It is rather a case of temperatures in the
foliage and corresponds to temperatures in tree tops in mature forests.

In the following table, taken from Mr. Fernow's report for 1889, the
maximum, minimum, and mean temperatures at various forest stations
are compared, the plus sign denoting higher the minus sign lower tem-
peratures than those observed in the field stations, in centigrade read-

ings:

TEMPERATURES IN TREE CROWNS AND AI?OVE TREES.

61

February-
April.

^
^

Central Italy*- - --
Eastein Francet....
Alsaiian Mountains J-

Bavaria^ :

Eastern Prussia ||

-0.8
-1.1

+0.8
-t-1.9

-0.5+0.2

'-0.7-1-0.1

0

-1-0.4

-o.y

—0.3

May-July.

—1.4

+ 1.0

+ 1.2
+ 1.9
+ 1.1
+0.5

a
a
o

3

—1.2
—1.0
— 0
—0.9
—0.4

August-
October.

a

r

3.6+1.1

2.6
-1.9
-3.2
-1.6

+ 1.3

+2.4
+1.6
+0.2

—1.3
—0.6
—0.2
-0.8
—0.7

November-
January.

3

—0.9
+0.9

0
—0.3

a

Tear.

-i-0.6— 0.11—1.9
+1.7,+1.3— 1.2

+ 1.2-f0.6
— 0. 2,-0. 2

S

+ 1.0

+2.0
+ 1.0
+ 0.2

a

-0.4

+ 0.4
0.3
0.4

* Station Vallaiabrosa, Tuscauy. t Station Bellet'ontaine, near Nancy.

; Station Melkerei, in tbe Vosges Mountains. § Stations See.sbaupt and Rohrbrunn.

II Stations Fritzen and Kurwien.

TEMPERATURES I:N^ THE TREE CROWN AND ABOVE TREES.

The observations used in the preceding- section were all taken at the
usual height, that is, about 5 feet (1.5 m). The Swiss forest observa-
tions are taken at a height of 3 meters or about 10 feet. I have not
been able to consult these observations except for a single year, that
of 1870. In Lorey's Handbook of Forestry there are given seasonal
means for twelve years from three stations, Interlaken, Pruntrnt, and
Berne. These give, in Fahrenheit degrees, for winter, ^Y — 0 0'^.92 for
spring, 1'^.33; for summer, 2°.72, and for autumn, 1°.57. These are of
the same general character as those taken at an elevation of 5 feet.
The large value of W — 0 for spring makes them resemble more the values
given in the jireceding section for stations at high elevations above
sea level. This is to be expected, as the Swiss stations are at the ele-
vation of 2,034 (020 m.), 1,476 (450 m.), and 1,946 feet, (593 m.) respec-
tively. Observations are also taken by theBavariau and German services
in the tree crown. A place is selected in the top of the tree, but under
the thatch of foliage. This is as nearly as practicable vertical over the
instruments i)laced at the surface, and observations on temperature are
here made at the same time as at the surface. The discussion of these
observations lies under two serious disadvantages. In the first place,
there is generally no station at the same elevation outside the woods,
and, in the second, the tree-top stations are necessarily at dift'erent
heights, thus making impossible a direct comparison between stations.

The first difficulty may be overcome by making' a fictitious outside
station and giving- the ground station temperatures a correction for
known mean variations of temperatnre with altitude. Unfortunately
this reduction of the surface observations to a given altitude can not be
made witliout considerable uncertainty. The change of temperature
with altitude alters with ivlinost every source of meteorological change,
with season and time of day, with the toi)Ography, the cliaractci- of the
soil and its covering, and with the weather, and the amount of change
is especially variable within the first 50 feet from the surface, within
which range must be made the correction we propose to apply. How-
ever, the correction is a small one, and over the territory occupied by
the German stations it is known with more accuracy than elsewhei'C
generally.

G2

FOREST INFLUENCES.

Average, - - - - -

St. Johann, - . - 39.

Melkerei, - - - 26.

Hollerath, - - - 2S.

Schmiedefeld, - - 31.

Sonnenberg, - - 25.

Marienthal, - - - 38.

Fritzen, . ... 26.

Hagenaii, . . - 52.

Lahnhof, - . . 38.

Kurwien, . . . 51.

Friedrkhsrode, - - 26.

Eberswalde, - - - 39.

Carlsherg, - - . 36.

Hadersleben, - - 41.

Neumath, - . - 36.

Schoo, . - . . 18.

Mhiimnm.

Mnxhmin

,

1

1

1

1

i

1

1

•

1

1

1

1

(0 (M V

r. , + + +

Scale. I I l_

(\l CO

X

_L

J

Pig. 23. — Forest teui]>eratare 4iffereuces for the year ati height of the tree top ( W—O).

TREE TOP AND OPEN FIELD TEMPERATURES. 63

The correction which is used is that given by Mr. Ferrel in his
"Recent Advances in Meteorology," and is taken from the colnnm
for "the Alps and Germany" in the table on page ISO. In no place
does this correction, as applied to the German tree-top observations,
surpass O'^.ll F., (O.OO^c) and it is always negative. After making
this correction the observations will be treated as if the li(;titious ele-
vated stations were real and the values of W — 0' (here primed to dis-
tinguish from those at the level of 5 feet) Mill be discussed as were
those of W — 0 in the preceding section.

Fig. 23 rei>resents the mean annual values of W — 0' tbr maxima and
for minima; the maxima always negative and below the zero line, the
minima always positive and above. The names of stations are placed
opposite the lines wliicli belong to them, and for each station is given
the elevation in feet above the surface of the ground. As in the pre-
ceding section, the length of the line indicates the reduction of range
of temperature in the tree top, as compared with that outside at the
same height, and the point in the middle of tlieline (marked by a short
cross bar) indicates the reduction of mean temperatures.

This diagram presents the same features as the corresponding one
for lower levels, but there is, if anything, more variation tor the indi-
vidual stations. Also, the mean temperature is here sometimes higher
in the tree tops than over open fields. This is the ca.se in five out of
sixteen stations.

It is noteworthy that tlie reduction af the annual temperature in
tree tops does not appreciably depend on the height of the station
above the ground. The lowest (Schoo) shows the smallest reduction of
temperatures. It should have the largest because nearest the ground.
On the other hand, the highest station (Hagenau) stands about mid-
way. The next highest (Kurwien) stands somewhat lower in the
series, while one of the two next highest (St. Johaun) leads the series.
Nor does any feature (reduction of maxima, minima, or mean tempera-
ture) show any distinct distribution with reference to the height of the
station.

Taking up the individual tree-top temperature diagrams we find
them strongly resend)liug those for the observations at the 5-foot
level, but with greater variation between the individual diagrams.
That for Friedrichsrode (Fig. 21) is a fairly smooth and representative
one, and it greatiy resembles those of the preceding section. The
minima are raised, the maxima lowered, the action is a cooling one,
and it is most marked in summer. These features are all visible in the
temperature diagrams at the 5-foot level. The cooling eft'ect is, how-
ever, sometimes distributed thrcjugh the year with some a])proach to
evenness.

The reduction of the minima is often relatively great (Fig. 26, St.
Johaun). It is not rare that this is greater than the reduction of the
maxima, leaving not a cooling, but a warming action on the part of the
woods.

64

FOREST INFLUENCES.
Tree-top temperature differences woods and open Jields ( Jl — 0).

_8UMMER.

Fig. 24. — rricdriclisrode.

+ 2"

|... ■- —

1

1

+ 1°

_..

.-■■—-'

-

0

,

„..

/ y f f f'}f

f .'? ?>-J

-^

- 1°

^^ii^

i^^iii^

"■>..

K//-^V'^

///////A

Y<//.'///

'///^'■//^'Z^i^^^i^

■:..'^///.:

-2°

^--.

.

y

/^

-3"

"^1

■ —

,

WINTER.

SPRING.

SUMMER.

AUTUMN.

Fig. 25.— Eberswalde.

\z'

_^

,--''

"\

+ z°

'"■"---.

--'

'-

^

+ 1°

^-^

/

0

^N

y

+ 1"

"■•^4^

W//A

^^i^

4^

W^

^^

^J

W//^

^^

+ 2°

\

^<ZZ2

i^-"

^^^"

+ 3"

N

^

y

'

- — '"

+ 4°

^^

■

^

J

+ 5°

/

+ 6"

^-.

^^^'

WINTER.

SPRING._

SUMMER.

AUTUMN.

Fig. 26.— St. Johann.

+ r

....^ ^

^^.---

--.,

. — — '

— ••**

""• '-"•.

..,.^

__ > "^

0

— ,

— 1 — ■

- r

—

—

^

— __^— ^

" "^^ —

,

_

— -'■

-£•

^-

"

WINTER.

SPRING.

SUMMER,

AUTUMN.

..MINIMA.

.MEAN.

, MAXIMA.

Fig. 27,— Carlsberg.

TREE-TOP AND OPEN FIELD TEMPERATURES. 65

Tree-top differeiias ( IT- — O).

+ 3"

.. --"'■

+ 2°

.--*"

'*'■"•- —

+ r

0

^-~-_

^

.

- 1°

.

-?"

~~

"^^^..^

^-"^

-3"

"^-.^

^^

— — .^

-^

— _,_

•

WINTER.

SPRING.

SUMMER.

AUTUMN,

FiQ. lis.— Souiienberar.

+ 2'

+ r

^^'"^

_,---"

'''-,^

_

0

' ~"

^^^

-1°

.

-— -^_

~^--.

-—

^^

^^-

_J>C

"^^^

"

^^-

,-.--'

-v^^

WINTER.

SPRING

SUMMER,

AUTUMN.

rio. 29.--Kurwien.

Fig. 30. — Hagonau.

0

1

^"^

s:= — .

-^-

i^^^^iy^

fm^-

U^

V7-777.VJ:.

- r

^7"

^ —

^-,

.-^

"-^^-"^

-z"

WINTER.

SPRING.

SUMMER.

AUTUMN. 1

Pia. 31.— Schoo.

+ r

•— ""'

V

0

^^^-^

p.^- "^

-'--TT

^"

\

■*"-.

- r

'^^A

WM

^^

W^

'W^

Z.^ li.^^

-2''

\

^ii^

-3"

\

^

— -"'''

^

-4-*

"V

-. — "

WINTER.

SPRING.

SUMMER,

AUTUMN.

....MINIMA.

12441— No. 7-

MEAN.

_MAXIMA.

Via. 3'2.— Ni'iiiiialb.

6G

FOREST INFLUENCES.

Fig. 30 is that for the .station with the greatest elevation above
the surface (Hagenau), and Fig. 31 that for the least (Schoo). They
show great differences, but there is no connection between the dia-
grams through those for the intermediate stations. They are alike,
and of especial interest, in that in both the reduction of the minima is
only a small fraction of a degree. Fig. 32 (Neumath) gives the only
case where the maxima were warmed decidedly more than the minima.
This increases the range and makes the temperature more excessive in
the late winter and early spring months.

+ z"

_...r —

'"^"•"•^

**^

+ r

0

.--*"*"'*

^-"^rr^

TZZTTT^^

- r

___^

m^

y/^y.v-V/^

b^^^

- z"

^"^

^~*^ —

- 3'

^X

/

- 4"

" V

"~^^

■

_,. ''

WINTER.

SPRING.

SUMMER.

AUTUMN.

Pig. 33. — Average dififerences of tree-top temperature, sixteen German Stations (TF — O).

The average differences of tree-top temi>eratures ( ^\'-0') are graphic-
ally given in Fig. 33. It is at once seen to be very similar to the corre-
sponding Figure (17) for the elevation of 5 feet. Tlune is the same
elevation of minima, the same lowering of maxima, the same cooling
effect, and the same exaggeration of all features in summer. The le-
duction of mininui is, however, more nearly equal to that of the maxima,

+ 2"

"^

+ r

"

>

-»■■

0

^

- 1'

^^

^^^

m

'0

•mm.

^M^

y^a^'^

- 2*

' ' ■^ —

^-v

y

/"

^

- 3'

\

N

- 4*

■^

^^*^

x"

- 5'

WINTER.

SPRING.

SUMMER.

AUTUMN.

Fig. 34.— Tree-top temperature differences for four stations with average height of 21 feet.

and there is, consequently, less cooling. It seems, therefore, that the
tree-top station is intermediate in temperatures between the base
station and that in open fields.

It will be interesting to see if there is any difference through the
year on account of elevation of the tree tops above the ground. To
ascertain this the four lowest stations were taken (two in evergreens
and two in deciduous trees) and tlieir values of W'-O' were c<mibiued.
The result is Fig. 3i. The average height was lil feet. The four

TREE-TOP AND OPEN FIELD TEMPERATURES.

67

highest (also equally divided between deciduous trees and evergreens)
were similarly treated, with the result shown in Fig. ."55. There ap-
pears no unmistakable ditfcrcnce l)etweeii them due to auything but
iudividual irrecularities.

+ 2°

+ 1°

_.*»

---.».

-•-.....

^^

"'•^^-«--. ^

0

J

-.

— --

- 1°

.^^

•--^

'A

C^^Hi^

■>^

j^

- 2"

~~~--,

*,^

^.^:v_

-3°

"-" — — —

y

V

, ^

/
/

f

- 4"

^-_

-

,•

WINTER.

SPRING.

SUMMER

AUTUMN.

Fig. 35. — Tree-top teuiperalure ditiennces lor lour stations with average lieight ot'4G feet.

The stations for deciduous trees and for evergreens were then aver-

+ 1"

+ 1°

'~..
'^-

0

- 1"

" -^ —

r:>^.^

^

"9\

^j^--^

-^

-2"

N

\

m^

/

- 3°

\

/

-4**

/
/

- s"

^v^

^

-'"

WINTER.

SPRING

SUMMER.

AUTUMN.

aged

KiG. ;!6. — Tree-top temperature dilierenees, German .statious ( W — O) — deciJuou.s trees.

separately, as for the preceding stations. Fig. ?)Ci is the diagram

+ 2°

+ l"

_,,--'

_-•-'*'

~~'--

0

- 1*

'i'/.''i\'/i ■■ ,

y/y' '////,

'-'/''/ ■y/'i

^^^ii-i^*-'

-2°

~~^-~

---^

^

^

-3*»

-^

—— ^

"——-_.

y
y
y

-4"

WINTER.

SPRING.

SUMMER.

AUTUMN.

Deciduous tics. Kvenjntu tracti. A viiitije of all.

FlQ. 37.— Tree-top temperature dilierenees, German stations (TV— (»— evergreen trees.

for deciduous trees; Fig. 'M Ibi- the evergreeus. The reductions for
evergreen trees are less in (juantity than for the deciduous, and are
distributed more evenly through the year.

68

FOREST INFLUENCES.

TEMPERATURE GRADIENTS.

So far the discussion has been as to change of temperature horizon-
tally, or the horizontal temperature gradient. The tree-top observa-
tions permit also the study of the vertical distribution of temperatures,
and this is a matter of considerable interest, because from it can be
di-awn some conclusions as to the stability of forest air; that is, as to
its tendency to originate meteorological disturbances or to take part
in those which approach it. This can be studied within the forest,
because there are two stations — one above the other; but it can not,
unfortunately, be studied outside the forest, as the station there, in the
usual meteorological services, is^ single. In order to get around the
difficulty that the stations are not at the same height, the gradients
are expanded to the uniform height of 100 feet; that is, the vertical
temperature gradients in the figure are those which would occur in for-
ests if the temperature continued to change at the same rate through
100 -feet (30.48 m.). The height of this assumed station is its height
above the ground, decreased by 5 feet (1.5 m.) (the height of the lower
station). As the irregularities in the vertical differences are exagger-

+2°

^
'^-

•>.. -ZT!;::::^

- —

■H»

_ ,,^

N_

•<'

7

s^

<,"-v

0

,^

Jl-J

N^

V--

' /'

' //

^-^

—

— '

■^^

"S;-

/<^

WINTER.

SPRING.

SUMMER.

AUTUMN.

Decichious trees. Kveigrecn tnes. Average of all.

Fig. 38.— Vertical temperature, gradient in wood.*, degrees Falirenheit for 100 feet.

ated in increasing them for 100 feet, only the data from the eight high-
est stations were used. These give a distance between lower and upper
instrument of from 31 to 48 feet, averaging 38 feet (11.0 m.). Three of
these stations were for evergreen trees (average di.stance43 feet) (13 m.)
and five deciduous (averaging 35.5 feet) (10.8 m.). These were arranged
separately.

The results appear in Fig. 38. The unbroken line represents the gra-
dients for the average of all eight stations, the broken line for the ever-
greens, and the dotted line for the deciduous trees. The vertical gra-
dients are surprisingly large, when compared with the average. In no
case would the latter be more than Oo.L'5, and it would always be nega-
tive. Here the gradients vary from OOOI F. per 100 feet in April (for
deciduous trees 0^.77, for evergreens 0^.35) to + 2^.50 in July (decidu-
ous + 20.62, evergreens + 2o,31). In summer the average gradient
under trees is about -f 2'=^; that is, it grows warmer as we ascend at the
rate of two degrees per 100 feet (31 m.). Outside in the^general aver-
age it grows colder by abont a quarter of a degree. In early spring the

TEMPERATURE ABOVE FORESTS.

69

g-raclients are somewhat alike in and out of woods. In other parts of
the year it generally grows warmer as one ascends.

A higher temperature of the air above than l)elow is called a rever
sal of the vertical gradient, and it appears that in the woods this re-
versal occurs especially in summer. It also occurs in the open air
regularly at night and often becomes very marked ou a clear summer
night, especially toward morning. The gradient is at such times gen-
erally three or four degrees and may, under exceptionally favorable cir-
cumstances, be several degrees more. The same thing is true through-
out the day in the winter montlis, but the gradient is then not gener-
ally large. The action of the forest, therefore, tends to produce a ver-
tical distiibutiou of temperature like that over snow, or over level
fields on clear sunmier nights. It should l)e noted that this arrange-
ment is in favor of stability of tlie air. The warmer air is the lighter
and is on top. Still this tree-top air is, in the warm season, usually
cooler, and conserpiently heavier than the air at the same level out
side, as is shown by Fig. 38. This is true of the entire column of for-
est air — that is, air in the forest — and this heavier air will tend to flow
out.

Fig. :i!).— Forest temiierature, (HHcrt-uce.s at Hulatte, iiiiilfr decitluous trees.

TE3IPERATURE ABOVE FORESTS.

Systematic observations above forests have been seldom taken.
Among them are those taken by M. Fautrat a few miles north of Paris,
in the forests of Ilalatte and Ermenonville. In the first the observa-
tions were under and over deciduous trees, oaks, and hornbeams, in
the second over pines. The soil at the latter place was a coarse quartz
sand, and the open field station was over a sandy plain. The surface
observations were taken at a height of 1.4 meters (4 feet 7 inches),
while those above trees were 14 meters (4() feet). The instruments
appear to have been close to the top of the pines, l)ut many feet above
the deciduous trees. The published observations are somewhat frag-
mentary, from two to ibur years being available. The: temi)erature dia-
grams near the ground are given in Figs. 39 and 40. The first is for
the deciduous forest and i.s of the familiar type. Fig. 40 exhibits some
striking peculiarities, c^hiefly in the exaggeration of the minima and

70

FOREST INFLUENCES.

the coustautly cooling mean temperature. These features may be in
part due to tlie soil. Otherwise the diagram resembles the correspond-
ing one of the preceding section. Fig. 41 gives the two forest tem-

Pio. 40.— Fure.st temperature, dittereneus at Erineuouville, iiudir iiines. Soil: coarse, bare saiid.

perature diagrams above trees from M. Fautrat's observations. The
line for mean temperature is omitted because of the narrowness of the
space. It would, in each case, hug the zero line. It is interesting to

+ 1°

DECIDUOUS TREES.

0

^,__

— . _ .

"~^

■ — T.rt

rr.^rr

-f.r

fc.-*^ ,

— ■-

EVERGREEN TREES.

+ r

0

^^^''

— — __

■— -

-^"'

~ V.

^

^^,^'

^^•««.,^

■

- 1°

—

,..--'

""^

-2°

'**'

WINTER.

SPRING.

SUMMER.

AUTUMN.

MINIMA.

.MEAN.

MAXIMA.

Fig. 41.— Forest temperature, ditiereiices above trees — from Fautrat's observations.

note how the temperature diagram, iu Fig. 41, has here contracted.
That for everg.een trees, with the instruments close to thetoj>s, is very
narrow, while that fo^ deciduous trees, instrument about 20 feet above

+ S^

_J

'^^

+ 4°

/'

— —

-^

"-^

4-3°

/
/

N

'.^"""^ .^

*\_

"^S^

\

+ 2°

/-

— — -^■"

"^^ —

^^

/
/

--"'

+ r

^:::^-

^

^..

y'

^\

"— —

0

--'■'

" — -*"

WINTER.

SPRING.

SUMMER.

AUTUMN.

. EVERGREEN TREES..

AVERAGE OF BOTH. DECIDUOUS TREES.

Fit). 42. — Vertical f iiiiperatiire •gradients finm (ibscrvations above trees.

the tops, has contracted so as to vanish at times. It is evident that if
the instruments had been put a few feet higher up it would have en-
tirely disappeared. Fig. 4"J represents the vertical gradients from M.

VERTICAL TEiMPERATUKE GRADIENTS.

71

Fautiat's observations. They are all reversed tlirougliout the season,
and tlie snininer gradients are about 4-4". Those over evergreens are
greater tlian those above the deciduous trees, or, perhaps more prop-
erl3^, the gradients decrease as we ascend above tlie top of the decidu-
ous trees.

The observations under tlie direction of Dr. Lorenz-Liburnau at Hied
were taken in the warmer months of 1888. Thej^ have been ])ub-
lislied (juite recently (1890), an<l came under the eye of the writer only
after the preceding pait had been finished. They were not systemati-
cally taken, but only at such times as seemed especially suited for the
study of especial features, and hence they throw light on some points
which the systematic observations do not sufficiently care for. One of
these is the change in temperature above and below the crown of foli-
age for the same station and time. Dr. Liburuau gives the following
mean results from many individual observations. The plus sign indi-
cates an increase of temperature upward, the minus a decrease.

Morning hours

About noon

A fternoon

Early night

Late night and early morning

Change-

From 10.4 to

From 30.1 to

30.1 feet.

50.8 feet.

°F. °C.

°F. °C.

+ 1.08 .00

+3.24 1.80

4-1.57 .87

+2.43 1.35

+ 1.40 .77

+2.10 1.21)

+ .18 .10

- .11 .00

— . 63 .35

— . 27 .15

The station at 3G.1 feet was in the crown, that at 50.8 feet above it.
The distances are not great enough to get good values for the vertical
gradients, but it is evident that these are very large, and greater
above the crown than below.

These observations were all taken with a clear sky and calm air.
Some interesting figures are also given for the differences between
field and forest temperatures when the air is still and when it is in mo-
tion.

The following table gives the values of W— 0 under these circum-
stances :

Morning bour.s:

Calm

Windy

Alioiit noon :

ralm

Wimly

Win.ly

Afternoon :

("aim

Wiuilv

Under trees,

10. 4 feet.

(5 m.)

First and lastnight hours (calm) .

-1.37
1.19

c J" c 0

-2.11 1.17

-1. 24 . 70

-2.47 1.37

-1.62 .90

-2.02 1.12

.76
.66

-2.50 1.42

Tree crown,

36.1 feet.

(11 m.)

°F. °0.

—1.53 .84

+ 1.1,7 .04

— . 54 .30

+ .02 .01

—1.13 .63

- .54
-1. 12

.30
.62

-3.00 1.70

Above trees,

50.08 feet.

(15. 5 m.)

+2.16 1.20
+ 2.29 1.27

+ 2.29
+ 2.54
1^1.30

+ 1.03
—1.28

1.27
1.41

.57
.71

-3.19 1.77

72 FOREST INFLUENCES.

Undei- the crowns the temperature is lower in the forest, but above
the crowns it is usually higher. The wind reduces the dilterence under
and in the crowns, but makes little difference above them.

Putting- the preceding- results together, it appears that the forests
reverse the vertical gradient, and that the gradient grows larger as the
surface of the forest is approached. Toward the ground the change is
slow; upward it is very rapid, so rapid that the special cooling effect
of the forest must disappear at no great height. The surface of the
surface of the forest is, meteorologically, much like tlie surface of the
meadow or cornfield. The isothermal surface above it in sunshine is a
surface of maximum temperature, as is the surftice of the meadow or
cornfield. From this surface the temperature decreases in both direc-
tions.

TREE TEMPERATURES.

The disturbances of temperature under trees, so far discussed, would
be the same in kind if the surface covering called a forest were inor-
ganic. They are due to the screen formed by the forest against the
sun's rays and against the radiation from the soil to the sky. The
kind of effect would be the same if the forest were dead, but the quantity
might be quite different. The shade under a dead tree is much less
than under a live one. This is due to a considerable extent to the fact
that its leaves are uo longer spread across the path of the sun's rays,
and the heat of the latter is not so much absorbed by the evaporation
of water. These, again, may be purely mechanical effects, and it is still
left uncertain whether the living tree affects the temperature through
its vital action ; whether in the processes of life the tree does not abstract
enough heat from the air, or add enough heat to the air, to be meteor-
ologically sensible. There is abundant reason to think that it may exert
such action, because all vital processes result in the absorption or radia-
tion of heat— and so do purely inorganic chemical processes. The tree,
as an organic being, is a source as well as an absorbent of heat; the
only question is whether the heat it absorbs or emits is enough in
quantity to make a sensible difference in the temperature of the air.
The analogy of the hot-blooded animal leads one to look first to the
tree itself for a solution of this question. Even the cold-blooded ani-
mal, when of a large size, has its own temperature slightly higher than
the mean temperature of the medium in which it lives. The tree is a
living body. Is the same true of it"?

The most elaborate series of observations of tree temperatures are
those taken at Geneva for the five years, 1796 to 1800. The thermom-
eter was introduced into the trunk ot a horse-chestnut tree to the depth
of 0 inches, or about half of the way to the heart. The temperatures
were read three times daily (sunrise, 2 p. m., and sunset), and observa-
tions were also taken at the same hours in the air and at several depths

TREE TEMPERATURES.

73

in the soil. In ¥\ix. 4.) is given ii grapliical representation of the mean
monthly temperature, for the tree (unbroken line), for the air (broken
line), and for the soil aib the depth of 4:^ feet (dotted line) The straight
lines across the diagram horizontally represent the mean annual tem-
peratures for these three positions respectively. It appears that the
mean temperature of the tree is lower than that of the air (by 0o.3),
and still more so than that of the soil at the depth of 4 feet ())y — l^.T).
The results are not easy to interpret. A part of the difference is
probably due to a retardation of diurnal nuixima and minima in the
tree. M. Becquerel found this retardation very marked, but his obser-
vations were of such a kind that they can not be satisfactorily tabu-
lated. He found that in small branches the summer diurnal maximum
was retarded by two or three hours, while for larger ones it might be

50°

-^

^,-.

40°

y^^

'y^

\

\

^■\

30°

'

i

/.

\N

20''

--^

■j

^\"

_

i6°

T^

//
//

#

\"''

IE"

*

^

8"

yi

-5^'

4'

y^

0

WINTER

SPRING

SUMMER

AUTUMN.

In a horse-chestnut. In the air. -/;( soil four and a quarter feet below surface.

¥m. 43.— Tree temperature — Geneva ob.servatious.

twice this. This being the case, the daily observations would not fall
at the same i)arts of the daily curve of temperature, and whether the
correction to be api»lied to bring them to a comparative condition
should be i»ositive, negative, or zero would depeiul on the hours used
and on the retiirdation in the tree. The differences shown in Fig. 43
are, therefore, not necessarily the true differences, and with the proper
corrections many of the irregularities of the curve would ])robably dis-
appear. The curve, however, indicates that the tree is relatively
warmer in early spring and late sumn^er and the opposite between
these dates, and also that Ihe sunimei' maximum is retarded as com-
X)ared with that of the air.

The trunk of the tree is so constructed as to be a very poor conduc-

74

FOREST INFLUENCES.

tor of lieat. The wood itself is a bad conductor; the tissue is filled
witli air or sap, both bad eouductors; and it is protected by a layer of
bark Avhich is an especially poor conductor. All these would tend to
prevent the access of beat to the interior of the tree and to prevent its
exit when there.

These conditions would explain the retardation of the maximum in
trees but would not explain the lower mean tem[)erature nor relatively
low temperature of winter. 8ome lij^ht may be thrown on their con-
dition l)y the values of the differences of temperature between tree
and air surrounding it [T—W). These observations were not made in
the German service but Avere in tlie Bavarian. It is necessary, there-
fore, to depend here on the nuich shorter series of published observa-
tions of the latter service. Observations of the temperature of the tree
were taken at the base and in the crown. The thernuuneter was al-
ways iutro<luced to the heart. For the stations at the l>ase this dis-
tance varied from 4 inches to 11.2 inches averaging 6.3 Paris inches

0

r

N

y"

0^"^

'~-^-v

v

N

\^^

\

\

— ^ar T.

^y-

—f

— \ —

\
\

3"

\
\

V

V _

\
\
\

__v_-

•

;."_'_

/

^

4"

•

"-. -•-'

•
•

WINTER.

SPRING.

SUMMER

AUTUMN.

In tree crown. Average depth 4.3 inches. In tree base. Average depth 6.3 inches.

Fig. 44.— Tenii)eratuie diflVreuces between tree aufl air.

(170 mm.). The observations iu the crown were made at an average
height of 31.4 feet (varying from 20 to 40). The depth of fhe ther-
mometers varied from 2.7 inches (73 mm.), to 8.5 inches (230 mm.), aver-
aging 4.3 Paris inches (110 mm.). The number of trees on which ob-
servations were made through the year was eight— five evergreen, and
three deciduous. The results are represented in Fig. 44, the broken
line showing the tree crown observations, and the dotted line those
near the base. The straight lines, parallel to the zero line, represent
the corresponding mean values. While in Fig. 43 the mean tempera-
tures are used, in Fig. 44 the monthly values of T—Wave shown.

In the arrangement of each one of the curved lines in Fig. 44 about
its mean value, the corresponding straight line, there is nothing but
what might be expected in observations iu the interior of any inorganic
mass, such as a rock, a mass of masonry, or a dead tree. In fact,
except for certain irregularities which are probably due to the shortness
of the series of observations, the curves are much like the value of

TEMPERATURE DIFFERENCES BETWEEN TREES AND AIR. 75

]\—() lor soil t(Mni)eratiives. Tlie tree, like tlie soil, is warmer in
wiiitiT and cooler in siunnier llian tlie mean value for W—0. It is
also noteworthy that for the smaller trunks and braucLes the varia-
tions from air temperature, aud the variations from its own mean, are
smaller, exactly as would be the case with any inorganic body.

In short, the variations of temperatures correspond to the varia-
tions of temperature in inorcanic bodies of similar character and
similarly i^laced. In fact, Bravais, in winter observations at Bossekoj),
La])laud, fouud the temperatures in a live and dead tree to be practi-
cally the same.

It only remains to discuss the average lower temperature in trees.
In Dr. Ebermayer's tables, this reduction of temperature at the base
of trees as compared with air is 2^.2, while at the crown, with an aver-
age diameter of limb two-thirds as great, it was 1°.2. This reduction
may be due to the temperatures of the water taken from the ground.

When the sap tiows most freely— iu spring and early summer— the soil
is cooler than the air aud the water drawn from the soil must have its
temperature. The tree is so thoroughly protected from the conduction
of heat that this temperature must change but slowly as the sap rises.
It must give a lower temperature to the interior of the tree and this
lowers the annual mean. At the same time, for reasons to be given in
the next section, the return sap is probably also cooled to some degree,
so that all the water circulating in the tree is, in the warm season at
least, cooler than the air.

The foliage seems especially arranged for the exchange of heat. Its
surface is very large for its mass, and it is so exposed to the open air,
to the sky, and to the sun's rays that it is adapted to receive and part
with heat easily and quickly. The temperature of leaves considered
as purely inorganic bodies should, therefore, be lower than that of
the air at minimum, and higher at maximum. There seem to be no
systemati(; observations on the temperatures of foliage, but botanists
have taken a good many occasional observations which confirm the
statement just made. The temi)erature of the leaves falls decidedly
lower than that of the air on ch^ar summer nights, the ditt'ereuce being
cenerallv several degrees and sometimes readiing eight or ten. On
the other hand this temperature rises in the sunshine to several, and
occasionally to many degrees of temperature higher than the air
around; in other words the maximum temperature of leaves iu sun-
shine is higher than that of the air, while the minimum temi)erature iu
clear nights is lower than that of the air, A leafy branch with the cut
end in water and exposed to the sunshine has been found to be cooler
than one not in water, suggestiug that the average teuiperature of
foliage is on the whole lower than that of the air about it.

76 FOREST INFLUENCES.

VEGETATIOIS^ AIS^D AIR TEMPERATURES.

It appears from the preceding section that the tree shows, of itself
by its own action, no sign of warming the air and but little sign of cool-
ing it. Systematic observations show this to be true of the framework
of tlie tree, but no observations have been taken for the foliage and
otlier exterior parts — the parts where most of the vital activity of the
tree takes place. The great volume of chemical and physical changes
which are produced here and which result in the growth of the tree and
the formation of its fruits make it an open question (not yet decided by
direct observations, which are not easily instituted) whether the tree
does not affect temperatures sensibly. So far as any observations go—
those in the tree-crown for instance— they show little clear evidence
of such a meteorological effect, but it w ill be of interest to see what can
be concluded theoretically. The changes which take place at the leaf-
surface, due to vital activity, are produced by the trees in common with
the other and lower forms of vegetation and consequently the problem
to be solved is double, namely: Do the vital processes of vegetation,
in which heat is ntilized or given off, produce a sensible effect on air
temperatures, and is the effect, if any, different for trees and for mead-
ows, mixed growths, or crops?

Vegetation has a complicated relation to heat in the various organic
processes. In the transpiration of water it utilizes large quantities of
heat, changing it from that sensible in temperature to the work of sus-
taining the w^ater in a condition of vapor. This latent heat becomes
again sensible when the water is condensed, but this may occur at a
distance from the place wiiere the w^ater is taken up. By unlocking
the combination of carbon and oxygen in carbonic acid, using the car-
bon and rejectin'g the oxygen, it reverses the action of combustion and
so takes up heat. On the other hand, in the various processes in which
oxygen is combined, it performs a process analogous to combustion and
gives out heat. This process is sometimes so active (as in the germina-
tion of some seeds and the flowering of some plants) that the tempera-
ture of the parts is raised several degrees above that of the outside air,
and it is sensibly warm to the hand. There are other more complicated
processes going on in the plant, the relations of which to heat can not
be foretold.

The quantity of heat used in these processes varies greatly with the
season, the temperature, the condition of the plant, and so on. It can
be ascertained only approximately even in the processes best under-
stood. Yet the problem has so many features of interest that even
such an approximate solution is desirable. For instance, the German
forester Ney has attemi)ted to show^ that the unseasonable frosts of
mid-May in central Europe are due to the amounts of heat absorbed i
by j)lants at that season. I

The transpiration of water by the plant is a vital process which ;
needs the stimulus of the sun's rays; it takes place in the green parts i

TRANSPIRATION.

77

only, and is especially active when these parts are young. It is tlie
process by which the plant gets rid of the surplus water after having
drawn it from the soil in order to extract from it the nutriment which
is present in only a very highly attenuated solution. Botanists have
made many measurements of its ainount and their results are ex-
tremely varied, due partly to the fact that this function varies much
naturally, and still more i)erhaps to the fact that the experiments are
generally made under conditions which are not natural to the plant.
Sachs says that it is no rarity for a tolerably vigorous tobacco plant
at the time of flowering, or a sun flower of the height of a man, or a
gourd plant with from fifteen to twenty large leaves, to transpire from
one to two pints of water on a warm July day; and, so far as may be
judged by the use of branches with the cut end in water, it may be
believed that large fruit trees, oaks or i^oplars absorb, transport
through their stems, and transpire from the leaves, ten to twenty or
more gallons of water daily. He also quotes Haberlandt to show that
the amount transpired by a stalk of Indian corn, in its season of 173
days, is 3 gallons; by a stalk of hemp, in 140 days, is 6 gallons; by a
sun flower plant, in 140 days, is 14^ gallons.

It is not generally practicable to comi)are the transpiration with
known meteorological phenomena, such as evaporation from a water
surface, or from the soil, or the precipitation, but some such compari-
sons have been made. For instance, comparing the leaf surface to an
equivalent Avater surface, Uuger makes transportation from the former
0.33 of the evaporation from the latter; Sachs for white poplar 0.36, for
the sun flower 0.42. Comparisons have also been made between the
transpiration from plants and the evaporation from the surface over
which the plants stand. Schleiden thought that the transpiration
from the forest was three times that of a water surface equal to the
territory covered by the forest. Schiibler thouglit it only a quarter;
and Pfafl', who studied a solitary oak in a garden, found that it varied
from 0.87 to 1.58. Comparing the transpiration of plants with the
evaporation from the bare soil which would be covered by them, Hartig
thought the transpiration of a forest less, Schiibler found it 0.0 for the
forest and 3.0 to 5.0 for sod. Marie-Davy found it, for firs 1.18, for
beeches 1.32, for sod 1.80. As to the influence of sunlight. Vines
quotes from Wiesner, who carried on his observations with special
precautious to prevent influence from other stimuli of the transpira-
tion. The results are given in the following small table, in which the
evaporation in full sunlight is taken as unity.

riant.

Transpira-
tion in sun.

In (lilTused
daylight.

III dark-
nosh.

Tho common broom

1.00
1.00
1 . 00
1.00

0.40
0. T,2
0.40
0 l-'

0. ;i7

0. :i3

litishy iiuilldw .

0. Xi

Ji)*li:tn com - - - -^ -

0.15

Means .

1.00 0. 30

0.27

78

FOEEST INFLUENCES.

From this it appears that it is the direct sun's rays tliat most pro-
mote trauspiratiou. When ill dift'iised light the phiiit transpires only
a third as much, and in darkness only a quarter. The application of
this to forests is evident, because there at any time of day a considera-
ble part of the foliage is in shade.

The estimates of transpiration are very numerous and it is not easy
to get from them an estimate of the amount for plants in terms of
evaporation or precipitation. In the following table those have been
selected which were most easily expressed in terms of quantities ob-
served meteorologically. The duration of the active season was taken
into account in each case.

Observer.

Hales .

Schleiden .

Vosel .

Hartis

Piafl'.':

Hriliiiel

Scliiibler

Haberlandt .

Plant.

Transpiration expressed as
rainfall equivalent in inches.

Daily.

Annual.

Inches.

Sunflower O.VS

1 Cabbage ! 0.12

j [Grapevine ! 0.03

[iHop 0.05

.jClover itoats j 0.11

( 4 vears beeches ] 0. OO:!

I 'lycars tirs 0.002

■) Wheat field ' 0. 62

I iBarley field 0.56

! 24 year.s mixed forest .... 0. 021

Oak

115 years beeches

Low spear grass

Oats ^

Barley

0.21
0.05
0.08
0.13
0.07

Mm.
3.30
3. 05

.76
1.27
2.79

.076

.051
15.75
14. 22

.533
5.33
1.27
2.03
3.30
1.78

Inchefi

1.5.2

14.4

4

5.6
13.6
0.43
0.28
4.4
3.9
3.8
193.2 (?)
9.2
14
9.1
4.9

Mm
381
361)
101
142
345

11

7

112

99

Bti.5

233
355
231
124

Authority.

Eberniayer.

Do.

Do.

Do.

Do.

Do.

Do.

Do.

Do.

Do.
Austrian Met. Journal.
Sachs.
Duchartre.

Do.

Do.

The best that can be done with these very variable measures is pro-
bably to take a mean of the values of Hartig and llohnel. This gives
the transpiration of the Ibrest as G. 5 inches (105 m.) for central Europe,
or about one quarter of the precipitation. That this is not much too
large is indicated by Pfafl's results, which appear to be excessive,
and that it is not much too small is indicated by those of Vogel.

The most elaborate investigations on transpiration of forest trees were
made by F. B. Hohnel, and since a discussion of these appears in the
report of the chief of the forestry division (Mr. B. E, FernoAv) for 1889,
I quote his language on this subject:

The quantity of water so used is as varial)le as the amoimt of precipitation and in
fact within certain limits dejyeuds largelj- upon it. That is to say, a plant will tran-
spire in proportion to the amount of water which is at its disposal. Transpiration is
also dependent on the stage of development of the plant, on the nature of its leaves
and amount of its foliage, on temperature, humidity, and circulation of the air, on
intensity of the sunlight, and on temperature and structure of the soil and on other
meteorological conditions. Kain and dew reduce transpiration, wind increases it.
The amount of transpiration depends considerably upon tiie thickness of the leaves,
therefore the surface of the foliage is not a reliable measure, but it should be com-
pared with the weight. '

With so many factors to vary them the values which may be given for the amount
of transpiration of various kinds of trees can only be appro.xinuitious of its range

TRANSPIRATION. 79

within wide limits. All the figures which have been published, based upon calcu-
liitions or experiments in the laboratory, are useless for jiractical purposes. Espe-
cially do those figures which represenb (he rer|nirementof the plant as exceediug the
amount of precipitatious, exhibit ou simple reHection, their absurdity.
' If the requircmeut per acre is cousidered, the density of the growth of plants must
also be taken into account.

The first careful and comprehensive investigations into the water requirements of
forest trees Avere made by the Austrian furest expt-rimeut stations in 1878 (F. B.
Hohnel), and full tables of the results obtained can be found in the records of those
stations.

An average of the many figures there presented would make the amount of tran-
spiration per 100 grams of dry weight of leaves in conifers 4,778 to 4,990 grams oT
water, in deciduous-leaved trees 44,472 to 49,553 grams of water. That is to say,
the deciduous trees transpired about ten times as much as the conifers, and compar-
ing the two extremes of transpiration, the deciduous tree with the highest rate of
transpiration utilized twenty three times more water than the coniferous tree with
the lowest rate. Ash, birch, and linden were found to be the most vigorous tran-
spirers, oaks and maples transpiring much less. Curiously enough, while in the
conifers shade reduced the transpiration considerably, in the deciduous trees it had
the opposite effect.

During the period of vegetation the following varieties transpired per pound dry

weight of leaves :

I'oiiiids
of watf r.

Birch and Linden 600-700

Ash 500-600

Beech 450-500

Maple 400-450

Oaks 200-300

iSpruce and Scotch Pine 50-70

Fir 30-40

Black Pine 30-40

The next season, which was more favorable to transpiration, the amounts were
larger; the deciduous trees transpiring from 500 to 1,000, the coniferous from 75 to
200 pounds, or in the projiortion of one to six.

The following actual amounts transpired per 100 grams of dry leaves during the
third season (1880), will show the relative position of the various species (European) :

Kilograms. | Kiloiiranis.

Ash 101, 850 Scotch Pine 12, 105

Birch 91,800 Fir 9,380

Beecli 91.380 Austrian Pine 7,005

Hornbeam 87,170 I Asjien 95,970

Elm 82,280 j Alder 93,300

Maple (J. crtWjfjr.sOT) 70,380 {Linden 88,340

Norway Maple (J . plafa- | Larch 125, 600

7ioidcs) 61, 180

Oak {(,>. robitr) 69,150 I Average deciduous trees .. 82,520

Oak ( Q. Cerrh) 49, 220 Average conifers 11, 307

Norway Spruce 14,020 i

The variability of transpiration from day to day is of Avide range; a birch stand-
ing in the open and found to have';iO0,O0() leaves was calculated to have transpired
on hot summer days 70i) to 901) pounds, while on other days its exhalations were
probably not mon; than 18 to 20 pounds.

A fifty to sixty y(!ar old beech was found to have 35,000 leaves, with a dry w(Mght
of 9.86 pounds; a iranspiration at the rate of 400 pounds per pound during the period
of vegetation would make the total transpiration 3,944 pounds per tree (about 22
pounds daily) ; and since 500 such trees may stand on 1 acre, the transpiration pet

80

FOREST INFLUENCES.

acre would amount to 1,972,000 pounds, while the precipitation during the same
period would be 2,700,000 pounds.

The transpiration of a thirty-hve-year-old beech with thinner leaves, pf which
there were 3,000, with a dry weight of 0.79 pounds, would under the same conditions
transpire 470 pounds per 1 pound of foliage, or 373 pounds per tree (about 2^ pounds
per day from .June to November) ; and since about 1,600 such might be found on an
acre, the total transpiration might amount to 593,800 pounds per acre, or consider-
ably less than the amount of rain-fall.

Calculated for summer months during .Tune, July, and August alone, the require-
ment of the two beech growths was 20,000 and .5,000 pounds per day an acre respec-
tively. Conifers, as was stated, transpire one-sixth to one-tenth of the amount
which is needed by deciduous trees.

The amounts transpired by agricultural crops and other low vegetation, weeds,
etc., have been found to be consideraljly larger, as will be seen from the results of
the latest investigations by Wollny, which I have calculated per acre to make them
comparable with the foregoing results :

Agricultural crops.

Winter rye

Barley

Peas

Red clover (first season) ...

Summer rye

Oats

Beaus

Red clover (second season)

Time of vegetation.

Apr. 20- Aug.
do

do ....:.

Apr. 20-Oct.
Apr. 20- Aug.
Apr. 20^Sept.
Apr. 2U-Sept.
Apr. 20-Oct.

3, 1879

1,1879

14,1880

14, 1880

10, 188J

1, 1880

Water con-
sumption
per acre.

Pounds.
2, 590, 186

2. 720, 2.38
.3,144,128

3, 070, 012
3, 000, 486
3, 422, 584

3, 1.30, 2.33

4, 109, 198

I repeat again that these figures can only be very rough approximations denoting
maxima of transpiration, and that the amounts transpired per acre depend largely
on the amf)Unts furnished by precipitation. Therefore our forest areas within the
arid region of the country- probably transpire a minimum of water, their scattered
growth and their coniferous composition, with the scanty rain-fall, reducing the
amounts to lowest limit.

Taking a rain-fall of 20 inches, which represents say 4,-500,000 pounds of water per
acre, a coniferous forest, assumed to trausitire one-sixth of the amount found for the
older beech-forest under most faA'orable conditions of precipitation, would require
hardly more than 330,000 pounds (presuming the same weight of foliage), or not 8
per cent of the total precipitation. To be sure, this amount must be available dur-
ing the period of vegetation.

THEORETICAL CALCULATION OF HEAT ABSORPTION.*

There is another way in which the average amount of transpiration
can be ai)proximated, perhaps more closely than by the method of direct
measurement. The water transpired comes mostly, if not entirely, from
the soil. On evaporating it leaves behind it the matters held in solution,
a portion of which is inorganic and is that which constitutes the ash of
the plant. Assuming that the absorption by the plant is not sensibly
selective in the average from a large number of individuals, then a
knowledge of the annual addition of inorganic material, i. e., ash, to the

* The computations in this section are made on the decimal system because of its
greater simiilicity.

HEAT ABSORPTIOIT.

81

forest, and the amouiit of mineral matter held in solution by the ground-
water will enable us to comijute the amount of water which has been
evaporated. The required data, particularly the latter, are often known
with some exactness. Taking, with Grandeau, the average amount of
of ash in the plant at 5 per cent and the quantity of mineral matter in
solution as two parts in ten thousand, the plant must transpire five
thousand times the weight of its ash or two hundred and fifty times its
own direct weight. Grandeau gives 6,497 kilograms as the annual pro-
duction of a beech forest (wood and foliage), 0,442 as that of Norway
spruce, and G,420 as that of pines, for every hectare covered. These
numbers multiplied by 250 would give the quantity of water transpired,
which, reduced to thickness of the sheet of water over a hectare, gives
the depth or rainfall equivalent. In these cases it gives 0.4, 0.3, and
0.3 inches, respectively. The numbers given by Grandeau are for cen-
tral Europe, as were the preceding. The results by the two methods
are 0.5 and 0.3 inches, which, by chance, are remarkably close to each
other. Knowing the amount of water transpired, and the temperature
at which transpiration takes place, it is easy to get the amount of heat
used up in the process. The evaporation of any given weight of water

. , . . . ... ■ 000.5 — 0.095^ ^.

would heat by 1° C. a weight of air which is times as

great, where t is the centigrade temperature of the water evai)oratcd.

As water is 773 times as dense as air at the standard temperature (32'^

F.) and standard pressure (30 inches), the evaporation of alayer of Avater

an inch thick takes up as much heat as would warm by 1° G. a layer of

, . , . , . C00.5 — 0.095 t ^^^
air of standard density and ot a thickness oi _ X < iS=

(006.5—0.095^) 3255 inches. For different temperatures the thickness of
the layer of a homogeneous atmosphere which might be cooled 1 degree
for each inch of the layer of water evaporated is as follows :

Tenii>era-
tiire.

Thickness
of air layer.

°C. '^F.

0 32

10 50

20 68

2liles.
31. 1
30.8
30.4

The cooling equivalent to the annual forest growth, is therefore
about 0.4 times that exi)ressed in this table and takes place through the
entire season, but is greatest in late spring in sunshine, least in dark-
ness in late summer. It acts, however, continually, and when the
enormous thickness of the air layer is divided by 150 days (about the
length of the active season) and this by the number of seconds in a day,
the result per secon«l does not appear so very large. As a convenient
general expression it may be said that the evaporation (►f any depth of
water would take up enough heat to cool by one degree (Fahrenheit in

12444— No. 7-

-G

82

FOREST INFLUENCES.

tliis case) one million times that depth or volume of air, the latter being
of average temperature and density.

In calculating the effect of the breaking up of the carbonic acid, we
may assume that as much heat is used as would be given out by burn-
ing the same quantity of carbon in oxygen— that is, by remaking the
compound. Grandeau gives as the annual product of carbon by
plants, per hectare, 3,000 kilograms for forests, 1,500 to 4,500 kilograms
for open fields, and, exceptionally, 15,000 kilograms for a field of giant
maize. This is probably all derived from the carbon dioxide of the air.
The burning of a kilogram of carbon produces 7,900 large calories of
heat according to Andrews, or 8,080 according to Favre and Silber-
mann.

The large calorie is the amount of heat necesssry to raise one kilo-
gram of Avater 1^ C. in temperature; it would warm to the same amount
1 -^ 0.237 = 4.22 kilograms, or 4.22 x 773.3 = 32G3.3 liters of standard
air. The burning of the annual carbon productions above given would
warm by 1° C. the thickness of air mentioned in the following table, in
Avhich 7,900 is used as the number of calories produced by the combus-
tion of 1 kilogram of carbon :

Forest.
Prairie
Maize .

Ciilorics

used per au-

iiuiii ])cr

hectare.

23,7 X 10 «

<, 11. K X 106

i 35. 5 X 10«

118.5 X 106

Liters of air
cooled.

77.2 X IQS

38. 6 X 10'

115.8 X 109

380.0 X 109

Thickness
of air
cooled.

Miles.

5.0

2.5

7.5

25.0

It appears that tlie heat absorbed by the entire carbon assimilation
for a forest is only from one-fifth to one-sixth of that used in evaporat-
ing a layer of water an inch deep. As the latter is only a sixth or a
seventh of the transpiration, it appears that the cooling which would
be caused by the breaking up of carbonic acid, in order to form forest
growth, is only from one-thirtieth to one-fortieth of that caused by the
transpiration.

It is highly probable that the heat used in the former process is more
or less closely made up by the heat produced in the oxidations that go
on in the plant. In any case the cooling due to this source may be
disregarded and only that of the transpiration be considered. It is es-
timated that the sun's rays, when the sun is vertical, pour out twenty-
five calories per minute over each square meter of surface exposed ver-
tically to it. Of this about 30 per cent is absorbed in passing through
the atmosphere, when the sun is vertical and the sky clear, and very
much more under other circumstances. The 17.5 calories, for a dear
day and vertical sun, would in a single minute evaporate a layer of
water 0.0011 inch thick. But the transpiration is about 0.5 inches
for the season of about 150 days, or 0.00003 per minute. Thus the

TEMPERATUKE IN WOODS, GLADES, AND PLAINS. 83

sun's rays under these favorable circumstaTices pour on the surface ^^^
times as much heat as is required for evaporation. If the maxi-
mum heat of sunshine is reduced by one-half for latitude, by seven-
eighths because of night and of low morning and afternoon tempera-
tures, and again by one-half because of cloudy weather, making its av-
erage value one-thirty-second that given above, yet there is more than
enough over each square foot of forest to effect the transpiration.

It is only where the transpiration has been very active during the day
and contiiuies into the night that its cooling could be possibly injurious,
but as the temperature cools the transpiration itself is checked, and
besides the moisture which it has poured into the atmosphere serves as
a screen to prevent rapid radiation from the soil, and its condensation
returns some of the heat that had been taken up. There might, however,
occur a combination of circumstances, very rapid transi)iration during the
day absorbing the sun's heat, a clear night, calm air and a low temi)er-
ature from other sources coming on Avith the evening, when the trans-
piration might cause the temperature to fall below freezing. This could
hardly happen, however, without a meteorological change toward cold,
and this change must come on toward evening, for otherwise it would
itself check the transpiration.

The combination would be more likely to occur over herbaceous
growths, especially over grass and cereal crops, than over forests.
Their transpiration, it appears from the table (pnge — ), is decidedly
greater than that of the forests, and they make rapid advances much
earlier in the spring when such low temperatures are otherwise possi-
ble.

TEMPERATURES IN^ WOODS, GliADES, AlVD PLAINS.

The foliage of the trees reflects a considerable part of the solar rays
which reach it, and this heat is reflected in all directions. That part
which passes toward the sky is probably lost, and plays only a small
l)art in warming the air. The part which is reflected longitudinally or
at downward angles has a very favorable path for absorption by the
air, more favorable in fact than that of the noonday direct rays of the
sun. The absorption of this heat should occur, for the most part, in
the vicinity of the forest. The temperature around the forest ought,
therefore, to reach a somewhat higher maximum on sunny summer days
than that of the air at some distance from the forest.

On the other hand, the surface at the stations near the forest is ex-
posed to nocturnal radiation of heat to the sky about as freely as is the
station at a distance, and, as the former is often in the lee of the forest,
its air is generally more stagnant. Tliis ])romotes cooling of objects at
the surface of the earth, and as the station near the forest is as much
exi)osed to celestial radiation and the relative stagnancy of the air
favors this cooling, and also favors the communication of it to the ad-
jacent air, the temperature at such stations should often fall lower than

84 FOREST INFLUENCES.

over a plain distant from woods. The forest-field or forest-glade sta-
tion is, therefore, not exactly like a station in an open plain. It is
likely to average a little warmer in the day and a little cooler at night,
giving it a greater daily amplitude not only than that under trees, but
greater, too, than that in open prairies. Whether the mean temper-
ature would be diflterent in the forest-field station would depend on the
relative increase in maximum and minimum. If the increase in the
former is greater, the forest-field station temperature would be higher
than that of the prairie station j if the minima exaggeration is greater,
it would be the opposite.

The phenomena mentioned above are especially noteworthy in forest
glades. The horizontal reflection from the foliage comes there from all
sides, and the heat rays not absorbed are reflected from the opposite
side, a part to the sky, a part to the opposite side, and this reverbera-
tion of the heat makes such places oppressively hot in the early after-
noon. This is well known to tlie hunter and to the pioneer with his
little clearing in the deep woods. Again, as the closed glade is in the
lee for winds from all directions, the air is always somewhat stagnant
and clear weather celestial radiation nuTst always make a lower mini-
mum temperature than where there is free circulation of air. The field
station, when there is a i^air, one in the forest, the other near by in
the field, must always have something of the glade character in its
temperatures, and the data given at this point must be somewhat af
fected by it. The relations for range of temperature, as between the
interior forest and its external station should show this, and the ranges
should be somewhat greater than between forest and plain. Whether
there is an exaggeration of the mean values of W—0 can only be told
by observation. Fortunately the plan of the Swedish observations
included and was rather devoted to this particular side of the problem.
The stations were really of four kinds. The first were under woods, of
which there were but three. The second were the parallel field stations
for the three forest stations. These were usually located in an '' endroit
libre" (opening), but the distance of the forest in all directions is given, '
so that this means a " glade"' of greater or less size. The third class
were stations in "clairiere" (clearing), with the distance from woods in j
all directions (and not very great). Each of these also signifles a
"glade," but large and unlike the preceding. It has no corresponding
forest station. The fourth class comprises stations in largeoi^en places,
far from forests. These are stations in the plain. This distribution of
stations attbrds the great advantage of permitting a study of the aver-
age amount of change due to the "glade" positionof a station and thus
enabling us to pass from therelations between ])arallel stations to those
between wooded and treeless districts. It has, however, the disad-
vantage that the stations are not strictly comparable, as in the " par-
allel " system, and as would be the case in the " radial " system. This
difficulty can be in large i)art overcome by the careful selections of

TEMPERATURES IN WOODS, GLADES, AND PLAINS.

85

stations, and this Dr. Hamberg', as appears from liis discussion of tlie
Swedish results, has endeavored to make.

The deductions of Dr. Hambcr!:;: have been taken without change,
excei)t in reducing tlie units to the commou English ones. The air
observations were taken at a height of 1.8 meters (.~).0 feet). The series
of observations discusse<l for temperature are usually from three to five
years' duration. The mean annual results are shown in Fig. 45. lu

Mi III Ilium.

2farliiiinii.

ANNUAL AIR TEMPERATURES.
Woods — Glade. - - - -

* WaoTls^ Plain. . . . -

ANNUAL SOIL TEMPERATURES.
Woods — Glcidp, (if 70 inrhc":.

Weeds — Plain,

Woods — Glade, at 5 feet.

Woods — Plain, ." "

1

M "^ o V N "n V
+ + 1 1 1 1
1 1 1 1 1 1 1

Scale.

Flfi. 45.— Dift'erenocs id' iiicaii annual temperature between wnod.s and glades and woods and iilaiiis,

Swedish observations.

the air temperatures the maxima are higher in the woods as compared
with the glades and as compared A\'ith the plains. They are, therefore,
erected above the zero line. The maxima arc reduced in botli cases,
and they are let fall below that line. It appears at ouce that the meau
range is larger (the total line longer) for the reduction by wood as com-
pared with plain. 1 f the <'om])arisou of woods with glade be ^V— G, and
wood with plain H— /*, then the following exhibits the dilferences:

I'l tHjiiritliiri (li()'rrrinrs in irouil.'i, i/liulis, diitl i)liiiii.'<.

Maxima
Minima.
Mean . . .
Kange . .

W-G.

Degrees.

—3. 0(i

1 l.5:i

—0. 70

4. .-.9

W-P.

G -P.

Deqrees.

-■.i. n

+0. 99

— i.i:i
4.'j:!

Degree.

-0.18

—0. 54

0. :{0

0. ;{ii

86

J'OREST INFLUENCES.

The range is exagQerated in tlie glade as comi>ave<l with the]>lain by
0°.3G. The niiiiitua are also exaggerated in the ])lain hy (K'.54, but it
ai)pears that in this ease the niaxininin is reduced.

The disajjpearance of this feature of glade temperatures is doubt
less dne to the large size of some of the glades, and also to the high
latitude of the st.ations, giving a low latitude to the meridian sun.
The relatively large reduction of the maxima causes the mean temper-
ature of the glade to be 0^.30 loMer than that of the plain. It ajtpears,
therefore, that the glade climate is l)y a small (iuantity more rigorous
than the climate of open plains. The forest tends, in so far, to exagger-
ate or sharpen the diurnal changes of temperature.

Fig. 45 also shows the values of W—G and W—P for soil tempera-
tures at the depths of 20 inches and of 5 feet. Tlie temperatures are
slightly higher under plains; G — F at both depths = +0°.l.

Fig. 40 shows the monthly values of ^y-G, W-P, and G-P. The

MINIMA.

MEAN.

MAXIMA.

Plfi. 4C. — Differences of ;iir- ti'iii])erataire in fore.st, glade, ami ]il!iiii (Swedisli).

stations are not necessarily the same for the three figures nor the same
as for the i)receding. Tlie first curve is of the same character as the
temperature figures in the j^receding sections, and shows the same gen-
eral features. The cooling etfe(;t is much more marked in the spring
than lias been the case with any examined before. The second shows
an exaggeration of the features of the first, except for tiie line of
minima, and these ai)pear in si)ring and summer to be much the same
in forest and jdain.

TEMPERATURE DIFFERENC'E.S IN GLADES AND PLAINS.

87

The third figure enables us to compare th<^ dKIereuces between glade
and pliiiii. They are small, but eurious. Tlie line tor minima (dotted
line) and for maxinni (broken line) hav(^ changed ithu-es. Tlie maxima
are reduced generally, though slightly, in glades as compared with
plain, but the minima are exaggerated by a quantity two or three times
as large. The result is that the glade is cooler in mean temperature
for three of the seasons, and that its diurnal ami»l.itude is sharpened
or its climate made more rigorous. Tn winter the advantages are with
the glade, in the other seasons against it.

To bring- out more clearly the advantage in range possessed by the
plain over the glade. Fig-. 47 has been constructed. The distance from

0

-]"

^

^,:::^

-z"

^"^

:x

/-

^^

.^ .

-y

X

//

::-i^

^^^

-4'

\:

:i=:

y

WINTER.

SPRING.

SUMMER.

AUTUMN.

Fio. 47.— Amplitude of W—G (lower curve) and W— O (upper curve), Swedish observations.

the zero line to the first curve gives the reduction of daily ami)litude
in woods as compared with plains, to the second of woods as compared
with glades. The distance between the lines is the difference in ampli-
tude favorable to the plain as compared with the glade. It does not
change sign during- the year, and glade and plain are alike for only one
month, that of July. Fig. 48 gives the temperature values of TT— 6^

0

,---''

• «^

-r

.-;>

-2°

\j

^^

Jy

?

-3"

\ \

v'V

-4°

\\

/

-5°

Vi

/■■''

-6°

N.^

^^_

y

-7°

\

y

WINTER.

SPRING.

SUMMER.

AUTUMN.

_ W-G
Fig. 48. — Dil'enMice ol' snil tiiniicniturc at 211 feet in depth.

W-P

(nnbr(»ken line) and IT— /' (hrokcn line) lor the soil at the depth of
liO inches. As is to be expected, the temperature of the glade soil is
intermediate between that of the woods and that of the jdain.

Dr. Ilamberg- also studied the ettect of clear and (doudy weather on
the differences between glade and i)lain. The following- table gives
tlu'inean values oC (i —V foi- cleai- and cloudy weather at the three
honrs (»f daily obseivation. The clonds cause the temi)eratures at the
two stations to ajjjjroach th(^ same degree, reducing' ii — V to zero:

88

FOREST INFLUENCES.

Tvmpi'raliire differences in yJade and plain.

Clear.

Cloudy.

8 a. m.

2 p.m.

Degrees.
)-0.i)

-fO.2
—0.2
+ 0.4

+o.y

9 p. m.

8 a. ra.

2 p.m.

!l |>. 111.

Winter . .

Degrees.
0

+0.5

U
+ 0.4
+0.2

Degrees.
—0.5
—1.4
-1.8
—1.3
—1.3

Degrees.
—0. 2

+0. 4
+0.2
+0.4
+ 0.2

Deqreea.
0
0
—0.2
0
0

Degrees.
+ 0.2

—0.2

0

0

Year

0

Again, J)r. Hamberg; studied tlie relations of this difference to the
direction of wind, and fonnd the effect to be very slight, perhaps not
appreciabh'. Tlie horizontal reflection of the rays of heat from tlie
foliage could cause an area of greater maxima and the shelter from the
winds an area of lower maxima about a forest. The two together
would cause an area of greater am])litudes or more rigorous climate,
and this has been directly observed by La (Jour in Denmark. Obser-
vations ^vel•e taken at a series of stations at dillerent distances from
the margin of the forest but at elevations above the sea, varying by
only 10 or 12 feet. The result of twenty-four days' observati(m in the
summer of 1807 in Jutland are given in the following table, where the
corrections for differences of elevation are already made. The ther-
mometers were arranged on a line from a northern to a. southern forest:

station.

200 feet witliin northern forest

At its ni:iri;iii

200 feet from iiiiirgiii

401) feet f 1(1111 iiiarji'iii

2lKI feet fioiii soutli forest

At its iiiiirgin

200 feet within it.'

Mean

tempera-

ture.

Degrees.

.'■)U. H

.'•>7. 2

58. ;i

."iS. 1

.08. ■A

57.7

57. 2

Daily

aiii]ili-

tude.

Degrees.
9.2
9.9
i:t.3
12.6
13.3
12.4
11.0

It appears that the least amplitudes are in the forest and the great-
est at something over 200 feet, not less than 400 feet, outside. In this
case the highest temi»eratures are at the same distance. A correspond-
ing series of observations were taken on the island of Zealand, and a
similar area of highest amplitudes was found.

OONCLITSIONS REGARDING TEMPERATTTRES WITHIN AND WITHOUT

FORESTS.

This completes the review of tiie direct comparative observations of
temperatures in woods and without. While the t)bser\'atioiis show,
like all other meteorological i)henomena, very great variations due to
soil, topography, latitude, and many other agencies, they permit us to
draw a series of couclusitms which nnu' be considered as definitely
established in their general features. Tliey are —

TEMrERATURES JK WO0I>ED AND TREELESS REGIONS. 89

(1) Tliero is one season only of iiiarked forest action and that is the
wann season. The action is most marked in the early summer months,

(2) In the forest durinj>' tliis season the maxima of temperatures are
lowered, the minima are rais(Ml, the mean is lowered by a degree or two,
and the (hiily range is reduced by several degri'cs. Tliis action decreases
slowly up to the level of the foliage, tlien rapidly, disappearing at a score
or two of feet above the foliage.

(3) The forest litter plays an important part in preserving the tem-
perature of the forest soil.

(4) The vertical temperature gradient through the forest is reversed ;
it is greatest at the level of the foliage, decreasing slowly downwards
but rapidly upwards.

(5) The diurnal amiditude increases from the margin of the forest
outwards to a distaui^e of a score or so of rods, where it reaches a maxi-
mum. The amplitude is also greater in glades. Hence the extremes
of temperature are exaggerated just outside the forest.

(G) As a result of the discussion of tree temperatures and the organic
sources of warming and cooling in the plant, it maybe added that —

(7) The changes of heat due to organic pi'ocesses are not sensible ex-
cept, possibly, the cooling due to transpiration. The heat used in this
process is an appreciable fi'a(;tion of the heat from the sun's rays and
the cooling due to it may lower to the point of fiost a temperature
already falling from general meteorological causes.

TEMPERATURES IN WOODED A:N^D TREEEESS TERRI-
TORIES.

In passing to the subject of temperatures and forests in its broader
as[)ect, namely, in so far as it relates to differences in wooded and tree-
less territories that are situated widely apart, the strict comparability
of the stations is lost. The stations to be compared are distant from
each other at different elevations above sea level, in diftVrent latitudes,
an<l, possibly, members of different meteoi-ologieal services, with all the
differences in instruments and methods which that implies. These dif-
ferences in the stations can not be so ])erfectly neutralized by ap]dying
correcti(»ns to the observations as entirely to satisfy the critical reader,
because the corrections are themselves unceitain and they often sur-
pass in quantity (for instance, the corrections for altitude) the amount
of change which forests might be expected to make. Besides, it is often
impossible to say when a ditference is found that this is due to the forests.
It may be due to other causes, for instance, the coolness of eva])oration
in regions of heavy rainfall, and the existence of the forest may be due
to the very difference found. These dilticulties may be evaded more or
less completely iu certain definite cases, and it is the published cases
of this sort which will now be given. The case of Vienna and its sur-
roundings, discussed by Dr. Ilann, is one of the most satisfactory. It

90

FOREST INFLUENCES.

is the case of the temperatures in open fields about Vienna and in the
Vienna forest. The elevations vary by less than 200 feet and the ob-
servations are corrected for the variation. Tho uiean temperatures
about Vienna are obtained from four stations.

Stations near Vienna.

Lati-
tude.

Eleva-
tion.

Janu-
ary.

April.

July.

a
67.8
67.6
68.4
68.9

Octo-
ber.

Tear.

48-1-
15'
13'. 9

5'
7'

Feet.
CG3
847

787
853

o

29.5
29.3
30
29.8

o

49.3
49.1
49.6
49.3

o
50.2
.50. 2
50.7
.50.7

o
48.7
48.6
49.3
49.3

INciif Stprnwarte * ....

Atoilliiit'" .........

Percbtoldsilort'

The best forest station for the comparison is Hadersdorf, IS miles
southwest of the Hohe Warte, and its altitude differs from this by 92
feet. If the second or cooler pair of stations in open country near
Vienna be represented by A, the first or cooler by B, and the observa-
tions at Hadersdorf by C, the comj)arison stands as in the following
table.

Latitude.

January.

April.

July.

Octo-
ber.

Year.

A

o /
48 6
48 14.5
48 13.

o

29.9
29.4

28.6
-0.8

o

49.4

49.2

47.7

—1.5

o

68.7

67.7

65.5

—2.2

0

50.7

50.2

48.7

—1.5

o

49.3

48.6

47.1

—1.5

B .

(7

C-B

This table shows that there is an appreciable cooliug- in the Vienna
forest and that this is greatest in summer and least in winter. The
forest consists of beech trees only.

A detailed comjiarison also brings out a curious fact. The tempera-
ture in the forest region is nearer that in the open region at the 2 p. ra.
observations than for that at the 9 p. m. observations. The monthly
means for the years 1875-1884 give the Hadersdorf temperature as 0o.2
higher to 0^.7 lower than the Hohe Warte temperature at 9 p.m., 1^.1
lower to 40.5 lower, the difference being greatest in the summer months.
The 7 a. in. observations are intermediate, varjdng from l^.l to 2o.2'lower
at Hadersdorf. The explanation of this is probably to be found in the
fact that the station at Hadersdorf is, in some sort, a glade station.
The thermometer is not under trees, but is on the north side of a build-
ing nnder the usual Austrian shelter. As ah-eady pointed out, the
glade has a greater range than the open field, and its maxima and min-
ima must fall at about the same hour of the day. If the 2 p. m, obser-
vations were deferred an hour or two, the temperatures at Hadersdorf
would probably be higher than at Hohe Warte.

The stations in forest regions used in making these general compar-
isons are not likely to be under trees. They are the ordinary meteoro-
logical stations, and while in wooded territories, they could not prop-
erly be established under the shelter of the compact foliage of a forest.

TEMPERATURES IN WOODED AND TREELESS REGIONS. 91

They are generally iu larger or smaller open places, generally on tlie
north side of the bnihliiig. Tliey ai-e, therefore, often of the nature of
" glade" stations. Whih^, they will not show the relation between the
interior air of forests and the surface air of prairies, tliey will bear on
the pracitically more interesting question of the ditt'erenccs between
these spots of a wooded region which are occupied by man and the
prairie or plain, or the treeless region. The results, however, will dif-
fer somewhat from those obtained by a system of parallel stations.

The meteorologist who has published the most elaborate study of
this subject is Dr. Woeikofif, who has devoted a chapter to it in his
book on the climates of the earth. A striking case given by him in
some detail is that of northern India. April to July are here the hot
months. From July on the heat is tempered by abundant rains. The
temperatures are compared, for the hot months, over a territory that
includes the three different characters of treeless, transition, and
wooded. In the first the vegetation is already burnt up by March, ex-
cept for small stretches where irrigation is practiced. TJie third is the
great forest region of the middle course of the Brahmaputra and its
southern tributaries, where the surface is covered with the densest for-
ests, and the openings are little else than islands in the sea of forest.
The transition region lies over the delta of the Ganges and northward
and has no extended forests, but groves and scattered trees and abun-
dance of bamboos. The table which follows is Dr. Woeikotf 's, except
that for the most of the stations the temperatures have been taken from
a later publication, that of Mr. Blanford's " Climates of India."

station.

Mean temperature.

Mean maxima.

Name.

Lat.

Alt.

Dis-
tance
from
sea.

Tear.

Apr.

MaJ^

June.

July.

Dec.

Year.

Apr.

and

May.

Open territory.
Anra

O

27

27

25.5

25.5

24

21.5

20

26

25
26
27

Feet.
555
369
307
183
66

26
110
125

104
384
333

Miles.
679
533
432
303
161

0

337
283

131
249
304

0

79

78
78
78
78

79

77
77

76
70
73

o

88

87

88

87

85

82
84
84

78
78
74

o
94
92
92
89
85

85
85
83

80
78
78

o

95

92

91

88

84

85
85
84

82
80
83

o
87
86
85
85
83

83
84
84

82
81
84

o
62
01
61
02
CO

67
63
03

60
05
01

0

110
114
116
110
109

'"'ios

105
99

0

100
104
106
101
101

Lncknow

A llahabad

I'atna

Burhanpur

Internit'diate territory.
Sangod

Durbm}i;a

96

98

87

Purneali

Porest territory.
Silchar

Goalpara

Sibsagar

99

86

The fall of the temperatures in the hot months i\s the forest is reached
is very noticeable, and it is still more so in the reduction of the maxima.
The cooling in the forest region even appears to surpass that due to
proximity to the sea. It is true that this ditfercnce iu temperature
may be in part due, as has been claimed, to other causes. The teni-

92 FOREST INFLUENCES.

peratiire in latitudes 20° to 30° N, seems to fall slowly from Western
India to China and Tonqnin, but Dr. Woeikoff finds tliat it rises in tlie
east, in tlie hot season, where there are extensive treeless regions, and
instances Hanoi, in 21° N., where the annual temperature is 75^.6, May
830.8, June 88o.2, July 8Co.9.

Dr. Woeikoff also gives series of temperatures near jiarallels of lati-
tudes across Europe and into Asia. A correction of 0o.9 is made for
each degree of latitude, and of 1°.3 for each 100 meters of elevation.
All temperatures are reduced to a uniform elevation of 200 meters (656
feet). The order is always from west to east.

Mean temperaiiires for Jnhj on the parallel of 52° N.

Degrees. Degrees.

Valentia 57.6 Orel and Kursk 67.6

Leipsic 62. 6 Poljiiulii 65.7

Warsaw 64. 7 Orenburg 69. 1

Tscheniigow 65.1 Akmollinsk 70

The mean temperature is lowest on the Atlantic coast. It rises up
to Poland by seven degrees, then remains steady over the forests of
the Bug and Dneiper, then rises over the black-earth region of Central
Russia with extensive cultivation (Orel and Kursk), falls again in the
forests over the sources of the Sura (Poljiinki), to become again higher
over the Kirghiz steppes.

July mean iemperature along 50° K.

Degrees. Degrees.

Guernsey 59. 5 Troppau 68

Brussels 62. 6 Arvavarallja (Cariiatbians) 64. 2

Wiirzburg 68 Lirow (Lemberg) 65. 5

Promenbof (N. W. Bohemia) 64.4 Kieft 66.2

Prague 68 Cbarkow 68.4

Hochwald 63.9 Seniipalatinsk 72.7

The temperature increases rapidly to the Main valley, then rises and
falls over the forests of Bohemia and the Carpathians, becoming hardly
higher at Charkow, where there remains some forest, than at Wiirz-
burg. Eastward from here it rises rapidly.

July mean temperaturea along 48° N.

Degrees. Degrees.

Brest 62.2 Bistnitz (Transylvania) 68

Versailles 65.5 Czeniowitz 68.9

Carlsruhe 66.6 Ekaterinoslaff 71.6

Vienna 67.8 Lngau 72.5

Debreczin (Hungary) 70.7 Irgis (steppes) 75.6

Rosenau (Hungary) 68. 9

Here the rise is about eight degrees from the Atlantic to the prairies
of Hungary (Debreczin). It is cooler in Eastern Hungary and in
Transylvania, where there is yet much forest. Czernowitz is near the

TEMPEKATURE IN RELATION TO PARALLELS OF LATITUDE. 93

South Eussia steppes, but it escapes their warming influence by the
beech forests around it. This is the eastern limit of the beeches.

Jiih/ mean temperatures on 46° N.

Degrees. Degrees.

La Rochcllo 66.7 Or<aviczaMountains,SontlieastHun-

Miliiu 62. 9 gary 67. 5

Trieste 72.7 Pojana Ruska (Hungary) 67.8

Agraii 71.1 Odessa 71.2'

Szegedin (Hungartan prairies) 71.6 Clierssou 72.5

Arad (Huugarian prairies) 73 Astrakah 75. 6

Eaimsk and Kasalinsk 76.1

The high temperature of Trieste agrees well with its bare and burnt
surroundings. In Croatia, far from the sea and where there are woods
in the jilain, it is cooler, warmer again in the Hungarian prairies, but
50 or i)° cooler in southern Hungary and Banat, where the vegetation is

luxuriant.

July mean temperatures along 42° N.

Degrees. Degrees.

Oporto 67.6 Poti 70.9

Rome 75. 2 Tiflis 78. 8

Ragusa 71.5 Kutais 73

Poti is in the midst of the swamps and forests of Mingrelia and its
temperature is 4° lower than in the Dalmatian towns on the Adriatic.
Even Kutais is cooler than Eagusa, though the former lies far in the in-
terior and the forests have been cut freely in its vicinity.

JhJij mean temperatures along SS"^ N. »

Degrees. Degrees.

Lisbon 70. 5 Smyrna 77. 9

Campo Major 76. 3 Lenkoran 74. 7

Palermo 76. 5 Krasnoodsk 82. 0

Athens 76. 2

The low temperature of Lenkoran is especially noteworthy because it
is deep in the interior of a continent and the cloudiness and rainfall
are light. Yet in the vicinity are extended forests Avith luxuriant
vegetation. It is on the Black Sea, while Poti is on the Caspian. Near
l)oth are extensive forests. If the July temperature is reduced to 42'^ N.
but not corrected for elevation it proves to be 73°.6, only 0^.4 higher
than that of Poti.

Dr. Woeikoff also calls attention to ih^ remarkable relations of
Bosnia, Herzegovina, and Dalmatia, along latitude 44° N. The tem-
peratures are here the animal means.

Degrees. Degrees.

Banjatella, Bosnia 70. 3 K lissa, llorzogovina 74. 8

Dolnja Tiisla,. Bosnia 68. 2 Mostar, Herzegovina 76. 1

Tra\vnil<, Bosnia 70.3 Knin, Dalmatia 73.6

Serajewo, Hosnia 69.3 Lissa, Dalmatia 72.3

94

FOREST INFLUENCES.

Here it appears that the deusely wooded Bosnia is 5° or 6° cooler tliau
rocky, nearly treeless Herzegovina, and even cooler by 2° or more than
the small island of Lissa, which is exposed to the full influence of the
Adriatic.

All this goes to show that in the western part of the old world the
presence of large forests has a very sensible influence on the tempera-
ture, so much so that the gradual rise of temperature from west to east
is almost invariably broken by it.

Comparisons similar to those of Dr. Woeikoff can- be made along
parallels in the United States with results similar in character but not
so striking in quantity.

The writer has made this comparison for the parallels which follow.
The data are taken from the annual reports of the Chief Signal Officer.*
They are reduced to a common elevation of 1,000 feet above sea level.
I^^o reduction is made for latitude, but the stations used are selected not
far to the north or south of the chosen latitude.

Temperatures alo»<j parallel SS"^ N.

Lati-
tude N,

Lynclil)urg . .
Louisville ...

Cairo

St. Louis

Leavenworth
Doilire ("ity. .
Las Animas .

37
38
37
39
39
38
38

Alti-
tude.

Feet.
652
551
359
571
842
2,517
3,899

AnTiual
mean

temper-
ature. Mean

June temperatures.

57
56
56
54
53
56
58

o

73
72
74
72
73
79
80

Max.

82
82
81
82
83
91
96

Min.

G4
64
66
64
63
67
66

December temperatures.

Mean. Max. Min

40
38
o7
33
30
34
35

47
45
44
42
40
47
52

30
30
29
27
24
26
24

Leaven Avorth lies near the margin of the wooded and treeless district,
but in no case are the woods heavy and unbroken. The mean tem-
peratures show no unmistakable effect of the woods, but there is an
evident difierence in the December temperatures and a still more
marked difference in those of June. The difference is most marked in
the extremes, as was to be expected. This is most clearly shown in
the mean ain[)litudes obtained by taking the minimum from the maxi-
mum. The amplitudes for December at stations in the order given in
the above table are 17°, 15°, 15°, 15°, | 10°, | 21°, 28°. The change at.
the edge of the plains is very noticeable. For June the amplitudes are
18°, 18^, loo, 180, I 20°, I 24°, 30°, with the same change as before.

This coiaparison is not entirely convincing of the effects of forests,
because with the change of condition as to trees there is a large change
of elevation, so large that the correction for it surpasses by far the dif-
ference which may be expected from the action of forests. This
difficulty can not be entirely overcome in the American stations, for the
plains form an interior table-land, but individual pairs of stations can

'1885, pp. 82-84, and 1886, pp. 408-411.

MEAN TEMPERATURES IN FORESTS AND OPEN FIELDS.

95

be selected, otlierwise suitable' and wliere the difterences ot elevation
are less marked. Such are yiven in the accompanying table. One
station is in the wooded district, the other in the plains, and they have
been selected with reference to being in the same latitude, to being as
near the maigin as possible, and to having at the same time only mod-
erate ditfereuces of elevation.

Mean IciiipcidtKri's for stationx i?i j;«ir.s.

Littli' Kock

Fort Sill

Fort Siuith

Flirt Elliott

Davenport

Korth Platte ....

Moorliead

Bismarck

St. ViiKiait

Fort Assinniboiiie

Lat. N.

34.7
34.7
35.4
35.5
41.5
41.1
46.9
4(5.8
48.9
48.5

Alti-
tude.

Feet.

288

Annual
mean

temper-
ture.

o
00.0

.Tune temperatures.

Mean. Max. Min

1,200

60.8

451

57.7

2, 050

60.7

615

49.1

2,841

53.3

926

36.4

1, 694

41.2

804

32.0

2,720

45.8

75.3
73.1
75.0
82.1
68.4
75.6
64.5
67.9
61.4
69.9

85.7
90.5
87.5
92.3
76.6
87.0
75.6
78.6
73.1
82.5

67.4
67.8
65.7
67.7
58.9
64.9
52.2
55.6
48.8
58.4

December terajieratures.

Mean. Max. Min.

43.4
38.2
39.2
38.1
26.9
28.1

9.5
11.9

4.0
21.5

50.9
53. 1
50.8
52.7
34.9
42.9
20.1
26.6
14.7
32.8

35.2

37.1

30.0

28.3

19.3

20.8

0.4

6.7

.5.8

14.1

The mean annual temperatures are persistently lower at that station
of the pair which is situated in the wooded district. The same is true
of the mean June temperature, while it is reversed for tlie December
one. The maxima are lower at the stations in the wooded area and
also, generally, the minima. The mean monthly amplitudes for each
pair for June are IS-^.a to 22^.7, 21o.8 to 24o.6, IJo.T to 22°.?, 2S°A to
23-3.0, 240.3 to 240.1 ; for December they are 15o.7 to I60.O, 20° 2 to
240.4, 150.6 to 220.1, 190.7 to 190.9, 200.5 to I80.7. They are generally
larger at the station on the i>lains — always so at the lower latitudes.

It appears, therefore, that wooded districts have an advantage in
temperatures over treeless areas. The forests lower the mean tempera-
ture slightly and also cut down the range.

There lias been much discussion over evidence of change of tempera-
ture at individual stations, or over more extensive areas, which might
accompany changes in the quantity or distribution of the forests. The
general opinion is that the data are not sutlicient to give satisfactory
cou<*lusious. To be sure of a cliange, one must be sure of a definite
and marked change in the amount of forests and must also be s*ire of
the temperatures before and after the cliange. Tlie Hrst requirement
is not without its difficulties, and the second is usually impo.ssible.
The early observations, if any were taken, aie always defective in
methods, instruments, and the care given to them. A variation in
the exposure of the tlnn-moineter alone might cause a greater difi'erence
of mean temperature than we would expect to find between forests and
prairies.

96

FOREST INFLUENCES.

RELATIO?^ OF EVAPORATION TO FORESTS.

The ainoimt of evaporation (Icpouds especially on the teiupeiature,
the Aviud, and the amount of vapor already present iirthe air. Tlie first
two are much changed by the presence of the forest, and it is to be ex-
pected that the evaporation in and about woods would show some
peculiarities. The evaporation varies very much also with the char-
acter of the surfoce from which it proceeds. Some surface must, there-
fore, be taken as a staniUird for evai^orations, and the measurements be
made with reference to this. There is considerable variation in the
standard surface selected, but in the German service it is that of water.
The reductions have been made, in this section, to terms of evapora-
tion from a water surface. At the same time, since the evaporation
is chiefly of meteorological interest because of its relations to precii)i-
tation, the amount of the latter is also included in the comparisons.
The observations were carried on in the open fields and in the forest.
The instrument emi)Ioyed is described on page 39.

Fig. 49 exhibits the comparison of the annual evaporation in fields

n Evaporation In Flel(te (E.O.)
and Woods (E.W.) compared
with Precipitation. (Pr.)

li Precipitation anJ

Evaporation.

Deciduous

C Precipitation and

Evaporation.

Evergreen.

d Precfpltation and
Evaporation,
Young Trees.

Fig. 49. — Evaporation and preriiiitation.

{E 0) and woods {E W) with the i)recipitation. It will be noted that
the evaporation uiuler trees is about one half of that in open fields.
The precipitaticni is that of the open fields and, for the German stations
and years reduced (the ten years, 1M79 to 18S8), its annual value was
34.0 inches (871 mm.). The evaporation in the fields for the same sta-
tions and time was 1-J.7 inches (3U2.5 mm.) anuually, and that in the
woods 5.4 inches (137 mm.). The corresponding percentages of evapo-
ration were 37 and 16, showing a saving of 21 per cent, of the precipi-

KELATION OF EVAPORATION TO FORESTS.

97

tation by the woods. Fig. 50 exliibitts the distribution of the ev;ii">o-
ration for the year. The upper line represents the progress for the
water surface in the open air, the lower that for the same in the woods.
The shaded space indicates the saving, in inches, of a sheet of water

Flu. 50. — Monthly cvapuratiou iu tliu liflds (upper curve) aud woods (lower curve).

which is affected by tlie cover of the forest in lessening evaporation.
The evaporation and its saving by the woods are botli greatest in IMay
and June and they decrease symmetrically on each side of these months.
The amount in the winter months is very sm;ill, and it is, moreover,
somewhat uncertain, as it is difficult to make the observations at that

P'lO. 51. — Percentage of evajjoratiou iu the woods as coiuparod with that iu open fields.

season. The annual evai)oration in the A\oods is 44 per cent of that in
the fields. Fig. 5L represents the percentage for the different months
of the year. The upper straight line stands for 100 per cent, or the
evaporation in the fields during the year, and the curve below repre-
sents the percentage in the woods. The shaded space is the percentage

100

Fio. 52.— Ratio of evaporation from water surface in field (upper curve) and forest (lower curve) to

precipitation (toj) line).

of lield evaporation which is saved by the foics-t. This covers a full
half of the diagram, with some addition to this from May to November.
It is also interesting to compare the ev;ii)oratioii iVoiii a water surface
with tlie precii>itation from month to month. Tliis is done in Fig. 52.
12444— No, 7 7

OQ

FOREST INFLUENCES.

The upper straight line represents preeipitatioii, Aviiile tlie u])per curve
represents evai)()ration in tlie fields, the hjwer that in the forest. L're-
Aious diagrams showed that the amount of evaporation increased rap-
idly up to May and June; this one shows that the percentage of rain
evaporated increases even more rapidly and reaches its maximum iu
April, a month earlier. From this point it descends more sh)wly to the
winter months, when the minimum i« reached. Almost the same course
is taken by the evaporation within woods. The hatched spaces below
represent the percentage of precipitation evaporated within woods and
without, the cross-hatched part that saved by the woods or evaporated
in open fields in excess of that evaporated in woods. The saving is
greatest at the time when the evaporation could dispose of the largest
eprcentage of precipitation, namely, in spring and summer.

It is interesting to see whether the evaporation iphenomena ])resent
any difference with different kinds of trees or with trees of ditt'crcnt
uges. Figs. 41)/> and c exhibit the annual ([uantity of precipitation
and evaporation for stations Under and outside of deciduous and ever-
green trees respectively. It will be noticed that the saving of evapo.-

100

/

_^

75

-^

~^^

-^

/

"^

\^

^

50

^— -

^

\

^^__„-.'^

25

r.^'C ,

- — ,

"la

--^^- —

— *

— —

— -'-"

--^^-

0

WINTER.

SPRING

SUMMER.

AUTUMN.

EVERGREEN TREES.

DECIDUOUS TREES

YOUNG TREES.

Fig. 53. — Pcrceutajic of cv;i point ion in -woods to tlKit in the opiai air

ration under trees is more marked for deciduous trees. The percent-
age of evaxjoration in forests to that outside is 41 for deciduous trees and
45 for evergreens. Fig. 41)J gives the same data for young trees. It
will be noticed that the j)ercentages are here entirely changed. The
evaporation in fields approaches the quantity of precipitation (which is
here small), while the evajtoratiou under trees hugs closely that in the
open. The evaporation in the open is here 73 ]»er cent of the rainfall
Avliile that under trees is 58 per cent, which is 80 per cent of the former.

Fig. 53 gives the percentages of evaporation in woods compared to
that outside for tlie various months. That ior dcciducms trees (the
dotted line) varies more during the months and sums up slightly less
for the year. The effect of the leafless condition in early spring is
faintly visible.

The unbroken line represents the i)ercentages for young trees. It is
very uneven (on account of the series of (►bservations, only seven years),
but it run>; much higher aiul ocu'asionally the evaporation under the
trees is greater than Hint onrsidc.

RELATION OF EVAPORATION TO FORESTS.

99

Dr. Ebermayer's series of observations incladedinore elaborate ones
ou evaiioratioii tlum those of the German service. In addition to the
evaporation from a water surface he also measnred tliat from bare soil
ontside the forest and within, and that from soil in the forest which
was covered by forest litter. The latter observations were made only
in the months from April to October. His instrnments were first a
simple one, somewhat like that of the German service already described,
and then the evaporometer of Lainont. The evaporation from earth,
either naked or covered with litter, was measured by an apparatus
which worked on the x)rinciple of the student lamp and Icept the earth
always saturated. The vessel containing the earth is of zinc. The
instruments are protected from the sun, rain, and snow, but the air is
given free access to them. The following table gives the percentages
which are derived from his results, with some corresponding ones from
the German stations to permit comparison. The evapf)ration from a
water surface is the mean for six stations for three years ; the other evapo-
ration percentages are for two, Avith the exception of two months. The
precipitation is for one year. Thepercentages are always made from the
averages for corresponding years. The data from the German observa-
tions are for ten years.

Evaporation in xooods in per cent of evaporation in the open.

Dr.

Ebermayer's results.

German observ

ations.

Water surface.

Bare soil.

Soil un-
der for-
est lit-
ter, nnd
witliiu
forest.

Eain-
fall.

AVatcr surface.

Rain-
fall.

Open.

Wootls.

Open.

Woods.

Open.

Woods.

.45
.43
.00
.35
.34
.33
.41

1.15

.91

1.07

.89

.87

.92

1.26

.64
.37
.38
• .34
.36
.39
.44

.27
.1(5
.14
.12
.11
.11
.18

1.75
.68
1.46
1.02
1.00
.59
3.45

1
1
1
1
1
1
1

.51
.47
.41
.38
.36
.35
.37

1.37

May

1.35

June ....... ....

1.91

fTulv

2.33

Auirust .--

1.98

Sf>pt6iuber

2.54

October

8.49

May September

1

.30

.93

.35

.13

.95

1

.39

2.02

This table brings to light several interesting facts. The ratio of
evaporation without and within woods is about the same in the two
services, giving confidence to the generality of these figures. It ap-
pears, however, that the eva])oration from a bare soil is about the same
as that from a water surface. This is very large and is probably due
to the fact that tlie soil is always kept saturated. The ratio between
evaporation from bare soil within woods and without is about the same
as that for the water surface.

The most striking feature of the table is the eftect of the presence of
forest litter on the evaporation from tlie soil within Ibrests. The char-
acter and depth of tliis litter would make nuich dilVerence, and it was
probably kept as nearly as possible like that lying naturally in undis-
turbed woods. In this series of observations, even with saturated soil

100 FOREST INFLUENCES

underneath, it permits an evaporation of only 0.13 per cent of a free
water surface. About seven-eightlis of the evaporation from the forest
is cut off by the woods and litter together.

The ratio of evaporation from a water surface to the precipitation
for the same months is surprisingly large. According to these obser-
vations 5 per cent more than the j)recipitation during the warm season
could be disposed of by evaporation from a water surface. Many sur-
faces permit a readier evaporation, and if we add to evaporation from
a meadow, for instance, the large transpiration from the grass, it would
appear that on the average more water must be evaporated than falls
during this season, which is very imi)robable. The German observa-
tions give a more probable value, that of one to two. Evaporation is
very readily affected by slight causes, and it appears probable that
Dr. Ebermayer's instruments, or their exposure, may give occasion to
an increase of evaporation as comjiared with the German stations.
The annual rainfall is about the same for the period for which results
are published from the two services — 34.3 inches (871 mm.) in the one
case, 34.5 (87C mm.) in the other. Tbe amount of water returned into
the air over a forest is a quantity worth knowing, even though roughly
approximated, especially if it can be compared with that from other
forms of vegetation and from bare soil. The forest air is furnished
with vapor by the forest tlirough transpiration from the trees, through
evaporation from the soil, and through evaporation from the trees.
The transpiration has already been estimated at Co inches. This (and
other special action of the forest) is practically confined to the warm
season, which may be considered as extending from May to September,
these months included. The evaporation from a free water surface in
the open at the German stations for this time is 8.39 inches (213 mm,).
The transpiration is, therefore, 6.5-^8.39= 0.77, or 77 per cent of this
evaporation.

The evaporation from the soil under forest litter is, according to
Ebermayer, 13 per cent of that of bare soil in the open in the warm
season. Dr. Wollny has carried on several series of observations on
evaporation through a forest litter, and from these it appears that the
litter reduces the evaporation by one-fourth or one-third. This would
give three-fourths of 0.39 or two thirds of 0.39, which are 0.30 or 0,20,
numbers considerably larger than Dr. Ebermayer's. As the lattei's
observations were made on an observational scale while Dr. Wollny's
were only on an experimental one, we will take Ebermayer's result of
0.13 as the measure of evaporation in woods.

Only 70 per cent as much rainfall reaches the rain gauge in the
woods as falls in the open fields, the rest is caught in the leaves or
branches and moistens these and the trunk.

Very little runs down the trunk, as has been shown by the observa-
tions at Nancy.^ This 30 i^er cent of rainfall nuist be again evaporated,

' This statement is based on observations made under «me tree, the kind and con-
ditions of which are not oven stated, and the conclusions have been refuted by tho

KELATION OF EVAPORATION TO FORESTS.

101

since the plant takes up little or no water in tliat Avay. We liave, there-
fore, to add to the other forest additions to the vapor of the air 0.30 x
2.02 = O.Gl per cent of evaporation from a free water surface in the open
air. To this should be added the amount of dew which is evaporated,
but this is uidvuown, and as it is condensed from the adjacent air at
night to be added to it again in the daytime, its omission will not make
the result less significant. The moisture added to the air is, therefore,
roughly, for the warm season :

By trauspiratioii, .77 per cent of the free surface evaporation.

By interior evaporation, .13 per cent of the same.

By evaporation of rainfall, .61 per cent of tlie same.

Total, 1.51 per cent of the evaporation from a water surface in open fields.

To get the total in terms of the i)recipitation, this result is to be
divided by 2.02, giving 0.75.

As to other forms of vegetation various estimates are given, but Dr.
Wollny within a few years has made a series of careful measurements.
From them, by reducing, we get —

For mixed crops and fallow, 1.44 of evaxtoration.
For clover, Avheat, oats, etc., 1.73 of same.
For sod, 1.92 of same.

These values are conservative, other authors giving generally larger
ones. The evaporation from bare soil varies much with the kind of
soil and its degree of saturation Avith water.

As Dr. Ebermayer kept the soil constantly saturated, his results do
not represent natural conditions. The evaporations from soil in open
fields will be cut down by general lack of sutficient moisture to satu-
urate them. Schulder's value of soil evaporation has, therefore, been
used. By it the soil evaporates O.GO as much water as a water surface
sheltered from sun and wind, but otherwise freely exposed to the air.
Tabulating the results, we find that the percentages of additions of
moisture to the air over different kinds of vegetation are about as fol-
lows for the warm season :

Evaporalious from various livds of vegetation.

Sod

Cereals . . . .

Forest

Mixed, etc
Bare soil. ..

Of evapora-

tioufroni a

Of precip

free water

itntion.

surrace.

1.92

.96

1.7H

.86

1.51

.75

1.44

.72

.60

.30

Austrian measurements of Ricgler, published in the Journal of the Forest Experiment
Stations in 1870, from which it would api)ear that the rainfall reai liin<; the soil is
increased by from 3.1 to li).fii)er cent through the water rnnnin<; oft" alonjj; the trunlvs,
accordinjLC to the kind of irees; the lirst fij^are referrini^ to spruce, the la.st to beech.
The total loss of water by interception may then be averaged to be 12 per cent of the
rainfall instead of 30. See also p. 134 of this bulletin.— B. E. F.

102

FOREST INFLUENCES.

It appears lliat tlie forest is of medium activity in tins regard. It is
usually surpassed by the ordinary forage crops, but surpasses ordinary
fallow growths and far surpasses the average bare soil.

Still, the forest exercises a strong, conservative effect on the waters
within and under it. It saves the most of its ground water from evap-
oration, and a great deal of the water above the surface. It differs
from other forms of surface covering in drawing its water from a
greater depth. It therefore does not dry out the surface so much,
and it is also less sensible to temporary droughts. Although less effec-
tive, on the average, in adding vapor to the air than are nmny less
herbaceous forms of vegetation, it is more steady in its action, and it
therefore adds its moisture when it is most needed. During con-
tinued dry weather the roots of the grasses dry out the surface soil
and exhaust its water. They then cease pouring moisture into the air.
At such times the forest may continue its transpiration and at the
same time, bj^ its preservative action, its soil is charged with moisture
and may continue to feed the springs tributary to it.

FORESTS AND HUMIDITY.

The psychrometer observations, within and without forests, permit a
study of absolute and relative humidity. The comparative results for
the absolute humidity are given in the accompanying table in the form
of the values of W—0. The plus sign indicates a greater amount of
vapor in the air of the forest; a minus sign the reverse. The units are
hundredths of a millimeter of mercury as it would be shown in a
barometer. The first part of the table is derived from the German
observations. To it are added the values calculated from the observa-
tions in forest and glade ( W—G), and in forest and plain ( IF— P), from
the Swedish service.

At base of tree:

G eneial average

Deciduous

E vergreens

Young

At tree crown :

General average

Deciduous

Evergreens

Swedish observations at base of tree

W-G

TT-P

Winter.

Spring.

Summer.

Autumn.

-I- 2

_ 3

- 5

+ 8

+ 2

rr

-11

+ 6

+ 2

— 3

— 5

+ 8

-h 9

+37

+26

+39

+ 1

-f 1

+ 10

+ 1

0

— 1

—10

+ 1

+ 1

+ 2

+ 10

+ 4

-f- V

+ 7

-fl5

+13

+ 10

-fl3

+ 7

+ 7

Tear.

+0
4-2
0
+37

-1-3
-2

-f-4

+ 10
-1-9

These numbers are very small, so small that, notwithstanding the
absolute humidity is itself of small amount, they are usually less than
1 per cent of it. Tliis is true of the German observations, except for
the young trees. Here the value of W — 0 is always positive and
reaches up to 4 or 5 per cent of the total absolute humidity. This

INFLUENCE ON HUMIDITY.

103

may be dnc to the siuTonii dings of tliis station, which is on the Liino-
bnrger heath, an ininicnse tract of moorland occnpying tlie eastern i>art
of Hanover. There is no distinct ditference between tree crown and
base nor between glade and plain. In the individual stations the mean
is frequently larger than the average of all, and is sometimes consist-
ently positive, in other cases regularly negative. It appears, as might
be expected, that there is no general diHerence between the absolute
humidity in woods and outside, and the differences in individual sta-
tions may be due more to the surroundings of the station in oi>en tields
than to the forest. The lack of a characteristic absolute humidity in
forests can be completely explained by the fact, already shown, that
the forest is not an especially active producer of the vapor of water. Its
position in this respect is intermediate; and even if it did produce a
peculiarly large or a specially small amount of vapor it would not be
apt to show a peculiar absolute humidity because of the rajiid connec-
tion and mixture of gases. A gas spreads in all directions and with
great rapidity usually from the point where it is produced. An obser-
vation by M. Becquerel illustrates the fact that a source of vapor need
not give the air about it a higher absolute humidity. lie compared, on
a summer's day, the psychrometer readings in the air at 5 feet, just over
a tree, just over garden vegetables, and just over a stream, and found
the absolute humidities to be practically identical. However much
these surfaces may hiive difl^n'ed in the production of vapor, the rapid
connection and mixture prevented this difference from being appreciable
as absolute humidity of the air.

While the absolute humidity depends only on the amount of the
vapor of water in the air, the relative humidity depends on this and on
the temperature also. As a difference in temperature between woods
and open fields has already been shown, there must be a difference in
relative humidity, and as the temperature of the woods is lower the
relative humidity must be higher, or the value of W—0 must be posi-
tive. The followhig table shows that this is the case:

Winter.

Spring.

Summer.

Autumn.

Tear.

At tliebase:

+2
+2
+2

+ 1
+ 1

+ 1
H-2

+ 2
+ 1
+ 4

+ 2
-t-1
+ 2

+ 5

+"

+ 5

+ 3

+ 2

H-t

+ 4
■Vi

+ 5

+ 0

+4

+2
+ 1
-t-3

+2
+2

+4

+ 4

Ever^ireens

In tre*' crown :

+4
-i-2

Dfcidiioiia -

+ 1

Kvfrtrro*^iis

-1-2

Swedish observations:

w O

+3

ir p

yi

There is no great diflerence in the different cases. The snrjdus in
the tree crown is a litth^ smaller because the temperature is higher
there.

M. Fautrat has carried on some observations of humidity above trees,

104

FOREST INFLUENCES.

but liis results, as pnblislicd in tlie ComptcH Rciidns^ are frngmeiitary.
They are given in tlie accompanying- table as tliey were i^nblislied.
Tlie results in absolnte humidity were not published. Tlie observa-
tions over trees are comj)ared with those at the same height over fields
and about a thousand feet distant. The psychrometer was several feet
above the trees in the case of the deciduous trees, but close to the top
of the evergreens.

Jamiai'y --
r»^bniaVy -
Mixicli ....

April

May

June

.July

Aii;;ii.st

S(])ti'iiil)tr
October.-.
November
December.

Fores!

of n.i

latte (lecidu-

Fores

t of E

rniPDonvJlle

■ Ills t

ree.s.

evcryreeu trees.

No. of

w.

0.

W— 0.

No. of

W.

0.

W— 0.

years.

year.s.

0

1

0
1

87

84

+3

80

71

+ 9

2

71

69

H-2

1

74

63

+ 11

*>

G4

62

+2

2

55

48

+ 7

2

(54

61

+3

2

57

48

+ 9

2

00

57

+3

3

61

05

+ 6

2

(iO

57

+3

3

57

48

+ 9

1

56

54

+ 2

9

57

50

+ 7

1

77

75

+2

*)

65

58

+ 7

1

80

77

+3

2

79

71

+ 8

1

82

79

+ 3

2

79

72

+ 7

1

83

81

+2

2

65

60

+ 5

The surplus of relalivc humidity over deciduous trees is not great,
but that just over tlie foliage of evergreens is so large as to be signifi-
cant. The pubhshed results do not admit of a ready comi)utation of
the absolute humidity, but the surplus of the relative is great enough
to snggest a surplus of the absolute, large enongh to be easily meas-
ured. The special, elaborate, and refined observations at Ried, in
Austria, show slightly greater variations in absolute humidity than
previous ones, and though they are nearly evanescent, some uniformity
can be traced in them. As they were taken at irregular intervals they
are not suitaljlc for a general tabulation ami their number forbids their
being quoted at length. The conclusions reached by the discussion of
them and given by Dr. Lorenz Liburnau are as follows:

The absolute humidity decreases in the forest from the soil upwards.
The rate of decrease is usually the greatest under the trees and the
least at the level of the foliage. The rate above the trees is interihe-
diate lietween the other two. This rate is least in the late hours of tlie
night, when it may be zero. It increases with the increase of the tem-
perature of the air, becoming greatest in the midday hours, when,
under exceptionally favorable circumstances, it may make a difference
of ]() per cent, or even more. Occasionally, in high winds, the absolute
humidity is greater over the trees. Over the field station the daily
IH'ogress of the absolute humidity was about the same as in the forest,
but the maximnm difference was only about half as great. The abso-
lute humidity in and al)ove the forest is greater than that over the open
fields, and there is some trace of an increase of this difference to the
time of maximum.

INFLUENCE ON HUMIDITY. 105

The relative humidity decreases upwards in forests, but there is au
occasional reversal of this at night. Over the oi^eu lields it decreases
upwards with wet soil, but increases with dry. It is almost invariably
greater about the forest, above as well as inside, over the field station,
and it is higher above the forest than it is at 5 or G feet above the
open field. The small significance of the vapor poured into the air
by forests in raising the absolute humidity suggests the question of
the rehitions of the forest to other constituents of the air. The forest
takes up large quantities of carbonic acid, retaining the carbon and
rejecting the oxygen. Does it sensibly change the relativ^e quantity of
either constituent so that the composition of the air is slightly differ-
ent in the woods? Repeated observation shows that each constituent
is curiously uniform in quantity in the free air. It has been thought
the carbonic acid is quite variable, but the introduction of better
methods of observation shows that, except in confined places where
the gas is produced, the variations are very small.

A study of the numbers involved will show that the carbonic-acid gas
taken u]) by a forest is a vanishing quantity compared witli that which
passes the forest in the same time with the moving air. Graudeau
gives the annual product of carbon by a fi)rest of beeches, spruces,
or pines as about 2,700 pounds per acre. This corresponds to 0,900
l)Ouuds of carbonic-acid gas, or 00,300 cubic feet. ISTow, if the average
motion of the air is 5 miles per hour (a light breeze, a low estimate),
and the layer of air from which tlie gas is taken be estimated at 100
feet thick, there would pass over an acre 050,01)0,000 cubic feet in one
liour. This air nuist contain about tliree parts in ten thousand of car-
bonic gas, and the total amount of tlie latter per hour is l(;r),0!H) cubic
feet. But this is 2|, or more tlian twice as nuich as that taken up by
the trees in the entire season, so tliat the air could provide in thirty
minutes for the wants of the trees for the entire season. If this sea-
son is taken at 5 months, or 150 (laj^s, the amount of air passing is
55 X 10' X 21 X 150 = 10S X 10'" cubic feet, and tlie amount of carbonic-acid
gas is 501x10*^. Tlie ratio of ('arl)onic acid used to that furnished is

•^"^•^ 5!I4000000' ^'" «"" ^''''' ^" ""'^^^•^-

The additions of oxygen to the air would form a still smaller percent-
age of the oxygen already present, for this gas makes up a quarter of
tlie air instead of a thirtieth of 1 per cent.

The case is somewhat different with ozone, as it is formed iu certain
definite localities and its eagerness for combination is sncli that it can
not pass far from its sonrce ))efoi-e it will disappear. The resins of
coniferous woods should be an abundant sonrce of ozone, and although
the decaying matter in and about woods provides abundant occasion
foi- its disa]>pearance before ])assing otf into the outside air, still the
amount of ozone in such woods should be sensibly greater than outside.
Observation has not shown this une<|uivocally, but that may be due to
the numerous imperfections of the method employed.

106

FOREST INFLUENCES.

RAINFAT.Ij in, above, ANT) NEAR FORESTS.

The first result from the parallel system of forest observations with
re.i^ard to rainfall is the determination of the quantity of rain under
trees as compared Avith that outside. Aside from any possible differ-
ence of rainfall over forests and ontside, the tree and its folia,i>e serve
as a cover for the space underneath. The result is that the rain
gauges under the trees fail to get as nnn^h precipitation as 'those which
are in the open fields. Generally the rain gauges are alike in the two
stations. Mucli dilference is caused by the position in which the for-
est rain gauge is placed, whether under a crown of dense foliage or
near its edge, or under the lighter fidiage of smaller or scattered trees.
Very little informntion is given on this ])oint. The rain actniilly caught
aiinnnlly under trees in the German service varies from 8!> i^er cent of
that in the open (at Carlsberg) to 52 per cent (at St. Johauu). The
average is 75 per cent, or three-quarters; and ten of the sixteen sta-
tions do uot vary 5 per cent either way from this mean. A twelve-
years series of observations gives, for the three Swiss stations, 90, 84,

100

75

^^

.^^

50

^ -^

25

0

WINTER.

SPRING,

SUMMER.

AUTUMN.

Fig. 54. — Vercontane of pieri]iitation, received uurtcr trees (all kind.s) — German oli.sei-vatioiis.

and 77 per cent, giving a mean of 84 per cent, considerably larger
than that given by the German stations.

The station at Bellefontaine, near I^Tancy, also gives 84 per cent as
the mean of eleven years of observation.

The proportion of the precii^itation which passes through the foliage
varies slightly with the season. Fig. 54 shows the relative amounts
(percentages of ontside precipitation) for the individual months. The
percentage withheld is somewhat greater in the warmer season, thongh
the difference is not great. Fig. 55 distingiiislies 1)etween the ever-
green and decidnous trees. Tlie values for the evergreens are rep-
resented by the broken line. Tlicy :ipi)ear very even, witli only a slight
tendency downwards in the warm season. Tlie dotted line is the curve
of the percentages for deciduous trees. It is less even than the other.
The dip downward is decidedly more marked, but it is not great. It
would seem that the advent of the foliage would have a much greater
effect than that represented by the slight turn downwards of the curve

PRECIPITATION UNDER TREES.

107

from April to June. The drop for deciduous trees during these months
is from a percentaiie of 77 to that of (K), then to 6~), making the foliage
catch hut 12 per (;ent of tlie iH^ecipitation, while the bare limbs and twigs
in March caught 14 per cent, and in April apparently 23 per cent. The
evergreen trees permit 73 per cent of the rainfall to pass through their
foliage in April and identically the same in June. Admitting (though
there arc some reasons for doubt) that the rainfall is actually the same
over a wood and a place outside but near, this small action of the foliage
as com])ared with the branches and twigs requires esidanation, and,
whatever the explanation nmy be, it must api)ly only to deciduous trees,
as evergreens show no difference in these months. No satisfactory
explanation occurs to me. The catch in winter is largely influenced
by the form of preci])itation, the snow being caught temporarily and
let fall later into the gauges below, but this would not affect the fall
from April on.

100

75

--""-1— "^

— ^^*<

^

— -^^^

50

^^

~'~-—

— — ~

__..-''

25

0

WINTER

SPRING.

SUMMER.

AUTUMN.

..,^. DECIDUOUS TREES. EVERGREEN TREES.

Fig. 55. — Percentage of precipitation, received under trees (dittereiit kinds).

The foliage catches more of the rainfall during the warm season, even
in the case of deciduous trees. The percentages for the German stations
are as follows, the warm season being taken from May to September,
five months :

(Jeiieral average

Average for deciduous trees
Average for evergreens

Warm
season.

70
65
74

It seems that the deciduous trees withhold more of the precipitation
(3 per cent) through the entire year than do the evergreens, and that
this percentage additional becomes 9 in the leafy season.

Less than two-thirds of the rainfall in open fields, during this season,
succeeds in reaching the gauge under deciduous trees; more than three-
quarters reaches it under evergreens. This difference may be easily due
to the character of the foliage in the two cases.

As an average from sixteen stations, and al)out 150 years of observa-
tion, it is found that, in the warm season. 30 per cent of the rainfall in

108

FOREST INFLUENCES.

the open fails to reacli tlie gauges under tlio trees. Taking all seasons
together this deficit under trees is 25 per cent. The disposal of this
water is easy to explain. It does not include the water that drips from
trees, for this is fairly accounted for by the ganges. It is the water which
moistens the tree and its various parts, and also that which flows down
the trunk. The latter part reaches the soil and is measurable ; theformer
is evaporated again without reaching the soil. At ^"aucy arraugements
were made for including in the catch of the gauge the part which flows
down the trunk. At two stations large gauges were put in; they were
of the same size as the tree crown and they embraced the trunk, a collar
about which directed the water into the recei\'er of the gauge. The
percentage of catch in the woods was here 92, the largest in stations
provided with ordinary gauges. The latter were used at the Belle-
fontaine station, which was not very distant, and the catch was here 84 per
cent. This is larger than for most other reported returns, but the difler-
ence between this value and that for Cinq-Franchees may perhaps be

100

^

75

,

.

^

■ —

-"^

~-

— ■

50

25

0

WINTER.

SPRING

SUMMER.

AUTUMN.

Fig. 50.— Precipitation through trees at Ciuq Franch^ea and Bellefontaine.

taken as an approximation to the quantity of rainfoll which flows down
the tree trunk. This would be 8 per cent. These are means from eleven
years of observation. Fig. 5G gives the distribution of the percent-
ages through the year. The upper curve is for Cinq-Franchees with
its large rain gauge; the lower curve is for Bellelbntaine with ordinary
gauges. They both show something of the dip for the warm season.
At Cinq-Franchees the precipitation under trees for one month (Febru-
ary) is slightly larger than that in open fields.

M. Fautrat's observations include those on precii»itation above forests.
The rain gauge over the forest was compared with one at the same
height and a thousand feet away, over fields. The two pairs of stations
were in the forests of Halatte and Ermenonville, and have been already
described. The accessible data were, for Halatte, the monthly values
for February to July, 1874, and August, 1876, to July, 1877; for Ermen-
onville, the monthly values from 1875 to July, 1877. These have been
changed into inches and condensed into the accompanying table:

PRECIPITATION OVER WOODS AND OPEN FIELDS.

109

"Winter .
Spring. .
Su:iiiner
Autumn
Tear . . .

Halatte, deciduous trees.

W

7.44
6.90
7.11

8.88
30. 33

O

7.35
6.54
6.69
7.32

27.90

W~-0

+0.09
+ 0.36
+ 0.42
+ 1.56
+ 2.43

Ermenonville, Evergreens.

IF

6.11
7.47
9.25

7.24
30.07

O

5.41
6.74
8.29
7.20
27.64

w~o

+ 0.70
+0.73
+0.96
+0. 04
+ 2.43

These results are of so great interest that it is unfortunate that the
published series of observations is so short. Kainfall observations, to
give satisfactory results, must be derived from a long series of obser-
vations. ]M. Fautfat's observations are very consistent among them-
selves, however, and the uniformity of signs of W— 0 and the distribu-
tion of the values through the season, show that they are not due to
accidental limits of heavy rains so arranged as to include oue station
and exclude the other. According to these observations, the simple
precipitation over woods is 9 per cent more than that over open fields
in both cases. The amount which reached the ground under the woods
was 90 per cent of that over the fields for one year in the case of the
deciduous trees, and this gives the large average percentage of 39 for
that which was caught by the foliage. For the evergreen trees and
the whole series of observations the amount which reached the ground
under trees was only 50 per cent of that in the open, giving here 53
per cent as the catch of the trees. The arrangements of rain-gauges
by M. Fautrat did not permit the measure of the amount flowing down
the trunks. Using the per cent obtained at Cinq-Franchees as the
measure of this, we have the following curious and interesting relations
between the precipitation over woods, in woods, and outside. The
data are percentages of the rain of all outside, in this case at the
height of the upper gauge, as that at the lower gauge is not given.

Deciduous trees
Evergreen tree.s
Average of two.

Fall over

woods.

Keached

gauge
Hudt-r
woods.

Keached

ground

under

woods.

1.09
1.09
1.09

0.70
0.56
0.03

0.78
0.64
0.71

Held by

the
foliage.

0.31

0.45
0.38

While Fautrat's observations show that there was a higher rainfall
over trees in the department of Oise than over comparative stations in
the open fields, they do not show that that occurs in Germany or else-
where, but gives enough of a presumption to that effect to make it
worth while to examine the subject more fully. It frequently happens
at the German stations that the heaviest rainfalls for the month (which
are always noted) are greater under trees than at the field station, but
as this could easily be due to heavy local rains it is not especially sig-
uificaut. It is otherwise with the monthly rainfalls. If the monthly

110 FOREST INFLUENCES.

amount under trees is greater than that outside, there is a fair pre-
sumption that the rainfall over the trees is persistently heavier, and
where this is found for several months, and in the warm season when
the precipitation is rain and can not be very long sustained in the
foliage, the presumption becomes very much stronger. There were a
good many cases in the two hundred years of observation at the Ger-
man stations when the monthly catch under woods was greater than in
the open fields. 1 will take a few of the most interesting, for the Avarm
season :

Carlsbcrg, May, 1881 (1.24), and September (1.11); also September, 1885 (1.20),
June, 1887 (1.20), and May, 1888 (l.U); Scliou, May to September, 1886 (1.05, 1.00;
1.04, 1.40, 1.40).

These are both stations among evergreens, and the summer reduction
of the catch under trees for this season, for the German stations, is .20
per cent. If this is added to the above they become —

For Carlsberg 1.50, 1.37, 1.46, 1.46, 1.40

For Schoo 1.31, 1.26, 1.30, 1.66, 1.66

and it appears that the monthly fall in the w(«)ds may be from a fourth
to two thirds more than in adjacent fields. If instead of the .20 per
cent of the German service, the .45 per cent found by Fautrat, by actual
measurement, as the difference in rainfall above and below evergreens,
be added, these values become .19 per cent greater, and the rainfall
observations at the forest stations generally become larger than those
in open fields.

But Fautrat's .31 per cent for deciduous trees and .45 per cent for ever-
greens are the results of brief series of observations at only two sta-
tions and can not be extended to the results of observations at other
stations without uncertainty.

They aflbrd, however, some grounds for the ijresumption that rainfall
is heavier over woods, and it is not hard to find reasons why it shonld
be. But it is most satisfactory to wait until the fact has been proven
beyond a doubt, and in the lack of sufficient observations on this point
lies the chief and most important gap in forest meteorological work.
For the completion of the theory of the action of forests on climate
exact observations are needed above trees at many stations and for
long series of years, and the most'imiiortant data to be obtained are
those relating to precipitation. The rainfall* caught under trees is of
minor importance — far subordinate to the amount caught above trees.

PRECIPITATION OVER WOODED AND TREELESS DISTRICTS.

Ill

precipitatio:n^ ofer wooded and treeeess dis-
tricts.

This branch of the forest meteorological problem is the most impor-
tant of all, has been most discussed, and the discussion has resulted in
all shades of opinion, from those absurdly favorable to trees to those
utterly adverse. Direct comparison of observations is hero not possible,
or at least comparable observations have not been carried on, and, as
in the case of temperatures over wooded and treeless districts, there is
always occasion for doubt whether the differences found at the selected
different stations are not due to other things than the presence or
absence of forests. The difficulty is, in fact, greater here than in the
case of temperatures, because the rainfall is especially sensitive to all
sorts of conditions and surroundings, and is also apparently very capri-
cious (changing from month to month or year to year by large amounts
and without any discoverable cause). In addition to all this, when a
difference of rainfall corresponding to a difference of forest conditions
has been found, there is still occasion for doubt as to which is cause
and which effect. There is every reason to believe that with increased
rainfall, other things being favorable, there will be an increased growth
of trees.

The facts at hand do not prove, with entire conviction, that forests
increase the rainfall. The historical method is lacking generally in
the character of the data for the beginning of the comj)arison. Be-
sides, where a change of rainfall is actually shown to be coincident
with a change in the forest growth it is not entirely certain that
the former is due to the latter; it may be due to what are called
secular changes of the rainfall, the reasons for which lie beyond our
knowledge. The geographical method is not entirely satisfactory, for
the reasons already mentioned. The entirely convincing method de-
l)ends on observations above forests and with systems of radial sta-
tions as proposed by Dr. Lorenz-Liburnau, and from these there is not
yet a sufficient amount of x^ublished results.

It will be of interest, however, to see what is the character of the
geog-raphical method so far as it relies on purely meteorological facts.

In lS0(i-'«)8 M. Becquerel took a series of observations of rainfall at
five stations, at Moutargis and Cliutillon-sur-Loing, and their vicinity.
They are in the department of Loiret, 50 to 75 miles (80-121 kilom.)
south of Paris, and contained within a range of 25 or 30 miles (40-48
kilom.) of a country with fields and woods interspersed. His observa-
tions gave results as follows:

station.

Condi tiou.

For 1800.

For 186e-'C8.

La Jacqnemini^ro

Woodfd

Wootled

AVooded

Inches'.
Xi. :iO

yj..')i
:ii.09

2-1.10

Inehcg. Mm.
'J7. 10 088

Lii Salvionniero

27. 0!> 088

!,(' (Jli;iriMn

(27. 08) 687
2(i 2') 060

f'lialilloii :..

Moiilargia

26 94 684

112

FOREST INFLUENCES.

At the end of the first year's observations M. Becquerel calls atten-
tion to the larger amount of rainfall at the more wooded stations; at
the end of three years, however, he ])ractically recalls this mild sug-
gestion. The observations were taken for two years only at Le Cliarme
and were corrected to make a mean corresponding to that for the three
years at the other stations. M. Becquerel also exiimined the relations
of rainfalls to forests iu Denmark, but the results were somewhat am-
biguous.

The observations made near ISTaru-y were arranged in such a way
that some conclusions could be drawn as to the relative rainfall among
forests and in open country.

Cinq-Franchees was in the midst of the large wooded ]>lateau called
the forest of La Haye, composed of deciduous trees about forty years
old.

The field station was in an open space of 10 or more acres, iu which
was the house of the forester. It was five miles southwest of Xancy.
Bellefontaine was on the edge of the forest, about four miles northwest
of Nancy. The field station was uncultivated land, devoted to nur-
series. It was in the bottom of a valley running northwest and south-
east. Six miles northeast of Nancy was the station of Amance, near
the summit of a hill, and surrounded by cultivated lands, not entirely
treeless, but devoid of forests. Observations were begun later at the
agricultural station near Nancy. It was in the open fields about
eleven miles from the forest of La Haye. There are only seven years
of observations from this station, but eleven from the others. The
menu annual precipitation at these stations is gi\ en in the accompany
ing table:

station.

Condition.

Cinq-Francli6es Forest .

Bellefontaine i Forest .

Amance Open . .

Nancy Open . .

Mean

Altitude.

of seven

years.

Feet.

Inches.

1, 247

H2.5

7S7

:il.8

1. 247

25. 5

712

3U.4

Mean

of eleven

years.

Inches.
31.5
.".0. 7
25.7

The stations at Cinq-Franchees and Amance are as comparable as
stations can be made, and the forest precipitation proves to be about
7 inches greater than that of the open fields. Bellefontaine, on the
edge of the forest, has an appreciably greater precipitation than Nancy.
These stations Avere all under one direction, and their results may be
considered as significant, especially as to the first and third, which are
at the same elevation. The position of the second is unfortunately of
such a character that its rainfall might be very sensibl}^ affected.

The rainfall at parallel pairs of field stations has been determined
by a series of careful observations. These stations are always near

PRECIPITATION OVER WOODS AND OPEN FIELDS.

113

forests, and it will be of interest to compare their results with those
given for extensive regions near them but without forests. Such a
comparison is made in Dr. Lorey's handbook of forestry; in repro-
ducing it the units of measurement are changed to feet and inches.
The stations are arranged in the order of their elevation above the
sea. The averages for the larger regions are those published by Dove
in 1871.

Pield station.

Name.

Schoo

Eberswalde . . .

Fritzi^n

Hadersleben ..

Lintzel

Kurwien

Marientlial

Hauenau

INeumath

rriedrithsrode

Lahuhoi'

Hollerath

Schueidet'eld . .

Carlsberg

Sonnenberg . . .
Melkerei

Eleva-

tiou.

Feet.

10

77

98

112

312

407

469

500

1,159

1,158

1,975

2, 005

2, 2;J0

2,400

2,549

3,071

Mean an-
nual pre-
cipita-
tion.

Compared with average over open regions.

Name.

Inches.
28.4
21.9
25.6
30.1
23.3
24.5
22.5
31.6
32.3
26.5
44.2
38.3
50.2
38.9
55.5
69.9

North Sea coast

Brandenburg

East Prussia

Baliic coast

ilanover

East Prussia

Thiiringen and Saxon Provinces

Alsace-Lorraiuo

do

Thihingen and Saxon Provinces

Westphalia

Rhino country

ThiiriDgen and Saxon Provinces

SilisianAIountaius

Harz

Alsace-Lorraine

Mean an-

nual pre-

cipita-

tion.

Inches.

27.5

21.8

24.1

26.0

26.9

24.1

23. 2

30.4

30,4

23.2

30.7

25.6

23.2

27.2

36.4

30.4

Surplus

over
■woods.

Inches.
-f 0. 5
+ 0.1
-f- 1.5
+ 4.1

— 3.6
-f 0.4

— 0.7
+ 1.2
+ 1.9
+ 3.3
+ 14.5
+12. 7
+27.0
+ 11.5
+19.1
+39.5

It seems from this that, where the results at the stations near forests
are compared with the general results in the section of country in
which the station is situated, the forest station usually shows more
rainfall. Lintzel is exceptional, because near young trees on an exposed

moor.

* In regard to this station Dr. Miittricli has lately published an interesting account
■which should correct the above result, and would go far to show a noticeable influ-
ence of the growing forest cover upon rainfall.

This station, situated in the Luneburg heath, was begun to be planted to forest in
1877, at the rate at first of 1,000 to 1,500 acres per year, afterwards more slowly, and
by this time over 8,000 acres have been planted to forest in that locality. Aroimd
the meteorological station a young forest of 10 to 12 years old, of pine and oak, has
grown up. The station is placed in an open field of about 75 acres extent sur-
rounded by the forest growth. The change of conditions immediately around the
station Lintzel, making Lintzel central for an area of about 25 square miles, is repre-
sented as follows:

After reforestation :

10 per cent field, mtjadow, and water.
lOjier cent heath, roads and openings.
80 per cent of forest.

Before reforestation :

12 per cent field, meadow, etc.
85 per cent heath.
3 per cent old forest.

There are now regular meteorological observations for nine years on hand.
The rainfall observations are compared with those from stations outside of the
forest conditions, but near enough to Lintzel to bo available for comparison in the

12444— No. 7 8

114

FOREST INFLUENCES.

There are, however, considerable variations in the elevations of the
stations employed to get the average values given.

following table; the values having in both cases been equalized to eliminate irregu-
larities by calculating the means of each three to four years by the formula

2a+J), a+2b-{-c, b-\-2c+d, etc.
3~" 4 4

Table 1. — Equalized values of precipitation, in millimeters.

Tear.

Lintzel.

Bremen.

Lintzel.

Hamburg

Lintzel.

Oslebs-
bausen.

Lune-
burg.

Garde-
legen.

1882

Per cent.
514.7
550.5
639. 3
620. 3
533.3
546.3
650.0
705. 0
668.0

Per cent.
797.6

798.8
821.2
756.2
636.0
508. 0
608.5
665.9
657.3

Per cent.

514.7
550.5
639. 3
620. 3
533.3
546.3
650.0
736.2

Per cent.

643. 7
650.9
681.5
650. 2
570.9
580.1
700. 1
781. 4

Per cent.

514.7
550.5
639.3
620.3
533.3
546. 3
606. 3

Per cent.
673.9
675. 6
676.3
579. 3
496.1
521.2
578.2

Per cent.
561.0
575. 1
619.0
588.7
512.4
530. 2
625.2

Per cent.
534. 7

1883

545. 0

1884

599.3

1885

56". 9

1880

407.1

1887

4.54. 4

1888

504. 9

1889

1890

From this table the percentage of rainfall in Lintzel, as related to the rainfall of
each of the other stations, is calculated as follows:

Table 2. — Bainfall at Lintzel calculated as percentage of rainfall at certain places.

Tear.

Bremen.

Ham-
burg.

Oslebs-
hausen.

Luno-
burg.

Garde-
legen.

1882

Per cent.
64.5

68.7

77.8

82.0

83.9

96.2

106.8

105. 9

101.6

Per cent.
80.0
84.6
93.8
95.4
93.4
94.2
92.1
94.2

Per cent.

76.4

81.5

94.5

107.1

107. 5

104. 8

104.9

Per cent.

91.7

95.7

103.3

105.4

104.1

103. 0

93.8

Percent.
96.3

1883

101.0

1884

106.7

1 885

109. 2

1886

114.2

1887

120.2

1888

120. 1

1889

1890

The author concludes that if the increasing reforestation at Lintzel had no influ-
ence on the amount of rainfall, the figures for Lintzel should have been nearly in
constant proi^ortion to those for the other stations, while from the percentage table
it appears that with reference to neighboring stations the precipitation at Lintzel has
increased with the increasing forest growth. The diiierences in the last years are
not so apparent, because the values could not be properly equalized. It is, however,
undoubtedly proved that at the beginning of observations the rainfall at Lintzel was
less than at any of the neighboring stations and that subsequently it increased from
year to year, until it was in excess of the other stations, except at Hamburg.

Comparing the rainfall at Lintzel witli that of the other stations, and calculating
it as percentage of the mean rainfall of the latter, the following series is obtained:

Per cent.

1882 81.8

1883 ? 86.3

1884 95. 2

1885 99. 8

1886 100. 6

. 1887 103.7

1888 - 103.9

This constant increase, going hand in hand witl^ the increase of forest cover in exr
tent and height, leaves Ijardlj^ any 4o«ht ag tq t"^^ close relation of the t^vo condjr
tious.-}}. e', F,

PRECIPITATION OVER WOODS AND OPEN FIELDS.

115

To eliminate this, compare the stations mentioned with other German
stations of about the same elevation. For this purpose the rainfalls at
192 stations given by Dr. van Bebber were used, and these and the
previous ones were combined for each hundred meters.

Elevationa.

Meters
1-100.

100-200.

300-400.

600-700.

100-800.

100-1000.

Average from forest field station

Averane Irom Van ]>ebber

Inches.

25.8

25.5

+ 0.3

1.0

I7iches.

26.2

22.9

+ 3.3

14.0

Inches.

29.4

27.4

+ 2.0

7.0

Inches.

42.9

36. 0

+ 6.9

19.0

Inches.

55.5

38.6

+ 16.9

44.0

Inches.

69.9

38.0

+31.9

84.0

I'erceiitage of surplus------- .-•

This comparison throws some light not only on the amount of sur-
plus over woods, l)ut on the distribution of it. It increases rapidly
with the elevation. For 328 feet (99.9 m.) or under, it is but 1 per cent;
from this to 1,.300 feet (396 m.) it is 7 to 14 per cent; between 1,950
(594 m.) and 2,270 feet (092 m.) it is 19 per cent; between 2,270 (692 m.)
and 2,028 feet (801 m.) it is 41 per cent; and above the last the sur-
plus appears to be 84 per cent, or the precipitation over forests is nearly
doubled.

It is easy to show that, in general, heavily wooded districts have a
higli rainfall, as compared with similar districts Avithout forests. This
is notably true in India, as shown by Dr. Brandis and others. It is
enough to say that the heaviest known rainfall occurs in one of the
densest and most extensive known forests, that of the middle and
lower Brahmaputra. But in all these cases it is more probable that
the forest exists because of the rainfall, rather than vice versa. A
case which is more favorable to the influence of forests on rainfall is
given by Dr. Woeikoff, depeuding on the rainfall distribution on Java
and Celebes. Java has extended and dense forests in the south and in
the southwestern interior, while the north coast has been largely defor-
ested. The station Tjilatjap on the south coast is at a distance from
mountains. Its annual rainfall is 182.3 inches (4,630 mm.). That for
three stations on the north coast (Batavia, Tegal, and Samarang) is
78.4 inches (1,991 mm.) or 43 per cent only. In this region the north
side of the islands is the weather side for the northwest monsoons and
should have, at least from December to March, more rain than the
south side, because this wind descends to reach the latter. As ^ mat.
ter of fact there falls in these months — •

At Tjilatjap 45.7 inches (1,161 mm.)

At the north coast stations 1.3.3 inches (1,100 mm.)

or almost as much on the lee side as on the windward, while there
should be very much less on the latter, .

116 FOEEST INFLUENCES.

Celebes, where there is no such defore>tiiig-. giv^es the normal annual
values :

Celebes:

Northern peninsula — windward, 110.6 inches (2,809 mm.); lee, 55.1 inches

(1,399 mm.)
Southwestern peninsula— windward, 136.2 inches (3,458 mm.) ; lee, 100.0 inches
(2,540 mm.).
West Centi'al Java :

Windward, 78.4 inches (1,991 mm.) ; lee, 182.3 inches (4,630 mm.)

It appears that in Celebes the windward side has a notably higher
rainfall than the lee side, while in Java it is reversed. Is this not due
to the relative lack of forest growth on the north side of Java? Where
forest growth exists the rainfall is higher, even there. For instance,
at Lormadjang, in a densely wooded level area, the annual rainfall is
73.7 inches (1,872 mm.), while at the two nearest stations on the same
side it is only 4(3.9 inches (1,191 mm.).

The literature of the historical treatment of this problem is very
great, and the most extreme and absurd opinions can be found registered
in it. It usually suffers under the ol)Jection, first of imperfect data at
the beginning, and always under that of uncertainty as to the cause of
any variation of rainfall which may be found.

The argument founded on changes in regions mentioned by ancient
writers deserves consideration only to point out its extreme unreliabil-
ity. The condition of the countries in question at the beginning of the
comparison is based on incidental references in the classics which are
not only incomplete but which are usually in ambiguous terms. The
classical terms for forest have an extensive range of meaning, as will
be seen by reference to any classical dictionary, greater than that of
our word forest, which is used in law English to designate areas which
now have no woods, and in some cases there is no evidence that they
ever had any.

Much more trustworthy are the more modern cases which depend on
meteorological observations; but in these the earlier observations are
not so trustworthy as to give assurance that any differences found are
not the result of imperfect observations. The Avriter endeavored, some
years ago, to show that along the parallels of latitude 40o and 42^1 iu
the Mississippi Valley the lines of higher rainfall were moving slowly
westward with the advance of settlements, but it was abundantly
shown that the sparsely scattered earlier observations were (luite un-
trustworthy.

There is one further case, which is quoted by Dr. Brandis, for India.
In the part of the central provinces between the Nerbudda River and
Nagpur and Rajpur, embracing a part of the Satpiira range of moun-
tains, much attention has been paid for several years to the care of the
forests, and specially to protection against forest fires. In consequence
a large territory, with scattered tree growth or entirely treeless, has

INCREASE OF RAINFALL WITH GROWTH OF FORESTS. 117

been covered with a dense growth of young trees. Over this region the
rainfall has been as follows, at the stations named:

Badniir

Chindwara.

Seoni

Mandla

Burha

Bilaspur . . .
Kajpur

Annual average .

1875 and before.

Years.

1867-
1865-

1865-
1867-
1807-
18^5-
186G-

1875
1870
1875
1875
1875
1875
1875

Mean
annual.

Inches.
39.83
41.43
52.07
53.58
64.51
41.85
51.59

49.27

1876

to

1885.

Inches.
47.83
48.48
54.76
56. 32
71.65
54. 81
54.41

55.47

Per cent

of
increase

20

17

5

5

11

21

5

13

From the 37-year series of observations at Nagpur, and the 40-year
series at Jubbnlpore, on the north side of the Satpuras, it appears that
the probable error of a 10-year series of observations is 5 per cent.
Farther, it appears from the combination of rainfall observations in
India that in 187G-1885 the rainfall was O.G() inch (17 mm.) greater
than in 1866-1875.

If one half of the above difference be attributed to error and to a gen-
eral increase of rainfall, there yet remains an appreciable addition which
might be attributed to the growing forest. At least it may be taken
as an indication which, combined with the results of observations above
trees, makes a good presumption that a forest does actually increase the
rainfall by an appreciable percentage. It should be noted that the for-
est fires practically ceased before 1805, and that at that date the forest
growth had been fairly started over small surfaces. By 1875 it had
extended over a surface of about 1,000 square miles.*

*NOTE. — It seems appropriate in this connection to quote the following extract from
Mr. Fernow's annual report for 1888:

" Blanford, who has recorded these observations in the above manner, adds that
these results, of themselves, are not proof absolute for the influence of forest preser-
vation, since possibly the earlier obscrvaiions were less reliable than the later, but
that these observations may be considered as addenda to the accumulating sigus of
the existence of such infliience on rainfall.

"But even this method, which would class with my retail" methods, although
seemingly simple, before it can be admitted as conclusive, must, as the writer says,
be guarded by those special precautious which are demanded by strict scientific
inquiry.

"The above figures were hailed wi(h satisfaction by those who are bound to prove
by statistics the forest influence on rainfall.

" Unfortunately, as this report goes to press, their value is entirely vitiated by the
following statements made in the Indian Forester, January, 1889, which again ad-
monishes us to bo careful in placing too much weight on statistics.

"Mr. Blanford, in order to assure himself of the value of the rainfall returns he
employed in the discussion of the Central Provinces, wrote to the chief commissioner
on the subject, to which the rejAy was received that ' the chief commissioner fears
that these records of rainfall previous to 188o can not be accepted as altogether reli-
able.' The commisioner explains the reasons why the records appear unreliable, and

118 FOREST INFLUENCES.

The balance of waters in forest regions can not be made up until
there is greater certainty as to the jirecipitation over forests.

It has already been shown tliat the forest preserves a considerable
proportion of the waters which reach its soil. This is due, for the
most part, to its lessened evaporation, and this alone would account
for the moisture of forest soil, for the preservation of bodies of water,
and of perennial springs in regions of dense forest. If the forest also
condenses more moisture from the air, then the same effects would fol-
low in regions more lightly clothed with woods.

FOP^ESTS, WIND, AND STORMS.

There remain a few points of more or less interest, which have
received considerable discussion, but have not been the object of sys-
tematic observation. Among these is the windbreak effect of the
forests. Like any other elevation above the surface, the forest pro-
tects from the wind objects Avhich are in its lee, and, over a very much
smaller area, those which are on its Avindward side. The outlines of
the protected area are detiued by the outlines of a snowbank before and
behind a fence. A section is given in the accompanying figure, where
A Bis the fence, T^ the snow to Aviudward of it, L that to leeward.
The windward protection is so small as not to deserve discussion in
this connection. As to the leeward protection, tlie chief point of inter-
est is the length of the line B 0 as compared with the height A B, or

adds : ' Hence one resvilt of the unsyHtematic registration of the rainfall in the Central
Provinces is to postpone the decision of the influence of forests on rainfall in that
area for another twenty years. It is only one of the many cases of the worthless-
ness of unsystematic observations.'"

It is of interest also to note the following from the same source:
" In the following extract from the same report Mr. Eliot refers to the observations
recently undertaken in the forests of the Saharuupoor district. For the reader un-
acquainted with the Western Himalayas it is necessary to explain that a rao is a
water conrse issuing from the hills, and having, generally, a broad, shallow bed,
which cojisists mainly of bowlders and shingle, and is therefore quite dry or almost
dry, except after a continuous heavy shower. Mr. Eliot has not mentioned that in
each rao levels are accurately taken every year along one and the same line, in order
to note the changes that may occur in the section of the rao in consequence of the
fire conservancy of the entire basin above.

"A different method has been introduced in the Saharunpoor forest division.
Twelve representative raos, between the Ganges and Jumna rivers, have been selected
for purposes of observation by the inspector-general of forests and conservator of
the school circle, and in each forest chowki a rain guage is suitably placed. Five
of them are located in the forest of Sakrauda, which is neither closed to grazing nor
protected from fire. The rainfall measurements will be made by the forest guards,
and the returns submitted to the meteorological department for critical examina-
tion. These observations will probably give a valuable series of data for testing
the effect of different forest conditions in modifying the amount of rainfall, and
hence also probably throw some light on the general question of the influence of
afforestation on rainfall."

INFLUENCE OF WINDJiREAKS.

119

tbe ratio of the width of the inotected strip to the height of the obstruc-
tion. M. Becquerel says that in the Rhone Valley a hedge -J meters
(() feet) high will protect delicate garden ])lants to a distance of 22
meters (70 feet). This is a ratio of 1 to 11, and this ratio is about that
given in the following cases gleaned from Dr. Hough's report on fores-
try for 1887. Judge Whiting, of Iowa, is quoted as saying that this
protection, " with almost mathematical precision, amounts to 1 rod on
the ground to each foot of height'' of the protecting trees. This is a
ratio of 10 to 165. Mr. Barnard, of Pawnee County, Nebraska, stated
that a wiiulbreak will protect an orchard a rod for every foot in height,
but Prof. Thomson wanted trees 25 feet high for every 10 rods of pro-
tection. The two latter cases relate to orchards which rise 10 or 20
feet above the surface. A probable estimate is that the forest creates a
calm area on its leeward side, which is, at the ground, ten to fifteen times
as wide as the forest is high. The protected area has the same relations

Fig. 57. — Snowliaiik lief'ore and behind a fence.

to temperature that a glade has. The air is relatively stagnant, and
temperatures rise higher in the direct sun's rays and fall lower on clear
nights. Evaporation is also somewhat decreased in this space, but its
chief advantage is found in the i^rotection from injurious cold or dry
winds. The rough surface of the forest must also decrease the velocity
of the wind, with the result of affecting the character of a storm which
passes over it, especially if the storm be small or local. Dr. Waldo
has given two interesting cases where the results are due undoubtedly
to several causes, one l)cing the retarding effect on the wind of a sur-
face of city buildings. At both New York and Boston there are three
meteorological stations, one out in the harbor, the second in the city,
the third fairly beyond the city (Central Park and Cambridge). The
mean velocities at these tiiree stations in order are, in ])ercentages:

Harbor.

City.

Suburbs.

New York

Jioston

1.00
1.00

0. 07
0.67

0.37
0.33

The reduction of the velocities as one goes inland through the city
is remarkable, and something similiar may be expected for a wind pass-

ing over a forest.

120

FOREST INFLUENCES.

As tlie interior air of tlie forest; is generally cooler, in the warm sea-
son, than the air outside, it must he heavier, and the difference of tem-
perature must be often so great that the heavier air will overcome the

Fig. 58. — Forest circulalioii.

obstacles to its flow and gradually pour out near the ground. Its place
will be taken by the air above which will settle, and thus there may
be set up a forest circulation, as represented in the tigure (Fig. 58)

Fig. 59. — Day circulatiou.

exactly corresponding to such a system of winds as is found in land or
sea breezes, or may be found over a lake at night.
During clear nights the forest air is generally warmer than that out-

*-« *-~«^ '^-^

•^-^

»—*^ •>— ► iS^^

Fir;. GO.— Night circulation.

side. When this happens a reverse circulation might be set up. This
circulation must be slow, and though somewhat enlarged on by writers
on tlie subject seems never to have been directly observed.

FOGS, CLOUDS, HAILSTORMS, AND BLIZZARDS. 121

INFLUENCE OF FORESTS ON FOGS AND CLOUDS.

The iiifiueuoe of forests ou fogs and clouds has been frequently men-
tioned, and observed in single cases. The fog seems to linger in the
woods after it has cleared off elsewhere. Trees also act as condens-
ers,* as gatherers of dew, hoarfrost, and ice, and the latter phenome-
non is especially remarkable in the so-called ice-storms, where the ac-
cumulation is so great as to overload and break the larger limbs and
branches. In these cases, however, the tree acts like inorganic bodies.
This is illustrated by a celebrated case on the island of Ascension, the
details of which are due to Prof. Cleveland Abbe, who in 1890 per-
sonally inspected the ])henomenon. This case is especially worth quot-
ing because its records have been so badly understood. The principal
water supply for the garrison of this naval station is gathered several
miles away, at the summit of Green Mountain, the upper part of which
has always been green with verdure since the island was discovered;
almost all of this water comes from slight showers and steady dripping
of trees enveloped in cloud-fog on the windward side of the mountain.
Every exposed object contributes its drip; these do not condense the
water, they simply collect it mechanically after it has been condensed
in the uprising cooling air. Whatever fog-drops are not thus collected
SAveep on over the mountain and quickly evaporate again. Thirty years
ago or more efforts were made to plant a few trees in the arid spot at
the garrison landing; none survived, but some few new shrubs were
added to the flora of the mountain top; extensive additions were also
made to the mountain reservoirs and drip collectors and pipes of the
aqueduct system. The few artificial scrubby plants have had no influ-
ence whatever in increasing the water supply.

INFLUENCE OF FORESTS ON HAILSTORMS AND BLIZZARDS.

M. Becquerel found, by a careful study of the destructive hailstorms
in some of the departments of Central France, that these storms show
a marked disinclination to enter forests. On the maps of the depart-
ments studied by him he has marked the severe storms by spots of color.
A glance at the maps shows how persistently the colors keep outside
the forests, especially the larger ones. To controvert these views,
cases of hailstorms in forests have frequently been pointed out, but the
question is not of absolute but of relative immunity of forests from
severe hailstorms, and that these cases do not affect. The subject re-
mains where it was left by M. Be<*querel. It still appears proven that
in Central France hailstorms avoid forests.

Many other relations of forests to storms have been suspected and
advocated. For instance, that the storm precipitation is more intense

See p. 124 of this bulletin.

122 FOREST INFLUENCES.

over prairies, tliat the heavy windstorms — sucli as blizzards — do not
long survive the traverse across a forest area, etc., but these relations
still remain in the condition of hypotheses only. Forests also have
broad and important relations to the flow of water on the surface, to
the protection of movable soils, and to many other features of practical
importance. These are matters for discussion by engineers rather than
by meteorologists.

ITL— RELATION OF FORESTS TO AVATER SUPPLIES.

By B. E. Feknow.

THE TOTAI^ WATER SUPPLY.

The water capital of the earth cousists of two parts, the fixed capital
and the circidating capital. The first is represented not only in the
waters on the earth, but also by the amount of water which remains
suspended in the atmosphere, being part of the original atmospheric
water masses which, after the rest had fallen to the cooled earth, re-
mained suspended and is never i)recipitated.

The circulatiug water capital Ls that part which is evaporated from
water surfaces, froui the soil, from vegetatiou, and which, after having
temporarily been held by the atmosphere in quantities locally varying
according to the variations in temperature, is returned again to the
earth by precipitation in rain, snow, and dew. There it is evaporated
agaiu, either immediately or after having percolated through the soil
and been retained for a shorter or longer time before being returned to
the surface, or, without such percolation, it runs throngh open chan-
nels to the rivers and seas, continually returning in part into the
atuiosphere by eva])(»ration. Practically, then, the total auiount of
water capital remains constant; only oue part of it — the circulating cap-
ital— changes in varying quantities its location, and is of interest to us
more with reference to its local distribution and the channels by which
it becomes available for human use and vegetation than with reference
to its practically unchanged total quantity.

As to the amount of this circulating water capital we have very
imperfect knowledge; in many cases an approximate estimate of the
amount circulating in any given area can not be satisfivctorily made
with the means of measurement at command, for often the precipitation
is so uuevenly distributed, as in the case of local thunderstorms, that
tNVo rain gauges a short distance apart collect varying amounts; hence
a record from one gauge alone would give a very erroneous idea of the
rainfall over the entire area. Furthermore, when improperly situated
or exposed to high winds a rain gauge may furnish quite inaccurate
records, and in the nuijority of cases the amounts collected by the
gauge will be iusutiicient. (See Trof. Abbe's paper in this bulletin.)

123

124 FOREST INFLUENCES.

A singular illiistration of these errors is presented in a hydrograpliic
investigation of tlie river Rhone and its watershed. While the amount
of annual discharge of the river corresponds to a rainfall of 44 inches
over the watershed, the rainfall records themselves for a certain period
gave a precipitation of only 27.6 inches. Even the close estimate of
the waters of the upper Elbe, according to which they drain one-fourth
of the total rainfall, calculated by the ingenious methods of Prof. Stud-
nicka, does not inspire confidence. The attempts of the U. S. Weather
Bureau to relate river stages to rainfall measurements have also so far
failed to yield satisfactory results. Again, much of the moisture which
is condensed and precipitated in dews escapes our observation, or at
least our measurements, entirely; * this is, however, so small a quantity
that it may be neglected in its relation to the total precipitation.

Late investigations have brought convincing proof not only that the
amount of dew is much smaller than was supposed, but also that the
larger part of the deposit is derived from the moisture of the soil and
not from that of the atmosphere; that, therefore, dew formation in
many cases can not be considered an additional water supply, but
rather an element of dissipation. While G. Dines calculated the
annual amount of dew at best equivalent to 38 mm. rainfall as against
127 mm. as formerly estimated for England, Wollny, upon careful
measurements, calculated it for Munich at 28 to 32 mm. or less than 3J
per cent of the total rainfall. The theories regarding dew formation,
according to which the moisture is deposited from the atmosphere in a
manner corresponding to the familiar phenomenon on the ice pitcher,
which was first antagonized in 1833 by Gerster and later by others^
among whom our countryman, Stockbridge, seem in part at least incom-
plete and needing revision.

The distribution of the circulating water capital is influenced by
various agencies. The main fiictor which sets the capital afloat is the
sun, which, by its heat and the air currents caused by it, produces the
evaporation which fills the atmosphere with vapor. Anything, there-
fore, that influences the intensity of insolation or obstructs the passage
of winds must influence the local distribution of the water capital, and
hence a forest cover, which withdraws a portion of the soil from the

*A few experiments on condensation of aqueous vapor made by L. Hampel witli

forest tree leaves are of interest : ^ ^.

Centigrams.

Austrian pine (4 needles) condensed per day in the average 4. 84

Linden (one leaf) condensed per day in the average 24. 40

Oak (one leaf) condensed per day in the average 25. o6

Spruce (a branchlet) condensed per day in the average 9. 80

The linden, of which one leaf condensed 24.40 centigrams of dew, had 1,763 leaves.
It would, therefore, if all leaves had done the same, which is to be sure not the
case, have condensed 430 grams (nearly 1 pound).

On grass the amount of dew per year was found by G. Dines to be 27 millimeters ;
t. c, if collected an amount corresponding to 27 millimeters (over 1 inch) height of
water would have resulted.

INFLUENCE OF FORESTS UPON PRECIPITATION. 125

influence of sun and wind, must have an influence, however small, on
the local water conditions, which may or may not become appreciable
in human economy.

The question of forest influence on water supplies may be considered
under three heads, namely: (1) Influence upon precipitation or upon
the distribution of atmospheric water; (2) influence upon the disposal
and conservation of available water supplies; (3) influence upon the
"run-off" or the distribution of terrestrial waters.

INFLUENCE OF FORESTS UPON PRECIPITATION.

The question whether forest cover has or has not an influence upon
the quantity of rain or snow that will fall over a given area is discussed
at length on pp. 111-117 of this bulletin and also upon additional
data in my annual report for 1888. It may suffice, therefore, to state
only the conclusions which may fairly be drawn from our present
knowledge and experience :

(1) Finding the air strata above forest stations moister and cooler,
although only slightly so, than over field stations, we would infer that
the tendency to condensation over wooded areas might be greater than
over open fields. Experience and measurements seem to sustain this
reasoning.

(2) These cooler and relatively as well as absolutely moister air strata
carried away by air currents must modify conditions near the forest
and possibly increase in its neighborhood also the tendency to forma-
tion of dew and to other precipitation under certain conditions.

(3) While the forest may not everywhere increase precipitation over
its own area and near it, yet the presumption is that large systems of
forest growth over exteusive areas alternating with open fields may
establish sufficient differences in temperature and moisture conditions
and in air currents to modify, if not in quantity yet in timely and
local distribntion, the fall.

(4) It must not be overlooked that the extent, density, height, and
composition, and relative position of the forest in making it a climatic
factor are important conditions, and furthermore that there are certain
rain conditions prevailing in climatic zones (rainy or rain-poor locali-
ties, with periodical, seasonal, or irregular rains) which are due to cosmic
influences and can not be altered, yet maybe locally modified by forest
cover. Hence experience in one climatic zone or under one set of con-
ditions can not be utilized for deductions in another.

(5) Altogether the question of appreciable forest influence upon pre-
cipitation must be considered as still unsolved, with some indications,
however, of its existence under certain climatic and topographical con-
ditions in the temperate zone, especially toward the end of winter and
beginning of spring.

As one of the most striking examples of an increase of precipitation,
seemingly due to forest planting, the experience at Lintzel, on the Liine-

126

FOREST INFLUENCES.

burg heath, may be recalled, where, with a definite increase in forest
conditions over an area of 25 square miles, a regular definite increase
in rainfall beyond that of neighboring stations to the extent of 22 per
cent within six years was observed, and a change from a deficiency to
a considerable excess over the rainfall of these other stations.

EESULTS OF OBSERVATIONS NEAR NANCY, FRANCE.

The following observations were made at two stations in the neighbor-
hood of Nancy, France. The instruments and their disposition were
identical at the different stations. A is a forest station; C, a field sta-
tion; B is on the verge of the forest and at a lower level. The follow-
ing table gives the amount of raintiill in centimeters for the seven
years 18G7, 1868, and 1872, 1873, 1871, 1875, and 187G:

Time of observation.

Station A

(forest

glade), :i80

meters.

Station B

(forest

verge), 240

meters.

Station C

(Held), 380

meters.

"Fphrnarv to AT>ril - -

Cm.
1.5.9
18.9
20.7
21.2

Cm.
16.2
17.1
17.2
18.8

Cm.
14.9
16.6
15.7
17.7

TVTav to fTiilv .-.

Aiipust. to ()ctober

November to January

Year -

76.7

69.3

64.9

RESULTS OIVIAINEI) IX I50HKMIA.

It may also be appropriate here to recall the method by which Dr.
Studnicka tried to establish the influence of forest areas upon the dis-
tribution of rainl^ill, an account of which I reproduce here from my
report for 1888 :

The latest, most valuable scientific v^ork whicli has been done to decide the impor-
tant but dilMcult question of the influence of forests on precipitation is the work of
Dr. F. J. Studnicka, professor of jnathematics at the University of Prague, published
under the title, "Basis for a Hyetosraphy of Bohemia," in which the results of
many years of observation at seven hundred ombromctric stations arc embodied, criti-
cal ly sifted, and scientifically considered. The author employs a wholesale method
which is quite novel, complying with Woeikoff's idea that it is necessary to reduce
the observed data to a common basis for comparison. To understand the sisinificance
of these observations an inspection of the map of Bohemia will be desirable, which
shows it to be a basin surrounded on all sides by high mountains.

The work of ombrometric observations, altliongh ])egun in Bohemia during the
last century, was neAvly organized in 1879 or 1880, when a systematic net of ombro-
metric stations was instituted, and in 1885 and 1886 it was extended to over seven
hundred stations, for the purpose of obtaining accurate data of the quantity and
distribution of precipitation over the Kingdom. Uniform ombrometers were used
and very carefully placed. As at present organized, there is one station for every 75
square kilometers (about 30 square miles). No other country, I believe, can boast
such a service. Although the time of observation at most stations has been short,
and the averages would have been more accur.itely represented by an extension of
observations for ten %q twelve years, yet the last four years of observations, for whicU

INFLUENCE OF ALTITUDE ON PRECIPITATION. 127

all Btatioiis furnish data, according to the author, represeut two extreme and two
average years, and are therefore quite useful.

The very large mass of material permitted a sifting out of doubtful observations
without impairing the number of available ones for the construction of a raiu map
of Bohemia, showing by isohyetal lines seven rain-belts or zones; the zones are
arranged so that the lowest shows less than 500 mm. rainfall, the three following
difler by 100 mm. each, the fifth and sixth by 200 mm., and the seventh by 300 mm. ;
the last showing, therefore, a rainfall of 1,200 to 1,.500 mm.

The central basin divides itself into two halves by a line from north to south,
running somewhat east of the middle Moldau, crossing the Elbe near the mouth of
the Iser. and following the latter river; the western half showing the smaller amount
of precipitati<m, namely, 500 to 600 mm. ; the eastern with 600 to 700 mm., continuing
in a small belt along the foot of the Erzgebirge and the Boehmerwald encircling the

first zone.

The other isohyetal lines do not embrace continuous areas, but follow in small
belts the trend of the mountains. The largest amounts of precipitation are found in
belts or islands in the higher altitudes of the mountains which surround this great
basin. The continuity of the zones is much interrupted, so that it would be difficult
to describe them without a map.

The maximum of rainfall, with over 1,200 mm., is observed in the south, near the
sources of the Moldau and Motawa; in the north, near the sources of the Elbe, Iser,
and Aupa. on the range of the Schneekoppe. In regard to the distribution through
the months, the experience has been confirmed that with increasing absolute height
the winter precipitation increases in greater proportion than the summer precipita-
tion, while those of spring and autumn are nearly equal.

SuflScient material was on hand from which to calculate the influence of altitude
on the increase of precipitations, although for altitudes above .500 m. the material
is not considered sufficiently reliable. Yet the general law is well shown that with
the altitude the quantities of precipitation increase in a retarded progression.
This progression is calculated by forming altitude zones from 100 to 1000 m., group-
ing the stations in each, calculating the mean elevation and also the mean annual
precipitation as observed for each class; then by dividing the difference of precipi-
tation in the neighboring two zones by the difference of altitude the amount of
precipitation which corresponds to each 1 m. elevation within that class is found.
With this figure the average amount of rainfall which theoretically belongs to each
station according to its absolute elevation can be approximated by adding to or sub-
tracting from the mean precipitations of the class as many times this amount as the
actual altitude diff'ers from the meau.

An example will make this clear : Tetschen, for instance, is situated 150 m. above
sea level. According to the table the average elevation of thirteen stations of the
lowest zone, to which Tetschen belongs, is 182 m., with an average precipitation of
506 mm. Now, Tetschen has an elevation 32 m. lower than the average; its normal
rainfall should therefore be 32 X 0.79 = 25.4 mm. less than the mean of the class;
hence, theoretically, according to its altitude, the quantity of rainfall for Tetschen
should be 506 — 25.4 = 480 mm. ; that is 248 mm. less than that actually found in an
eight years' average. By using the figures for the two extreme zones and dividing
by 100 the mean increase of precipitation for every 100 m. elevation is found to be

69 mm.

And now comes the application of this method to our proposition. The author
argues that if thp actually observed rainfall differs considerably from the theoret-
ically determined, this is ^,^ indication tbat special agencies are at work. He finds
pow that of the pne hundred a^d eighty-six stations which he subjects to scrutiny
(these offering the Jongest »pd mpst trustworthy observation), forty-eight show a
considerable difference between thp observed and the theoretically expected rain-
fall, and he finds also that these §t)a|;ipps ^^-e ^ituatpd iu the most densely wpode^
portions of ^hp Kingdom.

128

FOREST INFLUENCES.

The increased rainfall on the forty-eight stations is so considerable that a suffi-
cient quantity may without losing significance be ascribed to other local causes, as,
for instance, height and form of a mountain range in front or back, etc. Besides,
the greater amounts of rainfall at these stations have been used in calculating the
averages for the altitude zones, magnifying, therefore, these averages so that the
actually observed oue appears really smaller than if the quantities from deforested
and forest areas are compared.

Expressed in percentages of the height of precipitation, an increased rainfall is
shown for several localities in very large quantities, which will allow considerable
reductions for other influences w^ithout losing their significance for the main proposi-
tion. Especially important appears the fact relating to two st-ations near the rain
minimum, w^hichalso shows this influence of the forest.

Lastly, as a matter of interest I may state that the water balance is drawn for the
whole Kingdom, which is of special value, because the political boundaries coincide
with those of the upper Elbe watershed; therefore it is easy to determine how
much of the yearly rainfall is removed by the natural water courses. According to
the calculations made for the various zones by addition, the total precipitation
upon the area of 51,955.98 square kilometers (about 20,000, square miles) of the
Kingdom is found to be 35,398,670 cubic meters, of which the Elbe carries about one-
fourth, or 8,849,667 c. m., to the sea. This figure represents a mean rainfall for the
whole country of 681 mm., while the mean of observation is 693 mm.

RESULTS OF OBSERVATIONS IN INDIA AND BRAZU>.

The following" interestiug observations show the remarkable in-
fluence of forest areas in troiiical and subtropical countries (Annual
Eeport, 1889). Conditions in India are exhibited in the following table :

Ivflvence of forest areas on rainfall in India.

a

o
o to

2 ®

p

Mean temperature.

2*

S ci
*^ y.

Relative humid-
ity.

Precipitation.

Name of place.

P.

^
S

o
a

1-5

1-5

p.

36

60

1-5

<

^
a

6

a

Woodless country :

Lucknow

Benares

Patna

Ems.
847 30. 1
590 30. 2
445 31). 3

33.3

33.2
31.4
30.1

25.9
25.3

33.1

32.8
31.4
29.2

26.9

28.2

30.4
29.7
29.2
28.7

27.7
28.5

45.8
45.0
44.6
44.1

35.1
35.6

30
41

54
81

74
82

Cms.
0.5
0.5
1.0
5.6

14.8
25.9

Cms.
1.8
1.3
2.5

10.1

33.6
30.8

Cms.
13.3
12.9
16.9
24.2

64.3
39.5

Cms.
.39.4
32.4
27.8

Barharapnr

Wooded country :

Goalpara

Sibsagar

270

427
555

29.6

25.2
23.5

52

66

81

60

77

75

85
83

79

84
83

25.8

50.0
40.6

* Mean of two years.

A glance at this table will show that the presence of ivoodn has a far greater inftuoire
in mitigating the tcmperatnre during the hot and dry months of April and Mail than the
proximity of the sea. The same is true of the relative humidity, especially at Sibsagar,
{. e., in the niiddle'of the forests. Most striking is the effect of the presence of woods in
the diminution of the extreme maxima. The greater or less proximity of the sea has but
little effect, but as soon as tve reach the wooded region the extreme maximum falls 9 de-
grees. Thus in 1875 the maximum thermometer did notrise above 35.3 degrees at
Goalpara, while at Lucknow there was not a single day from March 14 to June 22
on which a higher temperature had not been observed. The great humidity of the
air even during the hot and dry months of April and May is the cause why, in the
forests, the raius begin early in March and gradually increase in intensity until

INFLUENCE ON TEMPERATURE.

129

June or July, while in the woodless plains of the Ganges the amount of rainfall
snclclenly increases from May to June or from June to July.

The Amazon hasin. — At the present time there are in the basin of the Amazon four
stations where observations are made ; this river basin is the most extensive forest
re"-ion on the earth. The middle and upper portion of the course of the Amazon is
over 1,000 kilometers distant from the Atlantic Ocean, while it is separated, by
mountains from the Pacific. Were it not for the forests we ought to expect, at this
distance from the sea and. so near the equator, very high temperatures and great
dryness. The following table shows the results of the observations:

Difference of temperatures of four stations in the hasin of the Amazon.

Name of station.

Para*

Manaos

Iquitos

Peruarabiico*

San Antonio on Madeira Kiver

Height

above

sea.

South
latitude.

Meters.

37
95

Degrees.
3

^

8
9

Distance

from
Atlantic.

Eiloms.

100

1,150

2,100

0

1,750

Temperature.

Annual
mean.

27.0
*26. 1
24.8
25.7
26.0

Mean of
hottest
month.

27.7
27.0
25.7
27.1
27.0

Extreme
maxima.

*35. 7
32.4
31.7

Relative

humidity

for the

year.

*80
83
72

* Ten months, from October to July.

t Pernambiico doe.s not beloiijr to the Amazon basin; its means are only given for comparison with
those of San iutoiiio. The shore line near Periiambuco is wooded, but a certain distance around the
city the forests are cut down to give way to fields and sugar-cane plantations.

Thus, owing to the vast primeval forests on the Upper Amazon and its tributaries,
the temiierature of the hottest month and the extreme maximum are not greater than
on the seacoast, and the extreme maximum is far from reaching the values some-
times observed in middle latitudes. It is also to be observed that there are few re-
gions on the earth where the "Trades" blow with such violence as on the coasts of
northern Brazil. Peruambuco is therefore subject not only to the influence of the
sea but also to that of a furious trade wind. Along the lower course of the Ama-
zon the " Trade" also blows with great force; but as soon as we turn into the side
valley of one of the tributaries running in a southerly and northerly direction calm
weather is found to prevail. The height and density of the forest arrests the wind.
There can be no doubt that the vast tracts of forest land on the Amazon, contribut-
iu'if to maintain the moisture of the air and weaken its motion, increase the amount
of waterfall. At Iquitos 284 centimeters fall in the course of the year. It must be
remembered that Iquitos lies in a plain 2, 100 kilometers from the ocean and 350 from
the mountains. Nowhere on the earth is the rainfall so great under similar circum-
stances.

SUPPOSED INFLUENCE OF FORESTS UPON HAILSTORMS.

In this coimection the chiimed iuflnence of forest areas upon hail-
storms may be discussed. While in France the conclusion was reached
that such influence existed, it is doubted by A. Biihler (Influence of
topography aud forests on frequency of hailstorms, 1890), who discusses
this question on the strength of observations over the whole Kingdom
of Wurtemberg through sixty years (1828-'87) in great detail. An in-
fluence either upon the frequency, intensity, or course of hailstorms, it
seems, could not be estabhshed.

There is, however, in these statistics and the conclusions to be de-
rived therefrom one factor, that must not be overlooked, and which the

12444— No. 7 9

130 FOREST INFLUENCES.

able author does not fail to observe. It is the fact that these obser-
vations refer to a nioiintainous region throughout. That means tliat the
influence which the forest areas might exercise in the distribution of
hailstorms is vitiated or wiped out by the more potent influence of
slope and elevation. It is, therefore, impossible to study the isolated
forest influence in such regions where other known but quantitatively
undetermined disturbing influences must be discounted. For such
studies the plain otters a more satisfactory held. The forest on hilly
or mountainous terrain is never effective by itself; it could, therefore,
here only be the question whether it plays a determinative or only a
subordinate part. The forest besides is not like hill and mountain, a
constant and evenly effective factor, since its condition and in most
cases its area is constantly changing. The author also recognizes the
fact that his material does not permit the discussion of the practically im-
portant question, whether and why, as is often observed, hailclouds will
pass by single fields or confined areas within the general route of the
storm. The careful observations of the route and extent of a hailstorm
on June G, 1891, in Thurgau, Switzerland, recorded in Meteorologische
Zeitschrift, 189 L, p. 403, would also show that at least in this case, in
mountainous regions and with extraordinarily violent hailstorms, no
effect of forest distribution was noticeable. The writer would recall
here a case of his own observation on the North German plain, where
the farm of one of his relatives of 5,000 acres is on three sides sur-
rounded by forest areas of considerable extent. The owner for a long
series of years has had no damage from hail, while his neighbor, only
one mile outside of this forest-inclosed district, reaps the benefit of his
hail-insurance money every other year.

INFLUENCE OF FORESTS UPON THE DISPOSITION OF THE WATER

SUPPLY.

(1) ELEMENTS OF DISSIPATION.

In analyzing the relation of forests to the conservation of the water
supplies, we shall examine first the factors of dissipation, or those
Avhich diminish the available supply. They are represented in the
quantity of water which is prevented by interception from reaching
the ground, in the quantity dissipated by evaporation, in the quan-
tity used by jilants in their growth, and in transpiration during the
process of growing.

IntercepHon.

The amount of rainfall and snow which is prevented by a forest
growth from reaching the soil varies considerably according to the
nature of the precipitation and the kind of trees which form the forest
as well as the density and age of the growth.

A light drizzling rain of short duration may be almost entirely inter-

INTERCEPTION.

131

cepted by tlie foliage and at once retnrned to the atmosphere by evap-
oration ; if, however, the rain continues, although fine, the water will
run off at last from the foliage and along the trunks. And this amount,
of which the rain-gauge takes no acconnt, rei)resents, according to
measurements from the Austrian stations, from 8 to 14 per cent, thus
reducing considerably the loss to the soil l>y interception.

While the careful measurements at the Swiss statious in a twelve
years' average show the interception in a larch forest as 15 per cent, in
a spruce forest 23 per cent, in a beech growth 10 per cent, the flgures
for the Prussian stations are for beecli growth 24: per cent, for spruce
at various stations 22 per cent, 27 per cent, and 34 per cent, respectively.
Altogether, for the rainfall conditions of the countries cited, a dense
forest growth will, on the average, intercept 23 per cent of the precipi-
tation ; but if allowance be made for the water running down the trunks,
this loss is reduced to not more than 12 ijer cent.* (See page 100 of this
bulletin.)

According to A. Matthieu's observations during eleven years at Oinq-
Franchees (Meteorologie comparee agricole et forestiere, 1878), only 8.5
X)er cent were retained by crowns (5.84 per cent in winter, 11 per cent
in summer). The more exact observations of Eiegler are compiled as
follows :

Species.

Beech.
O-ik...
Maple
Spruce

Bain on

Clown.

Liter.
20, 021
24, 273
30, 901
12, 014

Fell

through

the crown

upon soil.

Liter.

17,008

17, 873

26, 384

4,793

Run off
trunk.

Liter.

3,343

1,387

2,198

165

Reached
soil alto-
gether.

Liter.
20. 411
19, 260

28, 582
4,958

Lost by

evapor.a-

tion.

Per cent.
21.8
20.7
22.5
58.8

Percent-
age of rain
reaching
soil alto-
gether,

78.2
79.3
77.5
41.2

The results ibr spruce become vitiated by the fact that large aiuounts
of water run off from the tips of the branches, which have an inclina-
tion away from the trunk, and which could not be measured.

The intensity of the preci]»itatiou and the condition of folinge have
much to do with the amounts reaching the soil, so tliat sometimes prac-
tically all the rain reaches it and sometimes hardly any.

The amount of interception in the open growths which characterize

*The maxinmui rainfall ol>seive<l in rJermany is 4 iucbes in twenty-four hours and
2 iiiflie8 in one hour. In iSwit/A-rland there has bcou recorded a rainfall of 18 inches
in twenty-four hours, and 2k inches in three-iiuarters of an hour. This would equal
5, (MM) frallous per acre. Of such falls the foliage will retain only an inappreciable
amount. Intensity of rainfall in the United .States becomes clear from a few records :
Paterson. N. J., li inches in eight minutes; Sandy Spring, Md., 5 inches in two and
one-half hours; Clear Creek, Nebr., 4,50 inches in one hour and twenty-seven min-
utes; Castroville, Tex., 5.80 inches in twenty-four hotirs; Ellsworth, N. C., 13 inches,
of which 9 inches in three and one-half hours; and rainfalls from IJ to 4 inches per
twenty-four hours are quite frequently reiiorted in almost every month, especially in
the Western States.

132 FOREST INFLUENCES.

many of our western forest areas would be considerably smaller, es-
pecially as the rains usually fall with great force, and much of the pre-
cipitation is in the form of snow. Although branches and foliage catch
a goodly amount of this the winds usually shake it down, and conse-
quently but very little snow is lost to the ground by interception of the
foliage.

There is also a certaiji amount of water intercepted by the soil cover
and held back by the soil itself, which must be saturated before any of
it can run off or drain away. This amount, a part of which is eventually
dissipated by evaporation and transpiration, depends, of course, upon
the nature of the soil and its cover, especially upon their capacity to
absorb and retain water.

Altogether an appreciable amount of the precipitation does not run
off or drain through the forest cover but is retained by it; yet while this
is apparently a loss, we shall see further on that this moisture retained
in the upper strata fulfills an important office in checking a much greater
loss due to evaporation, and thus becomes an element of conservation.

Evaporation.

The loss by evaporation, after the water has reached the ground, de-
pends in the first place upon the amount of direct insolation of the soil,
and hence its temperature, which again influences the temperature of
the air. The nature of the soil cover, the absolute amount of moisture
in the atmosphere, and the circulation of the air are also factors deter-
mining the rate of evaporation.

A considerable amount of experimental data is available showing
the rate and total quantity of evaporation in different climates and at
different seasons. Eecent observations made by the U. S. Signal
Service and the Geological Survey show that the evaporation from a
water surface on the western plains and plateaus of the United States
may amount to from 50 to 80 inches, and in special localities to even
100 inches, in a year, while the rainfall (diminishing in reverse ratio)
over this area is from 30 to 12 inches and less. Thus in Denver, where
the maximum annual precipitation may reach 20 inches, the estimated
evaporation from a water surface during one year was 69 inches.

These experiments, made by the Signal Service and the Geological
Survey, have thus far referred only to the rate of evaporation from
water surfaces; the much more difficult question of the rate of evapora-
tion from soil surfaces has been the subject of tentative experiment,
but as yet without such results as can here be generalized.

If the loss by evaporation from an open field be compared Avith that
of a forest- covered ground, it will, as a matter of course, be found to
be less in the latter case, for the shade not only reduces the influence
of the sun upon the soil, but also keeps the air under its cover relatively
moister, therefore less capable of absorbing moisture from the soil by
evaporation. In addition, the circulation of the air is impeded be-

EVAPORATION. 133

tweeu tlie trunks, and this intlueucc npon available water supply, the
ivind-hrealc'mg poicer of the forest, must be considered as among the
most important factors of water preservation. Especially is this the
case on the Western plains and on those Western mountain ranges bear-
ing only a scattered tree growth and Avhere, therefore, the influence of
shade is but nominal.

The evaporation under the influence of the wind is dependent not
only on its dryness, but also on its velocity, which being impeded, the
rate of evaporation is reduced.

Interesting experiments for the purpose of ascertaining the changes
in the rate of evaporation elfected by the velocity of the wind were
made by the Signal Service in 1S87. The result of these experiments
(made with Piche's hygrometers whirled around on an arm 28 feet in
length, the results of which were compared with those from a tin dish
containing 40 cubic centimeters of water exposed under shelter) show
that with the temperature of the air at 84 degrees and a relative hu-
midity of 50 per cent., evaporation at 5 miles an hour was 2.2 times
greater than in a calm; at 10 miles, 3.8; at 15 miles, 4.9; at 20 miles,
5.7; at 25 miles, 6.1, and at 30 miles the wind would evaporate 0.3
times as much water as a calm atmosphere of the same temx)erature
and humidity.

Now, if it is considered that the average velocity of the winds which
constantly sweep the Western subarid and arid plains is from 10 to 15
miles, not rarely attaining a maximum of 50 and more miles, the cause
of the aridity is not far to seek and the function of the timber-belt or
even simple wind-break can be readily appreciated.

What the possibilities of evaporation from hot and dry winds may
be, can be learned from statements regarding the "Foehn," which is
the hot wind of Switzerland, corresponding to the "chinook" of our
Western country.

The change in temperature from the normal, experienced under the
influence of the Foehn, has been noted as from 28° to 31"^ F. and a re-
duction of relative humidity of 58 per cent. A Foehn of twelve hours'
duration has been known to "eat up" entirely a snow cover of 2^ feet
deep.

In Denver a chinook has been known to induce a rise in temperature,
of 57° F. in twenty-four hours (of which 36° in five minutes) while the
relative humidity sank from 100 to 21 per cent.

The degree of forest influence upon rate of evaporation by breaking
the force of winds is dependent upon the extent and density of the
forest, and especially on the height of the trees. For according to
an elementary law of mechanics the influence which breaks the force
of the wind is felt at a considerable elevation above the trees. This
can be practically demonstrated by passing along a timber j^lantation
on the wind-swept plains. Even a thin stand of young trees not
higher than 5 feet will absolutely calm the air to a considerable dis-

134

FOREST INFLUENCES.

taiice and lieiglit beyond tlie slielter. Unfortunately no accurate ex-
perimental data concerning this influence are at hand. According to
Becquerel, asiini)le hedge Ofeet in height will give j)rotection for a dis-
tance of 70 feet; and, according to Hardy, a belt of trees every 300 feet
will defend vegetation almost entirely against the action of the wind.
A.nother authority finds for every foot in height 1 rod in distance pro-
tected. (See also p. 14.9 of this bulletin.)

From many reports received in this office on the effects of windbreaks
upon agricultural crops, we may quote one from a farmer in Illinois,
which expresses the observations generally made:

My experience, is that now in cold and stormy winters wheat protected by timbet
belts yields full crops, while fields not protected yield only one-third of a crop.
Twenty-five or thirty years ago we never had any wheat killed by winter frost, and
every year a full crop of peaches, which is now very rare. At that time we had
plenty of timber around our fields and orchards, now cleared away.

The damage done to crops by the cold, dry winter winds is mainly
due to rapid evaporation, and plants are liable to suffer as much by
winter as by summer drought.

This is certain, that since summer and winter drought, i. e., rapid
evaporation, due to the continuous dry winds, is the bane of the farmer
on the plains, rationally disposed timber belts alone will do much to
increase available water supply by reducing evaporation.

How the forest cover, and especially the litter of a well-kept forest,
may decrease the amount of evaporation within the forest to nearly
seven-eighths of that in the open has been discussed on page 99 of this
bulletin. The reason for this important influence of the forest is due
not only to the impeded air circulation, but also to the temperature and
moisture conditions of the forest air and forest soil.

The stations of Prussia allow the following average for evaporation ;
the amount evaporated in the open fallow fleld being called 100:

Evapo-
rated.

Retained
more than

in open
fallow field

TTndftr lipprli o'rnwtU .. ..........

Per cent.
40.4
45.3
41.8
90.3

• Per cent.
59.6
54.7

58.2
9.7

TTiiflftr sTtriifMi PTowtli ..............

ITnflpr ninp « ro wtli . ............

From cultivated fio-lil .............

A balance calculation of the amounts of precipitation and the amounts
lost by evaporation for sixteen stations at varying elevations shows
that with increasing altitude the suri)lus of water remaining for the soil
increases, the mountain forest decreasing evaporation to its minimum
of 9 to 13 per cent, and leaving from 87 to 91 per cent to penetrate the
soil.*

* This is a larger amount than that given on page 101, where the water running
off the trunks of trees is counted as lost.

EVAPORATION.

135

Stations.

Altitude.

Surplus of i>recip-

itation over evap-

oratiou.

Of percipita-
tiou evapor-
ated.

Tn the
open.

In the
forest.

In the
open.

Tn tho
forest.

M.
3

30
34
42
95

Mm.

322. 5
387.5
495.8
142. 1
174.6

Mm.
343.6
322.5
481.4
237. 5
180.6

Per ct.
55
40
35
73
70

Per ct.

28

28

20

44

67

37

Average for the

0-lCiO

305.3

313.1

55

124
143
145

340.1
184.9
436.1

365.7
254.7
434. 3

44
68
46

26

37

20

Average for the

100-200

322.4

351.6

53 30

1 .

340
353

328. 5
291.0

510.9

385.8

60 1 23

57 26

region ^

Average for the

300-400

309.9

448.3

58 25

T.nliiilinf

602
612

G80
690

850
717. 5
1,408.2
718.8

685.2

490. 2

1,114.3

839.1

24
26
13

27

15

TT..ll,.,.„f U

21

Sclnoit'defcld

7
10

resfion .,..,■...-..,..-.

600-700

938.7

782.2

22 1 13

774
930

1, 190. 4
1, 142. 1

1, 093. 8
1,196.8

15

19

9

11

K

Altogetlier, it will have to be admitted that the factor of dissipation
represented in the evaporation from the gronnd is eon.siderably rednced
by the forest-cover; and since the rate of evaporation in onr Western
Territories is probably the greatest element in the dissipation of mois-
tnre, the greatest attention to checking it will be necessary in the hus-
banding of water supplies. This check to evaporation refers not only
to the preservation of the water supply where it falls, but also in the
natural and artificial channels through which it may be conducted or
in the reservoirs where it may be stored.

The surface exposed determines the amount of evaporation from
water-courses and reservoirs; but if the amount evaporated is related
to the available volume of water, it wOl appear that the smaller and
slower run loses proportionately more than the larger, which thus ex-
hibits the value and protective (character of accumulation.

Take a brook G feet in width and only a foot in depth ; this for a
length of 30 feet would contain 180 cubic feet of water. If from this
surface only one-tenth of 1 inch evaporates, the amount eva]>orated is
eqnal to 1.5 cubic feet or jl^ of the entire snpply. On the other hand
one-tenth of 1 inch evaporation from a river 00 feet broad and 12 feet
deep for a length of 30 feet, containing therefore 21,000 cubic feet of
water, would bring the loss to 15 cubic feet or only jtr o of the avail-
able supplies; the loss, in proportion to the supply, being twelve times
greater in the former case.

13G FOREST INFLUENCES.

Trauf^piralion.

All vegetatiou takes up a certain amount of water, a part of wbicli
is consumed in building- up its body, and a still larger part returned to
the atmospliere by transpiration during the process of growth.

The factor of dissipation having been fully discussed on pp. 9G, 130 of
this bulletin, it need not be further considered here, except to recall the
conclusion that forest growth transpu^es considerably less than other
kinds of vegetatiou.

Since this water is given off again to the atmosphere in the locality
where it has fallen — thus enriching the atmospheric moisture — and is
therefore only diverted temporarily for the i^urpose of doing diity in
producing useful substance and retaining it in the locality where it has
fallen for a longer time, transpiration may even be considered as an-
element of conserA^ation.

There is still to be considered a certain amount of moisture which is
retained and stored up in the body of the plant, partly as a necessary
permanent constituent, partly as a temporary constituent, being evap-
orated when the plant dies or the wood is seasoned. The amounts thus
retained vary considerably according to age, cax)acity for transpiration,
site, soil, climate, density, slow or rapid growth, weather, seasons, and
even the time of the day. It is therefore almost impossible to give
anything but very rough approximations, especially as also the differ-
ent parts of the tree vary considerably in the amounts of water
present.

The wafer which euters into chemical composition of the wood substance repre-
sents about 50 per cent of the weight of dry substance.

The water hygroscopically retained in the living tree varies within the wide range
of from 18.6 to 51.8 per cent in the wood, while the leaves contain as much as 54 to
65, and some even over 70 per cent while living; when dry, still 10 to 12 per cent.
The wood of deciduous hard woods, like oak, ash, elm, birch, beech, contain in the
average 38 to 45 per cent; soft deciduous trees 45 to 55 per cent, and the conifers 52
to 65 per cent. White pine when young may show as high as 77 per cent of its
weight as water, while larch, of all conifers, has the smallest water capacity,
namely, 45 to 55 per cent, ranking with the deciduous soft woods.

This hygroscopic water is reduced by seasoning to 10 or 12 per cent ; this amount
being retained even in well seasoned woods.

Given the entire mass of wood and foliage on an acre of forest, an
approximative calculation of the total quantity of water contained in
the trees will show that 56 to 60 per cent of the weight of the forest
must be attributed to water, while only 44 to 40 per cent is repre-
sented by dry substance. In agricultural crops it is known that the
amounts of water are still larger, reaching sometimes 95 per cent of
the whole weight. The production of dry substance in a well-kept
dense timber forest may amount annually to from 2,500 to 3,000 pounds
per acre, leaving, then, for the hygroscopic water, 3,750 pounds, and
the chemically fixed water, say, 1,250 pounds ; so that for this factor
of dissipation 5,000 pounds in round numbers as a maximum will
suffice.

TRANSPIRATION — SNOW. 137

ELEMENTS OF CONSERVATION.

In discussing tbe elements of dissipation as to tlie degree of their
effect under forest-cover, compared with the same elements at work
in the open field, we have seen that the shade, the low temperature,
the relative humidity, the absence of violent air-currents, the water
capacity of the forest floor, are all acting as factors of conservation.
We have seen that the quantity of water lost by evaporation— the
most fruitful source of dissipation— may be more than six times as
great in the open as in the forest. There is ouly one other element of
conservation affecting water supplies which requires special mention.
This is the retardation in the melting of the snow, which is due to
forest-cover. According to Dr. Biihler, of Zurich, this retardation in
Switzerland amounts to from five to eight days in general, and may,
according to weather conditions, be several weeks, thus giving a longer
period for distribution. The evergreen coniferous forest in this respect
naturally does better service than the deciduous one.

Effect of forests in case of snoiv.

Snow will lie in the forest more evenly and continuously than on the
open, wind-swept areas. Thereby not only the amount finally remain-
ing for drainage is increased, but the soil is prevented from freezing,
and is kept open for percolation when the snows melt. The retarda-
tion of the melting has been determind by Biihler in Switzerland to be
from eight to fourteen days.

Mr. R. U. Piper, in his Trees of America, states that an unobstructed
warm wind will dissolve the snow more than ten times as fast as when
it is protected from the wind, the temperature being the same, and he
adduces in verification of his statement the following experiments tried
by himself: In the first, a body of snow 1 foot in depth, protected from
the wind, but partially exposed to the sun, after a thaw of two weeks,
was not wholly melted, while another mass 6 feet in depth, more shel-
tered from the sun, but fully exposed to the wind, was melted in less
than a week.

In the second, equal quantities of snow were placed in vessels of the
same kind and size and exposed to the same temperature, one being
covered and the other having a current of air constantly passing over
it. The snow in the latter vessel was melted in sixteen minutes; that
in the former was not entirely dissolved at the end of eighty-five min-
utes.

In the third experiment, in a room with the temperature above
80O, the mercury in a thermometer rose from 32° to 80° when exposed
to a warm current created by a fan, or seven times as last as when the
heated air was still.

The conservative effect of the forest-cover is especially of value on
the western mountain ranges, which are liable to be swept by the

138 FOREST INFLUENCES.

cbinook, dissipating as if by magic the snew-cover over wliicli it
sweeps. (See p. 133.)

Even without this specially dry wind it is well known in Colorado
and other luountain districts that the regular wind sweeping over the
bare slopes above timber line, or where the forest-cover is removed,
while they drift the snow hither and thither, they " wear it out" at the
same time. By the blowing of the wind the snow is reduced to finest
particles, and, by the shifting, new surfaces are constantly exposed —
two processes which greatly fticilitate evaporation, and thus the snow
is literally worn out.

The proposition, then, to remove the forest-cover in order to allow
the drifting and compacting of the snow, from which possibly to secure
a longer period of distribution, even if there were no other objection,
must be considered a hazardous and ill advised expedient.

The influence of the forest upon the condition and driftingof the snow
is g-raphically related by Middendortf in his description of Siberia,
speaking of the Bnran or snowstorm characteristic of the treeless
plains of tundras:

As far as tlie forest readies ami impedes tie action of the winds the snow lies
everywhere evenly and loosely, so that in the beginning of winter one can travel
only on snowshoes. As soon as the tundra is reached there is no need of snowshoes.
The snow lies either like a thin carpet, or drifted together in incredible masses, so
compacted as to bear man and beast, etc.

General effect of forests in reducing evaporation.

The popuhir notion which ascribes to the moss-cover or spongy char-
acter of the forest floor a conservative function beyond that of retard-
ing evaporation and expediting infiltration seems to be entirely errone-
ous and needs revision. The idea that the moisture of the soil and the
flow of springs is increased by water from the spongy cover is alto-
gether in contradiction to physical laws, and can be shown experimen-
tally to be a mistaken one.

Water filters through the cover by the law of gravitation until the
whole mass has become fully saturated. With an addition of water it
will filter through to the soil, as long as the supply continues and as
long as the soil is not so saturated that it can not readily absorb any
more water. At last, the supply continuing, the cover will refuse to
convey it and will shed it superficially, leaving opportunity to reach
the soil only where the moss cover is interrupted. When the water
supply ceases, evaporation begins above, and by capillary attraction
the cover supplies its loss of water on the surface from the soil below.

To give water to the strata below, it would be necessary that these
should have become dry, or at least drier than the moss-cover before the
latter had lost its water. This may occur and depends naturally upon
the structure and nature of the soil. If the soil is strongly fissured,
thus rapidly draining the upper strata, then, if the moss-cover is still

GENERAL EFFECT ON EVArORATION. 139

saturated and an additional pressure is exerted by water standing or
falling- on it, a further supply of water may be given up to the soil; if,
however, the moss is only just saturated and no further access of water
takes place from above, then there is no physical law by which a sur-
render of this saturation water to the soil could take place as long as
the underlying soil is of a gravelly or nonabsorbing nature. If its
nature is like clay, marl, flue sand, capable of attracting water, then the
further process of water absorption depends upon the difference be-
tween the water capacity of the cover and that of the soil.

In a sand soil in which the upper strata lose their water rapidly to
the lower, the moss-cover, which holds water more tenaciously, can be
made to give up water to the soil as long as the capacity for absorption
by the sand is greater than the capacity for retention by the moss.

A loam or clay soil takes up water very slowly, but takes up a
great deal before it is saturated, and the process of filtration goes on
very slowly; if, therefore, a plentiful rain falls, there is formed
under the moss-cover a shallow, nearly saturated layer of soil, which
acts as an impermeable stratum. This layer is protected by the cover
against rapid surface drying, and since it gives up its water only slowly
to the lower strata, it remains moist so long as the moss-cover is not
dry. As soon as by evaporation the cover has lost its water, the soil
must give up some of its moisture by capillary attraction to supply the
deficiency in the cover. It must be noted here, however, that accord-
ing to Oltman's experiments, moss does not take up water from an only
moderately moist soil. A deficiency of moisture occuring in soil earlier
than in the cover can be presumed only when the water is utilized by
the roots and transpired, which is not likely to occur.

These are the extreme cases between which in nature many interme-
diary conditions occur. The litter cover does not act analogously to
the moss-cover or to a sponge. A difference must here be noted between
the newly fallen loose litter of the previous year and the closely packed
and felted litter accumulations of former years. The former allows a
rapid filtration ; the latter, according to Eiegler's experiments, is nearly
impermeable, and the water practically can enter the soil only where
the litter is intermitted. The compacted litter serves admirably to re-
tard evaporation. In reality there rarely exists an uninterrupted cover
of such litter or a cover of one uniform nature ; open spaces, moss-covers,
varying thicknesses of litter-cover interchange, and accordingly the
water penetrates readily, while the cover performs its duty as a conserv-
ing agent against evaporation.

There is an additional conservative action of the forest floor to be
noted, which will be more fully dis(?ussed farther on, as an influence
upon the distribution of the run-off. It is the mechanical protection
which the cover affords against the compacting of the soil by the fall-
ing raindrops; by this protection the soil is kept porous, permits ready
percolation, and therefore less water remains at the surface to fall a
prey to evaporation.

140 FOREST INFLUENCES.

It is, tlieu, the ijrotccHon against evapnraiion alone, duo to greater
relative huniidity of the forest air, to the shade, to the breaking of
the winds, and especially to the ])rotective soil cover, which males the
forest a conservator of moisture evcri/where, even where it does not by
its peculiar location increase the amount of precipitation.

Springs, then, may be influenced in the amount of their discharge by
a removal of the forest, not because the forest sup])lies them directly
with more water, but because by its removal the rate of evaporation is
increased.

SUMMATION OF THE CONSERVATIVE AND DISSIPATIVE INFLUENCES.

The total conservative action of the forest with reference to available
water supplies, aside from an increase of precipitation, is expressed by
the difierence between the elements of dissipation and those of conser-
vation; the former comprised in the loss of the water by retention or
interception, evaporation, and transpiration, the latter in the protec-
tion against evaporation. This balance is known to be in favor of the
forest cover in some localities and under certain given conditions; but it
will have become apparent that a general statement or quantitative
expression of the amount of benefit is in general quite impossible. Yet
in an ingenious manner a calculation for one of the Prussian mountain
districts is proposed by Dr. Weber as follows: Using the figures which
are exhibited in the table on page 113 he argues that the amount of
water left, over and above the amount evaporated in the open at low
altitudes, deducted from the amount left over and above evaporation in
the forests of high altitudes, will suffice to cover the amount of trans-
piration; thus, in the spruce forest at the station of Sonnenberg, the
surplus of precipitation above the water needed for evaporation had
been 1,093.8 mm.; deducting from this the quantity which was found
remaining in the open at Schoo, and which would suffice for purposes
of transpiration and plant growth, a balance for drainage of 771.3 mm.
results; for the beech forests at Melkerei and Hadersleben, the calcula-
tion gave a balance of 1,176.8—495.8=681 mm. for drainage. On the
average, therefore, 700 mm. of the precipitation in the mountain forest
in this locality are saved for the " run-off," that is, 100,000 cubic feet of
water per acre.

To get a conception of what these 100,000 cubic feet mean in the river
flow, it may be stated that with average water level the Rhine above
Manheim has a flow of 47,700 cubic feet per second, an amount which
would be yielded by 40,000 acres of mountain forest, provided all water
is drained into the river; and to keep the river continually flowing at
this rate would require, on the basis of these figures obtained experi-
mentally, a forest area of 23,472 square miles, a calculation which by no
means leads to absurd results for practical probability, since the drain-
age area of that part of the river is in reality about 30,000 square miles,
largely in forest.

DISTRIBUTION OF WATER. 141

I also recall here the water balance drawn for the upper Elbe Elver
watershed on p. 128, from which it appears that the river flow repre-
sents about 25 per ceut of the precipitation.

The amount of river flow, to be sure, does not permit a calculation of
the amounts of water originally available (after they have fallen and
been diminished by the factors of dissipation) for local use in both sub-
terranean and surface runs, since the river flow exliibits both kinds of
drainage, but not at the same time. Nor do the wholesale methods
sometimes employed to determine the relation of river flow to precipita-
tion promise a solution of either that question itself nor of the ques-
tion, how far surface conditions of the soil have a bearing upon drain-
age.

A detailed study of smaller and confined catchment basins alone
like those referred to hereafter (see p. 154) will give results that may
eventually lead to i)racticable methods of calculations.

DISTRIBUTION OF TERRESTRIAL WATERS.

The distribution of the available water supply is almost as important
and often a more important factor in the economy of the water than
tlie quantity of available supply itself, and the manner in which this
distribution takes place influences considerably the ultimate availa-
bility of the supply for human use.

In discussing the distribution of the water supply it is desirable to
folloAv the natural division of the waters into superficial and under-
ground drainage. The surface runoff gives to brooks and rivers all
their rapid variations of stage; the underground drainage gives them
their permanent regime.

ANALYSIS OF DETERMINING INFLUENCES.

The luoportionate division of these two classes of run-off, then, is of
the highest possible importance; we will, therefore, analyze the condi-
tions which determine their relative proportion in order to find how
the forest may influence the same. It is evident that the first condi-
tion is to be found in the amount and character of the precipitation.
A violent rainstoim will furnish more superficial run off than Avhen,
the rain falling slowly, time is given for the soil to absorb it.

Eainy and rain-i^oor or arid climates, short and insignificant rains,
short but violent, long and mild, or long, plentiful rains, also periodical,
seasonal rains and irregular raintalls, all these constitute differences
ill the nature and time of occurrence of the rainfall, which must neces-
sarily affect the relative amounts of the run off'. The effect is still
further complicated where the jirecipitation is i)artly snow, when not
only the inass of accumulated su])])ly but also the progress of melting
determine the result of the run-ofl".

We find, therefore, based upon this one factor, namely, the nature
and time of occurrence of iirecix^itation, differences in the run -off' which

142

FOREST INFLUENCES.

are dependent upon differences in climatic conditions. Thus tropical
rivers show one or two regular high stages of water according to
whether they have one or two rainy seasons; in regions of equinoctial
rains a spring and fall freshet is normal, while the rivers may be almost
dry in summer or winter; the frequent thunderstorms in the mountains
of Switzerland produce short but rapid floods during the summer,
Avhile the autumn is characterized by low water in the rivers. This
climatic difference in water flow it is important not to forget when dis-
cussiug the influences which may modify the discharge of Avaters.

W. Ule, in Meteor. Zeitschr., 1800, discusses the relation of water
stages in a stream and outflow in relation to precipitation. He comes
to the result that a direct relation from one to the other is impossible
to find, because of the complication of other conditions, which disturb
this a priori seemingly direct relation.

He finds that amount of precipitation and water stages or even water
stages and amount of water flow are not x)roportionaI, so that with the
same water stage different amounts may run off'. If this discrepancy
appears in the annual mean water flow, it is still more noticeable in the
monthly means. In the river Saale the mean water stage for March,
1886, indicated a flow of 378,000,000 cu. m., while the daily measure-
ments gave 508,000,000, or 34 per cent more. The mean water stage in
March was 2.13 m., in December 2.15 m. Yet the amount of flow in the
latter month was 23 per cent smaller than in the former. With the
stage twice as high it was found that three times the flow resulted.
From these moie detailed measurements it appears that changes in
the amount of river flow are not necessarily due to changes in amount
of precipitation.

In 1886 the amount of river flow was 14 per cent less than in 1884,
although the precipitation was by 1 per cent less in 1884.

The gre:itest influence on river flow is assigned to the distribution
through tlie year of the precipitation, at least in regions with persistent
frost periods, Avhere, as in the Saale catchment basin, the river flow in
Avintei' is tliree times as great as in summer, namely, 51 per cent of the
precipitation in winter against 17.3 per cent in summer. Hence the
annual mean river flow reflects more the Avinter precipitation than those
of summer, as the following figures show :

Tear.

Winter.

Summer.

1882-'83.

]884-'85.

1882-'83.

1884-'85.

1882-'83.

1884-'85.

10, 913
4,476

10, 757
3,182

4,4U

3,340

4,047
2,142

6,572
1,127

6,710

Iliver flow, cubic meters, millions

840

The larger amount of flow in 1882-'83 seems to depend on the fact
that much precipitation fell in winter, and, at the same time, the greater
percentage of flow in that winter is accounted for from the daily meas

DISTRIBUTION OF WATER. 143

urements sbowing repeated high rises, which, as has been shown, predi-
cate disproportionate flow. The winter besides was warm, most of the
precipitation was rain, or, if snow, was qnickly melted and carried off.
We see then that, besides the h^cal and timely distribution of precipita-
tion, temperature, direction and strength of winds, condition of topog-
raphy and soil, other dynamic influences are exhibited in the river flow,
which make this as an immediate expression of condition of precipita-
tion uncertain.

After water has reached the ground its distribution is determined,
first, by the character of the topography and, second, by the nature of
the soil and the surface conditions.

Tlie topography determines the rapidity of run-off and of collection.
The more diversified the country — cut into dells, coves, rills, and fur-
rows, steeper and less steep slopes— the larger the number of runs of
unequal length in which the water is collected, while the less diversified
the contour the more water must be carried off in each run. Yet
where the diversity of configuration is accompanied by steep slopes
the run-oft' nmy be so rapid that the valley river is filled more rapidly
than the river of the open plains country with even slopes of moderate
inclination.

Thus in some of the river valleys of West Virginia the watersheds
are scooped out into such an array of coves, gashes, and water courses
and minor watersheds, and so steep and rapid in descent that, in
spite of the forest cover, a rainfiill of a few days Avill induce a rapid
rise of the rivers, Avhile the same amount of rain will hardly wet the
ground in a prairie couutry like Iowa.

The second of the above-mentioned conditions determining distribu-
tion— the nature of the soil and the surface conditions — comprises a
large number of separate, though related, factors. The composition,
structure, and stratification of the soil itself, its water capacity, its
permeability, and other physical properties; the nature of the under-
lying rock and its susceptibility to disintegration under the action of
erosion ; the surface conditions of the soil cover, whether frozen or
sunbaked, cultivated or uncultivated, barren or covered with grasses
or forests; these are a part of the factors which affect the distribution
of the water supply and determine the proportions of surface and
underground d r a i n a ge.

On a given territory, then, with given geologic, topographic, and cli-
matic conditions, the only directly variable conditions influencing the
manner of di-aiuage are those of the u])per soil stratUi of the soil cover.
We are, then, ma inly concerned with the water capacity ol" soils and
soil covers, the inten.>ity of their water absorption and the amounts of
water which are drained through them in given times. We are inter-
ested in studying by what means the draining capacity of the soil is
increa.sed, and by what means altogether the run-oft" may be changed
in its nature from a superficial to a subterranean one and the reverse.

144

FOREST INFLUENCES.

WATER CAPACITY AND WATER CONDUCTIVITY OF SOILS AND SOIL-COVERS.

We must discern two kinds of water capacity in soils, namely, the
absolute or minimum water capacity and tlie full or maximum water
capacity. The former, also called ''moisture coefficient" in soil analy-
ses, is that water quantity which the soil will retain, no matter what
the drainage conditions, provided the supply is not deticient. It is this
water retained in the capillaries which determines the designation of
the soil as moist, wet, or dry, which latter is only comparatively so. If
the drainage is impeded by an impermeable layer underlying a porous,
permeable surface soil, and if the slope of the impermeable layer does
not favor rapid drainage, then if additional supplies reach the soil, all
interstices and h^rgest capillaries will fill up and the amounts of water
then contained in the soil represent its greatest (full or maximum)
water capacity. Any surplus above this greatest water capacity is
bound to drain ofl" at once either superficially or subterraneously.

Absolute or minivium water capacity of soils according to Von Elenze.
[Volume unit, 3.5 cm. diam. by 10 cm. height.]

Soil.

Sand, loose.... \

Loau), fine, loose \

[

Humose lime sand, loose, fine I

Height
of layer
in centi-
meters.

10

20
.30
40
10
20
30
40
10
20
30
40

Weight of
water ab-
sorbed in
grams.

44.6

68.0

83.8

92.9

43. ,5

76.2

106.8

138. 2

48.2

79. 2

113.1

129.2

Volume.

Per cent.
20.99
24. 04
20. 37
17.58
48.27
47. 59
42. 08
38. 42
49.13
47.37
45.11
43.60

Relation
of weight
to volume.

44.0
34.0
27.2
23.2
43.5
38.1
35. 6
34.5
48.2
39.6
37.7
32.3

Full or maximum ivater capacity of soils according to Meister.
[Volume unit 1 cubic foot].

Soils.

Clay

Loam

Humus

Peat

Garden mold

Lime

Chalk

Sand (82 per cent)
Sand (04 per cent)
Quartz sand

Weight of

water
absorbed.

Povndg.
21.99
26. 40
30.90
27.99
30. 33
. 24.09
21.74
19. 94
28.67
20. 42

Volume.

Per cent.
50.0
60.1
70.3
63.7
69.0
54.9
49.5
45.4
6.5.2
46.4

From the above tables it appears that with the depth of the soil the
minimum water capacity varies greatly, and the same may be expected
for the maximum water capacity. In sand it may sink within 15 inches

WATER CAPACITY OF SOILS. 145

to one-half of what it is at the. sniiace, wbile in loam, owing to smaller
capillaries, the reduction is only one-third. The finer the capillaries
the more water they can keep up proportionately in balance against the
force of gravity, and hence humus and garden mold, with their fine
capillaries, show the greatest water capacity.

The degree of such retention, as has been shown by Prof. C. E. Hil-
gard and others, is also somewhat influenced by the temperature of the
soil. The least retentive soil is a coarse quartz sand followed by finer
sands, and it is increased by an addition of lime loam, or vegetable
matter.

For a "second-class" Florida sand soil the "moisture coefficient" is
stated by Hilgard as 1.64 per cent of its own weight, while it is 23 per
cent and more in a peat soil; a pure clay rarely exceeds 12, while the
moisture coefficient of calcareous clay soils rises to 15 and 20 per cent.
The maximum water capacity may be many times that of the absolute
water capacity, depending on structure and more or less compact strat-
ification of the soil. According to Ebermayer, the amounts of water
held may vary between 3 and 88 per cent. Dr. Raman's investigations
show the maximum water capacity of sand soils of fine and medium
fine texture to be from 3 to 4 per cent.

According to Prof. Schuebler at Tiibingen, who experimented on
soils under natural conditions:

Sand soil may hold 25 per cent of its weight in crater; loamy soil 40 per cent of
its weight in water; clay loam 30 per cent of its weight in water; pure clay soil 75
per cent of its weight in water.

The most impervious soils, as was to be expected, vshowed the greatest
retentive power, and since by compacting the soil particles the capillar-
ity is increased, the imperviousness of such soils is increased.

Impermeable soil strata, such as loam and very fine sand, allow, when
a surface run-oft" is readily possible, only a passing and inferior retention
of water after rainfall, in springtime taking up no more than 10 or 12
per cent of their weight, while a stratum of sand of medium grain, 20
to 25 feet deep, Avas calculated by Kannxn to be capable of taking up
and holding the entire annual precipitation of 24 inches.

The capillarity of sand soils of the North German plain investigated
by Raman was not capable of raising the ground water higher than IJ
feet, so that the upper strata of the .soil, which was within reach of
ground water, did not show any greater amount of water than the soil
which had no ground water to fall back upon.

Mr. F. H. King ( Agric. Expr. Sta. Wise. Ann. Rept., 1889) has investi-
gated the water capacity of soils with special reference to the function of
groundwater in plant production. Six thousand observations of ground
water stages under varying surface conditions show insufliciency of
capillary actions to supply water for plant trans])iration rapidly euough.
He found that a column of 1.5 in height of natural soil (consistiug of
loamy marl, red clay, sandy clay^ and fine sand) could hold an amount
12444— No. 7 10

146

FOREST INFLUENCES.

equal to 54 cm. precipitation, which is equal to three-fifths of the mean
annual precipitation at the station, of which the author presumes not
more than 30.5 cm. are used for an averajie crop of agricultural grains.

Coming- now to the influence which a forest cover may have upon the
water capacity of the soil, Ave have to record the results of Dr. E. Eber-
raayer, who has investigated the water capacity of a heavy loam soil,
both protected and unprotected by a forest cover.

The following table containsthe measured amounts of water contained
in such a soil under a forest of spruce, twenty-five, sixty, and one hun-
dred and twenty years old, and a naked soil at IG inches (40 cm.) and
32 inches (80 cm.) depth.

Water contents of a loamy sand; results hif seasons expressed in percentages of the iveight

of the soil.

Spruce.

Naked soil.

Season.

25 years old.

60 year.s old.

120 years old.

16
inch.

32
inch.

Aver-
age.

16
inch.

32
inch.

Aver-
age.

16
inch.

32
inch.

Aver-
age.

16
inch.

32
inch.

Aver-
ago.

Winter (January and
February)

Spring (March to May)..

Summer (June to Aug iLst)

Fall (September to No-
vember)

20. 23
18.62
15.10

16.57

17.00
18.02
16.22

17.57

18.61
18.32
15.96

17.07

18.06
15.29
14.42

13.49

17.76
16.28
17.03

16.52

17.9]

15.78
15.72

15.00

19.75
17.47
17.78

14.88

22. 44
20. 83
20. 90

19.46

21.09
19.15
19.97

17.17

19.96
20. 66
19.37

20.04

24. 73
20. 51
19.98

20.20

22.35

20.58
19.97

20.12

These figures show that a loam soil under forest cover is apt to be
drier in the depth of the root region, and that at all seasons, than in the
open field at the same depth, less so under an old and scattered growth
than under a younger growth or thicket.

A repetition of these experiments, in which various depths from the
top to 32 inches were included, gave during two years the following
averages of water cai^acity, expressed in percentages of the Aveight of
the soil:

Averages of water capacitg, expressed in percentages of the weight of the soil.

Spruce.

Unshaded
soil.

Depth.

25 years
old.

60 years
old.

120 years
old.

Per cent.
30.93

Per cent.
29.48

Per cent.
40. 32

Per cent.
22. 33

'6to8iii("h(^s

19.19
19.10
18.40
17.91

]8.99
16.07
16.26

17.88

19.30
18.28
20.16
21.11

20. 62

12 to 11 iiirlips .

20.54

19 to 20 inches -

20. 14

30 to 32 iLches

20.54

A verao'ft .

18.65

17.30

10.71

20.46

Ebermayer combines the values for depths from 0 inches down to 32
inches, and then concludes that the forest soil is less moist, due to the
transpiration of Avater by plants. This conclusion is, however, not at
all warranted. For if one combines the figures f\)und in all the strata
from the top to 3 inches down, they figure as follows: Sx^ruce 25 years

WATER CAPACITY OF SOILS.

147

old, 24.79 percent; spruce 60 years old, 23.39 per cent; spruce 120 years
old, 30.01 per cent; naked soil, 22.39 per cent.

Hence, take it altogether, the naked soil contains considerably less
water than the forest-covered soil. But the distribution of the water
through the difterent layers of the soil is different in the two cases; the
naked soil, due to rapid evaporation, no doubt, contains the least
amounts in its upper strata, where the forest soil with its absorptive
coA'cr preserves the largest amount. Measurements of the stratum
from 2 to 6 inches would probably have shown the preservative effect
still more prominently.

The water capacity of soils and soil covers in general has been re-
ferred to as an element of interception. With reference to the run off,
this capacity becomes influential in determining the manner of run-off".
As soon as the soil cover and the upper soil strata are saturated, and
especially when the latter are more or less impermeable and the rain
continues, either uo water or only a small part gradually can find en-
trance into the soil, and the run off" becomes in the main suiierlicial, or,
if the ground be not slo])ing, stagnant water results.

For every watershed, no matter what its conditions, there may, there-
fore, come a time when the rain or snow-melting continuing, the entire
run off becomes superticial, the soil being unable to take up more.

It is evident that this time must occur later in the forest than on
the unfoiested and especially naked soil, because the water capacity
of the soil cover as well as of the protected soil in the forest is greater
than that of the naked soil or that covered with field crops.

The water capacity of litter, which Wollny investigated, depends on
its nature and, of course, its thickness to a certain degree, and is quite
considerable, much greater than that of soils.

The water capacity of various litters was found to be as follows in
volume per ceut:

Depth of litter.

Oak
leaves.

Beech
leaves.

Spruce
litter.

Pine
litter.

Moss.

Calcare-
ous sand
soil.

When 2 inches deep

Wlien 12 inches deep

50. 77
45.42

38. 9.3
41. 65

""""36.' 28'

19.82

39.78

24.93

No soil cover was found so variable in water contents as moss, while
litter would hold two or three times as much water as moss and twice
as much as the soil.

The variation of water capacity at dififerent depths appears from the
following figiues :

Depth of litter.

Oak
leaves.

Spruce
leaves.

Two inches

50. 77
52.99
53.09
45.42

38.98
47.76
41.03
41.65

Kiffht inches

Twelve inches

148 FOREST INFLUENCES.

That is to say, the increase in water capacity ceases with about 8-inch
depth.

The quantity of water which the soil cover can contain appears from
the followiug measurements of Dr. Ebermayer : On the 17th of August,
1885, after rainy weather, the moss cover in a 60-year-old spruce
growth contained 72.33 per cent at the top, 76.64 per cent on the lower
side, and 71.67 per cent in the humus soil beneath.

After a rainstorm lasting one and a half days, on September 9, 1885,
the moss cover contained 80.45 per cent at the top, 74.61 per cent on the
lower side, and 74.42 per cent in the top soil.

In this connection the following note regarding the enormous water
capacity of moss covers in Alaska may be of interest:

In the interior plateau of the Cordillerau and St. Elias regions of Alaska, accord-
ing to Mr. C. W. Hayes, surface degradation is greatly retarded by the luxuriant
growth of moss, which covers practically tbe entire surface of the country. The
annual precipitation is largely confined to the winter mouths, and the water from
the melting snow is held by the sponge-like moss, which remains saturated through-
out the short but hot and dry summer. Thus, with a rainfall which, in lower lati-
tudes, would condition an arid region, a large part of the surface is swampy, quite
irrespective of slope— that is, wherever the material composiug it is sufficiently com-
pact to become impervious to Avator on freezing. On account of this slow and
imperfect surface drainage, the slopes are not cut into the ravines and arroyas so
characteristic of arid regious.

WATER CONDUCTIVITY OF SOILS.

Of still more importance for the runoff than the water capacity is the
water conductivity of the soil, or the rate of water absorption— filtra-
tion.

The rapidity with which the water is conducted from above down-
ward must necessarily influence the nature of the run-off.

Gravity tends to drain the water downward, capillarity to carry it
upward; the difference of these two forces in the main must, besides
the mechanical obstructions of the soil particles, determine the rapidity
of drainage. Experiments to establish the rate under various conditions
are very few and unsatisfactory.

The capillary conduction from below has frequently been made the
subject of investigation, but the downward movement has not yet been
studied with sufiBcient detail, and it has hardly yet been recognized by
the experimenters that this depends upon the difference of gravity and
capillarity as two opposed forces. According to E. Wollny's experi-
ments in 1883 and 1884—

(1) Water is conducted downward the more rapidly the larger the soil particles
(i. €., the less capillary attraction exists).

(2) The noncapillary interstices of the soil accelerate the downward movement of
the water (t. e., the less mechanical obstruction of soil particles).

(3) In granular soil the water penetrates faster than in powdery soil (i. p., pene-
tration is the slower the denser the stratification). It is most rapid in quartz and
slowest in clay; in humus at a rate between these two, but in a mixture of clay soil
and humus faster than the average of the two.

WATER CONDUCTIVITY OF SOILS.

149

(4) Tlie l-apidity of drainage iu a grauular soil i.s indepeudeut of the size of the
grain.

The experiments were made with soils of varying grain in tubes 110 centimeters
deep, the water drojjpiug on top constant] j"; the results ai'c exhibited iu the follow-
ing two tables :

Water conductivit}) in soil with varying size of grain.

"Water sank to a depth of—

Soils.

After 5
minutes.

After 10
minutes.

After 15
minutes.

After 25
minutes.

After 45
minutes.

After 65
minutes.

After 120
minutes.

In soil of grain :

0.01 to 0.71 millimeters . . .
0.071 to 0.114 millimeters.

Cm.
8.8
18.0
28.3
45.0
84.0
11.0

Cm.
32.8
27.0
48.0
82.0

Cm.
16.2
37.0
65.0

110.0

Cm.
21.3
52.5
96.0

Cm.
30.0
79.0

Cm.
36.7
103. 0

Cm.
52.0

0.114 to 0.195 millimeters-

0 175 to 0.2 millimeters ..

Mixtureof various grains .

19.0

24.5

33.2

50.8

C5.5

106.0

Water conductivity in granular soils.

Soila.

Loam powder :

0 to 0.25 millimeters.
0.5 to 1.0 millimeters

Loam grannlfs :

1 to 2 millimeters

2 to 4 millimeters

4 to 6.75 millimeters-
6.75 to 9 millimeters -
jUixturo

Water sank to a depth of-

Afterone
half hour.

Cm.
9.0

18.8

After 1
liour.

Cm.
32.1
32.2

19.0

32.2

19.3

32.0

18.8

30.4

18.5

31.0

4.0

8.0

After 3
hours.

Cm.
20.2

82.4

83.1
81.5
77.5
80.5
11.0

After 4
hours.

Cm.
23.4
100.0

100.0
100.0

99.6
100. 0

19.5

After 23
hours.

Cm.
57.4

After 59
hours.

Cm.

97.6

24.1

100.0

According to Fesca the downward movement proceeds quickest in a
dry dust, only slowly in clay soils; tlie same amount of water being
drained through the former in one hour which it took two days to drain
through the latter.

The influence of a soil cover on the physical condition of soils has
been investigated directly by Wollny ; he comes to the result that veg-
etation and cover with dead material (straw, litter, etc.) tend to i)re-
serve the loose granular structure of the soil.

The forest cover, then, has a tendency to preserve the granular, porous
structure of the soil, which is favorable to tiltration ; and as, moreover,
the roots furnish cliannels for unimpeded drainage, it must have the
tendency, other things being equal, to allow a more rapid tiltration
than the naked, mostly compacted soil, or even that of a field of crops
after cultivation ceases.

The temperature, too, appears to have an influence favorable to rapid
filtration in the forest; for, according to Pfaff", in the field during winter
three-quarters of the precipitation will sink to 2 feet depth in the soil,
and not more than 10 to 30 per cent in summer.

Unless, therefore, the forest cover itself had a tendency to retard

150

FOREST INFLUENCES.

penetration, which we will see is not the case, the influence of the fof^
est upon the intensity of water absorption would be in the direction of
diminishing superficial flow and facilitating- subterranean drainage;

This factor is of tlie utmost importance in the discussion of the
causes of floods. Without a consideration of the water capacity, and
still more of the intensity of water absorption, it will never be possible
to draw conclusions as to probable floods from the amount of precipita-
tion alone.

The influence of various soil conditions and soil covers upon the
amount of water that will filter through has been investigated by
Wolluy and Ebermayer in an extended series of experiments.

Experiments of this kind which will yield results a])plicable to nat-
ural conditions are exceedingly difficult to arrange, and require not
only many precautions, but must be continued for a long time before
generalizations can be attempted. One of Wollny's series of experi-
ments was intended to show the influence upon filtration of a grass
cover on different soils. The results calculated for 1 acre are as follows :

Kinds of soil.

Calcareous sand with humus

Quartz saud *

Loam .soil '

Peat soil

Amount of filtration.

Fallow
field.

Po%inds.
h 593. 216
3, 044, 250

1. 529, 671

2, 018, 124

Grass
covered.

Pounds.

782. 334

661,548

59 105

405. 162

* From May to November.

The grass cover, therefore, reduced considerably (by 50 per cent and
more) the percolation of water. Ebermayer experimented with boxes
43 square feet surface (4 square meters) and 4 feet deep, filled with
garden soil, leaving one bare, covering another with moss, and two
others each planted with 6-year-old plants of beech and of spruce, with
the following results, arranged according to seasons:

Itain.

Filtration water in height—

Tear.

Under
beech.

Under
spruce.

Under
moss.

Naked
soil.

188G.

mm.
156. 98
560. 22
114. 45
126. 30

mm.

12.65

15.89

1.12

9.73

mm.

10.52

12. 09

0.76

5.98

mm.

16,96

31.00

7.17

11.40

mim.
10.93

26.13

September to November

3.27
9.08

Total

957. 95

219. 20
210. CO

39.39

.10. 61

2.50

29.35

5.05
1.49

67.13

14.40
13.00

49. 41

1887.

9.97

3.91

In these experiments it is remarkable how small a percentage of the
rainfall was filtered through, which would lead us to look at the results
with caution, namely:

WATER CONDUCTIVITY OF SOILS.
Of tlie total rainfall was filtered —

151

By soil covered with moss

By soil naked

By soil covered with hecch growtli
By soil covered with spruce growth

1886.

Per cent.
7

5.1
4.1
3

1887.

Per cent.
C. 2
3.5
2.9
1.5

In regard to the amount of filtration which various soil-cover.s allow,
we have the following- very instructive results from the experiments of
Wollny, in which the amounts of rain and corresponding filtration on
62 square inches sui face are given :

May to September,
1886 — total rain
28,529 grams.

Amount,
grams.

Oak leaves :

5 centimeters .
10 centimeters ,
20 centimeters
30 centimeters

Spruce litter;

5 centimeters ,
10 centimeters
20 centimeters
30 centimeters

Pine needles :

30 centimeters

Moss:

5 centimeters

Bare soil:

30 centimeters

17, 591
19, 482
21. 160
21, 061

17, 793
19, 277
19, 523
19, 467

19, 734

14, 993

11,610

Per cent
of rain-
fall.

61.7
68.3
74.1

73.8

62.4

67. 5

68. 3
68.2

69.2

52.5

40.7

April to September,
1887— total rainfaU
18,652 grama.

Amount,
grams.

Per cent
of rain-
fall.

7,894

7,353

12, 954

13, 272

8,653

7,356

14,611

13,912

9,784

7,260

3,636

42.3
39.4
69.4
71.2

46.4

39.4
78.3
74.6

52.4

38.9

19.5

These figures show that a litter will filter considerably larger amounts
of water than a soaked soil of the same depth, and that the moss cover
allows less Avater to filter than the litter. This is accounted for by the
soil needing a larger amount of water to supply thB moisture evapo-
rated than the litter which remains moist.

Notable is the influence which the thickness of the cover exerts upon
the amounts of drain water and also the relation of the amount of pre-
cipitation to the amount of filtration.

It will be noticed that with a thicker cover to 1 foot in depth (30
centimeters) the amount of precipitation hardly changes the amount of
drain water, while the lighter covers have mucli less power to preserve
a small precipitation, for of course the amounts not drained are evapo-
rated.

E. Ebermayer (Sickerwassermengen in verschiedenen Bodenarten
Wollny 1890) from a long series of experiments comes to the conclusion
that besides clay, it is especially humus, which imbibes almost all pre-
cipitation and gives up very little water below.

A layer of garden mold of 1 m. furnished only 3.2, 5.7, and 7.1 per

152 t-OREST INFLUENCES.

cent filtration water from i^recipitation in fall, spring, and winter. '"'If,
therefore, be says, our earth were covered with a humus soil of 1 m. in
depth, subterranean drainage would be so slim that springs would be
scanty and continuously flowing springs absent.

From these exi^eriments it would then be conceivable that the forest
floor could be of such nature as to prevent rapid filtration to the soil
(close uninterrupted moss carpets, or compact humus), when with sud-
den large masses of rain falling less water would become available for
underground drainage than without the forest cover. Such conditions,
however, are exceptional; the possibility of their occurrence, on the
other hand, makes it necessary in every additional case to ascertain not
only the nature and stratification of the soil, but also the nature of
the soil cover or forest floor, before we can determine whether or not
the presence of the forest is conducive to practically greater percolation.

There is another element favorable to the absorption of water by the
soil, and to percolation and subterranean drainage, which, as far as I
know, has not been elsewhere noted. It is the fact that snow will lie
in the forest more evenly and continuously than on the unprotected
surface. This element of conservation not only increases the amount
finally remaining for drainage, but also prevents the soil from freezing,
keeping it open for percolation when the snows melt in spring.

In open fields the snows are not only apt to be dissipated by ca apora-
tion, but the soil is more apt to become incrusted with an impermeable
surface stratum which would turn over the melting snow waters into
surface drainage.

It is these snow waters, preserved to the subterranean draiimge,
which above all account for the continuity and equality of flow in
springs far away from the catchment basin, the waters that fell in the
winter and melted in the spring reai)pearing in summer.

A further element tending to increase the amounts of subterranean
drainage waters lies in the retardation of the surface flow, by which
the time is lengthened during which the soil may take up and filter
through rain and snow waters.

The forest floor offers such imi^ediments to surlace flow not only in
greater degree than any naked soil but than any other vegetation. An
advantage over other kind of vegetation is also found in the deep i)ene-
tration of the roots of trees, which increases the chance for perc(jlation,
while the more compact soil cover of a green sward would be rather
opposing percolation. All that has been said regarding evaporation
and transpiration within and without the forest needs also be kept in
mind in the discussion of the amounts of drain waters for underground
disposal.

The conclusion, then, is, that in general a forest floor, although retain-
ing much of tlie water in its upper strata, allows less water to run off
superficially, and by rendering the soil more permeable larger amounts
of water are turned into subterranean channels.

FORMATION OF SPRINGS. 153

While tbis increase in drainage waters is the general tendency of the
forest cover, geological stratification may be so favorable to drainage
(deeply fissnred vertical or tilted-np strata) that on the score of per-
colation at least the effect may become irrelevant, or it may be so
unfavorable (horizontal unfissnred strata) that the intiuence upon per-
colation may become practically of small value.

While underground, one i)art of the filtered water becomes station-
ary as soil moisture retained in the capillaries of the soil, and finally in
part to be returned to the air by transpiration from the foliage and
evaporation from the surface of the soil. The other part — the surplus
above the water capacity of the soil — continues to filter through the
soil, gathering into definite channels, collecting in beds or basins, and
finally reappearing as springs. It is obvious that the first part can not
be more than the water capacity of the entire soil layer, unless there
be standing ground water which would replenish the loss by transpira-
tion and evaporation, sustained by the soil moisture.

Now, by reasoning from the statements regarding the greater ease of
percolation in soils kept granular under the protection of a forest-floor,
as well as by the experiments regarding water conductivity and water
capacity of soils under varying conditions, and regarding actual meas-
urement of filtration waters from such soils, we are forced to admit
that in general the quantities remaining for underground runs are not
only larger in the first place, but remain so during their subterra-
nean existence, suffering less loss by evaporation under the forest cover.
This effect, especially apparent on shallow soils, will be more sensible
the further away from the catchment basin the water reappears as
a spring, that is, the longer the subterranean run.

FORMATION OF .SPRINGS AND CONDITIONS AFFECTING THEIR FLOW.

Finally we come to a consideration of the conditions which determine
the final reappearance of the underground drainage in the springs.
The place where a spring appears is, of course, i)rcdicated in the first
place by the structure and topography of the soil and rock strata.
The question of the location of a spring is, therefore, a dynandcal one,
on which the soil cover can have but little influence. Yet even here,
an indirect influence may be found in the amount of water to be
drained, and in the looseness of the surface soil, both of which c()udi-
tions would tend to ])roduce more numerous outlets and a wider distri-
bution of reappearing underground waters.

The following elementary explanation of the formation of si)rings
may serve to show how geological conditions influence to a large ex-
tent the manner in which the wateis falling on the watershed are dis-
tributed in undergiound channels, collected and discharged, and that,
in spite of favorable forest conditions, a region may be poor in springs
and that, without any disturbance of the forest cover, a change in the
location or even in the run of springs may occur.

154

FOREST INFLUENCES.

Springs. — A spring is water which has penetrated the soil and reappears collected on
the snrface. Springs are in most cases the heginnings-of brooks and rivers. Accord-
ing to the manner in whicli tlie percolated water reaches the surface, springs may be
classed as standing or running springs.

The standing or ground-water s])riugs are such as collect water in some depression
of the soil and overflow only as long as the water reaches the lower level of the out-
let. Their formation is easily understood from the acccompauying figure (61), in
which (1) represents a hillside of massive rock, continuing under the overlying
strata at a. The latter consists of impermeable 8trata/2, 2) clay, loam, marl) ; above
this a layer of gravel or coarse sand und rock material (3), and above this a stratum
of soil (4), whicli at X is absent, leaving an open bowl where the gravel layer becomes
visible. All the rain water falling on the plateau op and on the slope o a running
down, when arriving at the impenetrable strata near h, will be diverted into the
gravel bed and spread in this, being prevented l>y the underlying impermealjle strata
from ?iinking. When sufficient water is supidied the water level rises until it appears
at X, and if there is an outlet over the rim of the bowl and sufficient slope of the
ground the spring begins to flow, forming, it may be, the beginning of a brook.

Such a standing or ground- water spring ceases to run if precipitation ceases for a
length of time sufficient to reduce the water level lielow the outlet. Similar condi-
tions can occur alongside of rivers when the seepage of the river supplies the
water to a spring Ijelow the river level, and the level of these seepage waters rises
and falls, of course, with the rise and fall of the river level.

Of running springs, there may be distinguished, according to the manner of their
formation, three kiud.s — soil or surface springs, tissure sitrings, and cavern springs.

A surface spring originates when a more or less imiiermeable soil lorms part of or
lies near the upper soil stratum, allowing the water to enter only imperfectly and
to an iuconsideralde depth, and, passing through the looser parts of the soil, to col-
lect and como to the snrface at some point where the top soil is absent. These
shallow- soil springs naturally vary quite 8ensil)ly, according to the physical condi-
tions of the surface, and are dependent directly on tlie ])recipitatii)n ; ilry uj) easily
if it does not rain or if the soil is exposed to in.solation and is de])rived of shade;
they are warm in summer and freeze out in winter. They are usually found in local-
ities where the rock consists of easily disintegrated clay slates and sandstones,
capped with a shallow layer of decomposed rock, or in the neighborhood of loam
hills. An addition of broken rock and stones to the soil facilitates tlie penetration
of the water aiid increases the comparative tlow of tliese springs.

Wlnde districts along the foot of the Alps in Switzerland. Bavaria, Austria, and
the Carpathians in Galicia, etc., have hardly any other kind of springs.

The second class, conveniently called " fissure" si)rings, originate from the waters
which have deeply peneti'ated the soil and rock through the fissures, rents, and splits,
or numberless cleavage strata of the ujtper rock torniations, and ultimately reach a
deeper lying inclined rock formation, which prevents further penetration and causes

SPRINGS.

155

tlie Avater to tnh along its upper plane until the formation somewhere comes to the
surface and with it the collected water of the spring. These conditions are illustrated
in the aeccompanyiug cut (Fig. 62), in which h c d of represents the upper fissured
formations through which the rain and suow waters penetrate to the lower imper-
meable strata below the line bf, necessarily gravitating to point/, where the oppor-
tunity for discharging as a spring exists; a smaller spring might occur at c. Such
conditions exist where lime or dolomite rocks overlie hard sandstones, compact clay

Fig. 62, — Fissure spring.

slates, or clay beds. These springs, as a rule, are much less dependent on the changes
of precipitation and temperature; they are mostly continuous and even in their flow
and their temperature.

The third class of the running springs may properly be called "cavern" springs,
from the fact that while their waters are drained like those of the second class, they
aie first c(dlected in some subterranean basins or caverns, and appear on the surface
as overflow of these basius.

In the accompanying figure (63), a h c is the catchment basin, from which the vari-
ous fissures conduct the water to A, overflowing at X into B, and from there over-
flowing and appearing at the surface at Y.

Fig. 63. — Cavern .spring.

This kind of spring is found frequently in limestone formations, and since the
waters of such often come from great distances from above their discharge at the sur-
face, they are usually of very cold and even temperature; they are ;ipt to run low
"when the soil is frozen and when precipitation is small; and their discharge is more
or less intermittent. The (tbstruction of the old and ope.iing of a new outlet by a
fall of rocks at A' and V, and the widening of a fm-merly insignificant fissure at- or t,
may reduce the flow or 8toi» the original spring entirely, opening a new one in an
entirely difl'ereut part of the locality.

156 FOREST INFLUENCES.

While we have here cousideied condition;; under which spring's are
formed, there are also conditions under which their formation is ex-
cluded ; such might be found in extended plains or low hill lands, with
a compact, impermealde soil, which may give rise to pools and morasses,
but not to springs. Plateaus of fissured limestone dolomites or of com-
Ijact gneisses or granites may also be poor in springs, their waters sink-
ing at once to such depths that no discharge is met in the immediate
neighborhood of the catch-basin, or else shedding the water at once
superficially.

The most direct influence of a forest cover ui^on the discharge would
be noticeable on the surface springs, since in these the catchment area
and the place of discharge lie close together, while the underground
run is not only short, but lies near the surface, and hence experiences
most sensibly the effect of the j^rotection against evaporation which
the forest cover offers. Deforestation here would no doubt i*educe or
cut off discharge eutirel5^

In cavern springs an influence could be exercised only in the indi-
rect manner, by the increase of filtration over the catchment-basin. The
same pertains to fissure springs, whose sources of supply are usually
quite removed from evaporative influences, and only where these come
nearer the surface or when the spring is only small, may the removal
of the shade of forest cover reduce the outflow.

With reference to ground-water springs, which come to light at a
considerable distance from the catchment-basin, the conditions of the
latter as far as the influence, increase, and preservation of water suji-
plies, and of the area over and under which the waters run collect, is
of considerable importance, while the surface condition of the area
within which the spring lies {a e of Fig. 61), if of impermeable strata,
is of less consequence, except that a forest growth may lower the ground-
water level by transpiration, should the water quantities furnished
from the catchment-basin not be continuous and sufficient. If these
strata consist of permeable soil, they would act as a second catchment-
basin, and the effect of the soil cover upon the quantity of drain waters
(precipitation, evaporation, and transpiration) would be directly no-
ticeable. W"e have seen that the tendency of the forest cover — trees,
foliage, litter, moss — is to change a certain amount of surface draiimge
into subterranean drainage, or, in other words, to reduce the surface
waters where they have fallen. Eventually, however, the subterra-
nean waters come to the surface again, and add their stores to the
surface waters that are carried away in open runs, brooks, and rivers.
Finally, then, all the water that falls on the catchment basin, except that
which is returned to the atmosphere by transpiration or evaporation,
becomes surface water; but the manner in which it runs off is the im-
portant pointo

SURFACE DRAINAGE. 157

-INFLUENCE OF FORESTS UPON SURFACE DRAINAGE.

Surface waters, like springs, may be considered from three i)oints of
view, namely their quantity, the course and manner of distribution,
acd finally their behavior when collected in rivers. All that has been
said regarding the conditions of nnderground drainage has, of course,
a bearing upon the quantity of surface drainage. The amount of sur-
face run-oft" is the complement of the amount drained oft' by springs,
audit follows that where surface drainage is the rule the supply to
springs is scanty, and vice versa.

It is evident that the conditions for a superficial run off are to be
found, first, in the amount and nature of precipitation, and next in the
soil and surfiice conditions. A violent rainstorm will naturally furnish
more superficial run-oft' than when, the rain falling slowly, time is given
for the soil to abvsorb it; a snow cover, fallen on frozen ground, is apt
when melting to shed its water over the frozen surface without pene-
trating the soil.

Nature of soil and soil cover and t opography determine, with equal
amounts of water to dispose of, what the nature of the run off will be.
An impermeable soil takes up sufficient water to make it plastic and
then sheds all additional water superficially; a permeable soil continues
to take up water and conducts it into the depth. This difference of
behavior must influence and determine largely the conditions of any
riverbed; for if it run for some distance through impermeable soil
even insignificant rainfalls will rapidly collect and swell the river,
while the jiermeable soil would have taken up and held all or parts of
the precipitation and would only gradually have given it up.

The topography determines the rapidity of run-off and of collection.
The more diversified the country — cut into dells, coves, rills, and fur-
rows, steeper and less steep slopes — in the greater number of runs of
unequal length is the water collected, while the Ifess diversified the con-
tour the more water must be carried off in each run. Yet where the
diversity of configuration is accompanied by steep slopes the run oft'
may be so rapid that the valley river is filled more rapidly than the
river of the open pi ains country with even slopes of moderate inclina-
tion.

Thus in some of the river valleys of West Virginia the watersheds
are scooped out into such an array of coves, gashes, and water courses
and minor watersheds, and so steep and rapid in descent that, in spite
of the forest cover, a raiutall of a few days will induce a rapid rise of
the rivers, while the same amount of rain will hardly wet the ground
in a prairie country like Iowa.

As regards soil and surface conditions it is ()l)vious that the less
permeable the soil or soil cover the less the absorptive capacity of the
same, and the fewer mechanical obstructions are met the more water
runs oft' superficially.

158 FOREST INFLUENCES.

EETAEDATION OF THE WATERFLOW.

It is in the first place this mechanical obstruction which a forest
floor more than most other kinds of vegetation offers, which changes
the distribution in time of surface waters, that constitutes the forest an
influential factor in water flow.

Direct measurements as to the dilference in time which it takes for
water to run ofl" from watersheds of different conditions are difficult or
almost impossible, because it would be necessary that not only the same
amount of water should fall upon the two areas under comparison, but
also that the topography, angles of slopes, and length be the same, and
a ready means of measurement be found. In fact, the absorption and
obstruction to surface flow acting in the same sense, it would be impos-
sible, and for practical purposes also irrelevant, to credit each with its
separate quota of influence. We shall, therefore, have to be content
with general reasoning and more or less inaccurate observations to prove
this retardation of surface drainage. But without any evidence fur-
nished by experiments, we can at once understand that the surface runoff
IS impeded by any kind of mechanical obstruction, such as is offered by
the vegetation of a meadow or of a forest.

The great number of inequalities which the forest floor otters, in ad-
dition to the trunks and stumps and fallen trees, forces the run-oft' to
many detours, thus retarding its flow and its collection in the open runs

and brooks.

The retardation iii the waterflow begins even before the rain has
reached the soil, for the leaf canopy catches and reevaporates, as we
have seen from 12 to 25 per cent of the total fall, and certainly retards
the fall of the water to the ground, as can be readily observed; long
after the rain has ceased the water keeps on dripping from the foliage.
Thus, although most of the water reaches the ground at last, except in
case of very light showers, yet the devious ways in which it reaches
the soil makes the flow of water from a forest-covered hill longer in time
than if the rain had fallen on a bare slope. As the result of a long-
continued precipitation, it would be under the same conditions by an
unforested slope, but this stage occurs in the forest later than on un-
forested soil and later still than on naked soil.

The great importance of the factor of time in surface drainage, both
as regards dangers from freshets and erosion of soil, will be more readily
appreciated when we remember that the dangerous waters in the moun-
tains are generally of short duration.

A difterence of 1,000 to 2,000 cubic feet of water per second from a
square mile of watershed may often determine whether a dangerous
flood is experienced or not. And since a square mile of moss-covered
forest floor is capable of absorbing from 40,000,000 to 50,000,000 cubic feet
in, say, ten minutes (a humus cover is capable of taking up 50 per cent
of its own weight), nearly all of which the naked soil would give up
Bome tNvelve to fifteen hours earlier, the surface conditions of the water-

THE FRENCH ALPS BAD LANDS OF MISSISSIPPI. 159

slied must in many cases be determinative in the excesses of water flow
in rivers.

Tliis important fact should at least be recognized, that the surface
conditions of the soil of a watershed are the only controllable factors in
the problem.

Amount of precipitation, topography, and character of the soil are
the practically unchangeable other conditions which determine the
occurrence of freshets and floods. With a forest floor in good condi-
tion, small precipitations are apt to be absorbed readily and entirely
prevented from running off superflcially; with excessive rainfalls, topo-
graphical and soil conditions have eventually more influence than the
forest floor; from steep declivities and an impermeable soil waters will
be shed sui)erflcially in spite of and over the forest floor as soon as the
latter is saturated at the surface.

Yet even so a diflerence in. the run-ort" will be experienced by the
fact that the well protected forest soil prevents erosion, the formation
of detritus and the carrying of debris into the runs and brooks below.

EXPERIENCE IN THE FKENCII ALPS.

•

By this protection of the soil the so-called torrential action of water
is prevented, which, as the history of some departments in southern
France has shown, is capable of devastating thousands of acres offer-
tile land by carrying the detritus into the valleys and depositing it
there. At the same time the reforestation Avorkof the French Govern-
ment has also progressed far enough to furnish proof that the recloth-
ing of the denuded hills is the practical remedy against these torrents.

Not only were the mountain sides themselves devastated and made
useless by the destructive action of the water, but fertile farms for 200
miles from the source of the evil were ruined by the deposit of the
debris and the population pauperized and driven out.

According to M. Demontzey, forest administrator of France, it was
estimated, in ISGO, that the area of denuded mountain lands needing
reforestation was 2,904,000 acres. The Government has taken hold of
the restoration of the most needing area, some7<S(),000 acres, on which
so far .some $10,000,000 have been expended, while private owners and
communities have increased this expenditure for the repair of past fol-
lies to over $30,000,001), which is estimated to be about one-half of what
is necessary.

flow in our own country this erosive and destructive action of water
is at work even in the hill country is thus graphically described by Mr.
McGee, in speaking of the bad lands in the State of Mississippi :

With the moral revolution of the early sixties came an iiulustrial ovoliitioii ; the
planter w;ih imjiovcrished, liis sonLS were slain, his slaves were liberatecl, and he was
fain either to va<ate the plantation oi- jircatly to restrict his operations. So the cul-
tivated acres were altandoncd by tiiuusands. Then tlie hills, no longer j)rotected by
the forest foliage, no longer bound by the forest roots, no longer guarded by the

IGO FOREST INFLUENCES.

balk and brush dam of the careful overseer, were attacked by raindrops and rain-
born rivulets and gullied and channeled in all directions; each streamlet reached a
hundred arms into the hills, each arm grasped with a hundred lingers a hundred
shreds of soil, and as each shred was torn away the slope was steepened and the
theft of the next storm made easier.

So, storm by storm and year by year, the old fields were invaded by gullies, gorges,
ravines, an<l galches, ever increasing in width and depth until whole hillsides were
carved away, until the soil of a thousand years' growtli melted into the streams,
until the fair acres of ante-bellum days were converted by hundreds into bad
lauds, desolate aud dreary as those of the Dakotas. Over nnich of the iipland the
traveler is never out of sight of glaring sand wastes where once were fruitful fields;
his way lies sometimes iri, sometimes between gullies and gorges— the ''gulfs" of the
blacks whose superstition they arouse, sometimes shadowed by foliage, but oftener
exposed to the glare of the sun reflected from barren sands. H<3re the road winds
through a gorge so steep that the sunlight scarcely enters, there it traverses a narrow
crest of earth between chasms scores of feet deep in which he might be plimged by a
sino-le misstep. When the shower comes ho may see the roadway rendered impass-
able, even obliterated, withina few minutes; always sees the falling waters accumu-
late as viscid mud torrents of brown or red, while the myriad miniature pinnacles
and defiles before him are transformed by the beating raindrops and rushing rills
so completely that when the sun shines again he may not recognize the nearer
landscape.

This destruction is not^ confined to a single field nor to a single region, but
extends over much of the upland. While the actual acreage of soil thus destroyed
has not been measured, the traveler through the region on horseback daily sees
thousands or tens of thousands of formerly fertile acres now barren sands; and it
is probably within the truth to estimate that 10 per cent of upland Mississippi has
been so far converted into bad lands as to be practically mined for agriculture
Tinder existing commercial conditions, and that the annual loss in real estate
exceeds the revenues from all sources. And all this havoc has-been wrought within
a quarter century. The processes, too, are cumulative; each year's rate of
destruction is higher than the last.

The transformation of the fertile hills into sand wastes is not the sole injury.
The sandy soil is carried into the valleys to bury the fields, invade the roadways,
and convert the formerly rich bottom lands into treacherous ciuicksands when wet,
blistering deserts when dry ; hundreds of thousands of acres have thus been de-
stroyed since the gullying of the hills began a quarter century ago. Moreover, in
much of the upland the loss is not alone that of the soil. i. e., the humus represent-
ing the constructive product of water- work and plant- work for thousands of years;
the mantle of brown loam, most excellent of soil stuffs, is cut through and carried
awav by corrosion and sapping, leaving in its stead the inferior soil stuff of the La-
fayette formation. In such cases the destruction is irremediable 1»y human craft —
the fine loam once removed can never be restored. The area from which this loam
is already gone is appalling, and the rate of loss is increasing in a geometric propor-
tion.

The formation of detritiis and deposit in the river beds may also
become the cause of dangerous floods in the hirger streams for the
amount of rock material and soil which the rivers carry is one of the
most potent factors in their water flow.

Since this detritus is deposited wherever the velocity of the water
sinks below that necessary to carry it, forming sand banks and rubbish
heaps which obstruct and change the direction of the run, it plays
quite an important part in shaping the bed of the river, besides in-

RIVER DEPOSITS AND FLOODS. 161

fluenciiig the whole system of dependent brooks and rivers. The na-
ture and shape of the detritus — whether fine sand or earth, smaller or
larger rock masses, stones, roundish, square, or flat — cause much difter-
ence; and this in turn depends upon many conditions, geological and
climatical.

According to the nature of the rock from which it is deriA'ed, the
detritus appears in different shapes, and again changes in form dur-
ing its further transportation by the waters in diU'erent ways, and
therefore exerts n varying influence npon the run. Thus the detritus,
which appears in large plates or shales, is carried more easily than the
square or round rocks; the former, even when deposited, hinders the
flow of water between the plates but little, and therefore gives less
cause for stow water than the heavy square rocks, which resist the
transportation and obstruct the flow more efl'ectuall^'.

Sand and gravel detritus is easily carried, easily accumulated, and
again with a new flood easily removed; it otters, therefore, little resist-
ance to the flow of water, but becomes objectionable in filling the lower
channels of rivers, etc.

Clay detritus, although easily carried, is apt to compact and cement
the rock detritus together, and thus becomes one of the worst impedi-
ments of water flow and is the cause of the worst dangers from flood
waters.

Fr(mi tlu'so exani])les it is apparent that two rivers, although under
similar conditions of rain fall, i)hysical conditions of soil and topog-
rai)hy, may yet show different behavior, according to varying char-
acter of the detritus.

As we have stated in the beginning of this chapter, the underground
drainage gives to rivers their permanent regime, while the surface
riiii off gives to them their rapid variations of stage,

According to whether a river is mainly supplied by surface waters
that riin iu trackless courses and open runs over the slopes into its
bed, or wliether it is supplied by underground waters, its water condi-
tions must follow a different course, fitful in the first, even in the second
case. Few rivers depend on one of these sources of sui>ply alone;
Oiost of them are supplied in the different parts of their course by both,
so that a streanj nuiy begin as a torrent and later in its coarse find
additional supplies from springs or ground wat<n's — or the reverse may
take place, being originated by a spring, it may have no other addi-
tions except from surface waters,

The greater the proportion of the supply by surface run -off the more
lialde to disturbances must be the flow. Finally in a larger river, like
the Ohio or Mississippi the question of floods becomes still farther
com]>licated. For here not cmly the regime of the main river, but also
that of all its affluents and the topographic, stratigraphic, climatic, and
surface conditions of their catchment basins become elements of dis-

12411— No. 7 11

102 FOREST INFLUENCES.

turbance. Here the comparative lengths of the affluents alone m ay-
become all important, since the simultaneous or uonsimultaneous
arrival of flood waters may determine the occurrence or nonoccurrence
of high floods. As far as forest cover is concerned in such cases,
deforestation in one of the side valleys and consequent rapid discharge
may become an advantage for the water flow in the main river, by allow-
ing its removal before the arrival of the flood waters of another affluent.
In view of these considerations it would, therefore, be folly to assign to
the condition of forest cover in the catchment basiws an all determinative
function. jSTevertheless, in general the influence of favoral)lc forest
conditions in the catchment basin upon river flow can not be doubted,
although it may become practically of no account in abnormal floods.

The first cause of abnormal floods is the occurrence of abnormal rain-
I'alls or the sudden thawing of abnormal masses of snow. If the former
occur after the soil has been saturated, or the latter when the soil
remains frozen, the forest cover will be powerless to influence the run-
off and will shed the water as rapidly almost as the open ground,
although even the brief retardation of the confluence of water masses
which the obstacles of a forest growth cause may be of moment.

But in its further course the drainage of this Avater, collected in the
rivers is favorably influenced by the presence of the forest, it having
prevented the formation and deposition of detritus in the river bed.

The beneficial influence of the forest in case of abnormal floods can
then i^robably be claimed only in so far as it protects the slo])es
against abrasion and the formation of debris or detritus with which
the upper head waters are filled, and which carried down into the
rivers gives rise to sand banks and changes in the river bed which may
increase the dangers of the next flood.

EXAMPLES OF THE IXFLCENCE OF FORESTS ON WATEKFLOW.

Besides the detail experiments, which are to furnish explanation of
the physical laws upon which differences in water drainages occur,
there are constantly accumulated experiences. Especially in France this
question of forest influence on waterflow has been answered by practi-
cal demonstrations. To bring it, however, to a final solution, special
hydrographic surveys and statistics, such as are now contemplated in
Germany, will be necessary. A first attempt at such a work is the hydro-
graphic description of the lihine, a model work compiled by the agency
of the various governments whose lands border on that river, published
in 1889. This, however, was too comprehensive a field, and more de-
tail measurements and observations would promise more striking re-
sults. The conclusions from this survey, and a more detailed one of the
Hauensteiner Alp in the southern Black Forest mountains, are that the
comparative absence of damage from high floods in this Alp country,
when compared with neighboring valleys, may be ascribed to the forest

INFLUENCE ON WATER FLOW. 163

cover, occupying 51 i:>er cent of its area, and the general observation is
confirmed that the presence of well kept large forest areas at the head
waters in the mountains has a favorable influence upon the water
stages in the water channels and per cent of erosion and formation of
detritus to an appreciable extent.

One interesting detail hydrographic survey, which has in view to
establish relation of forest cover to river, is published by Danckelman.
It refers to the river Wupper, an affluent of the lower Rhine from the
Khenish state mountains. It was made in connection with the con-
struction of a dam to regulate water supplies. Three catchment basins
were under consideration, two of these, although containing 32 and
39 per cent woodland, are reported as devastated and especially robbed
of their soil cover of litter and humus, the third containing nearoU per
cent of well-kept dense forest.

From the detailed observations and measurements there was not ob-
servable any influence of forest conditions on the average daily and min-
imum flow, but in case of high water the daily flow was most decidedly
influenced, namely, a diminution or retardation occurred in the well-
wooded basin: in July, by 55 per cent; in August, by 34 per cent; in
November, by 28 per cent; in March, by 21 per cent. It is stated in
addition that the well-forested basin had a larger amount of rainfall
and steeper slopes, a narrower valley, and was, therefore, comparatively
more unfavorably situated.

An interesting note as to the amount of retardation which may be
produced by the artificial means employed in the French Alps for reg-
ulation of waterflow, namely, forest planting in connection with overflow
dams, is given in M. Mathieu's work Reboisement in France.

The two basins of Faucon and Bourget were visited by a terrible
downpour of rain of twenty-five minutes' duration. In the upper moun-
tains there fell 42 millimeters, in the lower regions 12.3. The torrent
of Faucon (which was in a devastated, deforested condition, but other-
wise topographically similar to that of Bourget) was at once filled with
flood waters which were estimated to consist of 60,000 cubic meters of
water and 180,000 cubic meters of rock material or detritus, the flood
subsiding in two hours.

In the torrent of Bourget, which had been reforested and corrected
in its bed, a simple, somewhat turbulent run of water was observed,
which at the overflow reached the licight of 45 centimeters (18 inches)
and lasted about three hours. The report continues:

These facts show the iiuportauce of the forest cover. Thanks to the dense forest
growth planted, the flood waters, divided in nnraberless runs and retarded constantly
in their movement over the declivities in the upper basin, arrive only successively
and little by little in the main bed, instc^ad of those formidable masses of water and
debris which, rapidly agglomerated, rush into the channel ; the brooks called to re-
place the torrents receive only pure water; flood waters flowing off gradually and
made harniless by the regulation of the torrent bed and of the slopes.

164 FOREST INFLUENCES.

In the department of L'Heranlt, in the Oevennes Mountains in
southern France, the following calculation was made of the amount of
water retained by the forest cover after a heavy storm.

The basiu of Lampy, comprising l,60f> acres, of whicli more than 50 per cent are
under forest, the rest in grass and field (6 per cent), rests on impermeable granite
and quartz roclc with a layer 2 feet deep of rich humus soil. All the water falling
must i)ass into a reservoir formed by closing the valley with a dam which stores the
water to a heiglit of 51 feet. The reservoir, nearly 100 years old, has never needed
to be dug out, which is mentioned as a sign of the absence of soil erosion. When
full all the water must pass over the dam in an overflow race, which permits a toler-
ably exact calculation of the discharge of water from this reservoir. The discharge
through pipes and overflow in one ye.ar (1860) amounted to 4,060,038 cubic meters,
while from the record of one rain gauge at the reservoir situated, therefore, at the
base of the elevations which rise to 1,000 feet above it, the rainfall was calculated
6,837,350 cubic meters. prol)ably on account of the position of the gauge an under-
statement, showing, therefore, that not less than one-third and probably as much as
cue-half of the rainfall had been retained in the soil or evaporated.

On two days (July 28 aud 29, 1863) there fell on this area of 1,600 acres, according
to the rain gauge. 530,500 cub. m. of rain. Before the rain the reservoir was full to
high-water mark, and the overflow, the only means of discharge open, had been
delivering for days 3,936 m. or 45 liters per second, which must, therefore, be con-
sidered the natural discharge of the basin.

After the rain the water level rose 8 inches and the discharge was:

Cubic meters.
July 29, moi-ning and evening 30, 504

30, morning and evening 28, 864

31, morning and evening 7, 872

Aug. 1, morning 7, 872

2, evening 4, 920

Total in live days 80,032

At the end of the fifth day the water level had returned to its former height and the
discharge the next day was again 3,936 c. m., Avhich it maintained for three mouths.
If we deduct from the discharge of 80,032 c. m. the water that would have been dis-
charged during these five days Avithout the additional rainfall, namely (5 X 3,936=)
19,680, there remain 60,000 c. m. round, which without doubt were furnished by the
two storms, and since the total fall had been at least 500,000 c. m. it follows that
more than eight-ninths of the rain was absorbed and held by the soil to be delivered
gradually.

In a neighboring basin, that of Salagou, of 10,701 acres extent, with
just 10 per cent wooded, also on impermeable rock (permian), bnt the
soil otherwise considerably washed and thin, observations could not
be carried on with precision. But while the discharge of this much
larger basiu in ordinary times Is calculated at not more than 20 to 25
liters per second, after a storm the discharge into the river is almost
immediate and has been observed to rise to more than 000 c. m. per
second.

Of examples in this country we may give the following as coming
from good anthorities and well substantiated.

Jn a report to the Chief of Engineers of the War Department in

RELATION OF FORESTS TO FLOODS. 165

regard to a survey of the Savannah River, made by order of Con-
gress, it is said:

Reports upon the Sixvaiiiiah Valley and River at Augusta of about the year 1775
show the Savannah to have beeu a clear, rapid stream, full of excelleut fish and sub-
ject to no sudden or marked changes of height. This was previous to the destruc-
tion of the forests and the opening of large tracts of land to the plow. Now the
stream has become turbid ; the fish, I am told, have nearly deserted the lower waters,
and sudden and marked changes of the water level are the rule.

Maj. Charles W. Raymond, in a report to the Cliief of Engineers, on the
west branch of the Susqnehanna River and the practicability of improv-
ing its navigation and of confining its waters, in times of great flood,
to the general course of its channel, attributes as the principal cause of
the excessive high and low stages of this river the progressive destruc-
tion of the forests from the mountain crests and slopes of its watershed,
and in discussing the means of protecting the river valley from inunda-
tion in the future, says: "Most important of all, such forests as yet
remain upon areas not valuable for cultivation, especially near the
head waters and the upper slopes of the basin, should be protected."

The New York Forest Commission, speaking of floods in the Adiron-
dack region and the influences of forests in relation to them, say:

In the uplands of the preserve there are many densely wooded tracts adjacent to
others from which the forests have been stripped. The residents agree that in the
former Hoods are unknown, while in the latter they are a yearly occurrence. Their
appearance was coincident with the disappearance of the woods. It was then
noticed that the bridges, which for many years had sufficed to span the streams
during heavy rains, were no longer safe, and new ones with longer spans became a
necessity.

They- refer also to the effect of the removal of the forests in the
Adirondack watersheds upon the navigation of the canals of the State
and the whole system of inland commerce. They say:

With the clearing away of the forests and the burning of the forest floor came a
failure of canal supply that necessitated the building of costly dams and reservoirs to
replace the natural ones which the fire and ax had destroyed. The Mohawk River,
which for years had fed the Erie Canal at Rome, failed to yield any longer a suffi-
cient supply, whereupon the Black River was tapped at Forestport, and its whole
volume at that point diverted southward to assist the Mohawk in its work.

The superintendent of public works of the State has also called
I>ub]ic attention to this subject several times. In the report for 1882
he says:

The importance of the preservation of the woods in the Adirondack region in con-
nection with the water supply of the canals can not be overestimated. With the
continual cutting away of the forests and the burning of the forest floor, the decreas-
ing water sujiply lias become painfully apparent. Should this continue, the result
on the canals would be disastrous.

Another interesting and impressive exampleof changes in conditions
of water flow is given in connection with the Schuylkill River. During
the last sixty or sixty-five years this river has shown a marked diminu-

166 rORfeST INFLUENCES.

tion iu its minimum flow. In 1816 this flow was estimated at 500,000,-
000 gallons per day; in 1S25, at 440,000,000: in 1867, at 400,000,000, and
in 1874, at 245,000,000. In regard to tbis a commission of engineers
say in their report in 1875:

This remarkable decrease, uot being accompanied by any great change iu the
rainfall, nor probably in the total annnal discharge of the river, is no doubt largely
due to the destruction of the forests in the drainage area, whereby the conservative
action of the woodland has been lost, and the rainfall is permitted to descend rap-
idly to the bed and pass off in a succession of freshets.

hi the year 1881 the State of New Hampshire established a forest
commission, wlio were instructed to inquire, among other matters re-
lating to the forests, into "the effect, if any, produced by the destruc-
tion of our forests upon onr rainfall, and consequently upon our ponds
and streams." In theii- report, made in 1885, the commission presented
a summary of the large number of rej^lies to their inquiries. These
replies came from all parts of the State. From the summary the fol-
lowing citations are made:

Beginning witii the southern portion of the State, and with the town of Rich-
mond, attention is called to a small stream there, which in 1865 furnished sufficient
power for four sawniill« nearly all the year, but which began to dry up with the
more rapid reinuvil •)<' the timber occasioned by the introduction of steam as au
auxiliary power, 'ilie water and the woods have disappeare<l together, and the
same is the case in other ))(>rtions of the town.

In Fitzwilliani and Rindge the same results have been reached all the more rap-
idly because of the nearer proximity of these towns to a market. Well-known
trout streams, once abundantly stocked with fish, are now dry half of the year,
and the treeless ground and naked rocks along their banks and about their sources
are considered a sufficient explanation.

The chairman of the board of selectmen in Henuiker, who has given much atten-
tion to the subject, is confident that the water iu the Coutocook River has decreased
fully one-third within even twenty years, and that the triltntaries have fallen otf
still more, many being nearly dry in the summer. During this period $75,000 worth
of timber has been cut within this one town. In the surrounding towns, also, the
timber has disappeared Avith equal rapidity, and the water supply has seriously
decreased.

The report from Bow, which covers a period of fitty years, within which most of
the timber has been cut otf, and that from Hoiikinton, which covers a period of
sixty years, both tell the same story of naked hillsides and diminished streams.

At Hanover the Connecticut River for many years has been decreasing in volume,
aud with increasing rapidity the timber from its head waters has beeu tloatiug by.

In Canaan sixty-five years ago there were nine or more mills of ditterent kinds;
abundant water power all the year around; no thought of reservoirs or double
dams, or precautious against drought. Canaan street, now covered with a firm, dry
sod, was laid out through a swamp, impassable but for the humuiocks and fallen
trees, while dense forests of giant trees covered the hills. The writer who fur-
nishes the above facts, a native of the place, returning after an absence of thirty
years, found the hills and rocks bare, the springs choked up, and the mills obliged
to resort to steam power or lie idle.

The great mountain region of the State lies in contiguons parts of the conuties
of Grafton, Carroll, and Coos. The nuuibeiless streams originating in this region,
protected by the primitive forest, might be thought to be beyond any disturbing
causes, but such is not the case. The town of Littleton depends upon the Ammo-

EFFECT ON STREAMS IN NEW HAMPSHIRE. 167

noosnc foi' its water power, but three of its oldest citizens testify that this power
has diminished one third within lifty or sixty years. The mountain forests during
this same period have been encroached upon as never before, and it is not surpris-
ing that so ooiiimonly these two facts are associated as cause and consequence.

Coos County ct)ntains more of the first growth of timber than any other portion
of tlie State. In tlie midst of this region are the sources of the Connecticut, An-
droscoggin, Saco, and tlieir many tributaries, and a diminished water supply at
this point is felt throughout the course of these important streams. The report
from .Jefferson is that the older inhabitants agree that the streams are smaller than
formerly. An intelligent observer at Berlin, on the Androscoggin River, maizes the
following important statements, covering a period of twenty-six years. Within a
radius of 4 miles from his residence are eight streams or brooks and two ponds, and
the water in each during the above period has materially diminished. As an illus-
tration of the connection between the removal of the woods and this diminished
supply, he adds that '* six years ago he supplied his stock with water from what was
then an unfailing brook, by means of an aqueduct which furnished 300 gallons per
hour. Now that the trees along the stream have been destroyed by the woodman's
ax and by forest iires, his water supply is cut short in summer by drought and in
winter by frost. Hundreds of acres of timber have been cleared within these six
years in the s.ame vicinity."

At Lancaster, the county seat, on the Connecticut river, an old resi-
dent reports —

an alarming decrease in the water of the streams and springs during the past sixty
years, and especially during the last twenty-five years, within which period the
smaller timber also has been removed. Israel's River in his boyhood was a large
mill stream 8 or 10 rods wide, with sufficient water to carry a very large amount of
machinery the year round. Now it is an insignificant stream, with, from May to
November, not more than half the water it had fifty years ago, and not more than
two-thirds there was twenty-five years ago. Other streams have suffered in the
same way, and the springs have, if possible, suffered more than the streams. Many,
once thought to be never-failing, are now for long periods dry. That the cutting
oft' the forests accounts very largely for this change he considers as sure as that
ertect follows cause, and the result is hastened by the reckless methods in use. In-
stead of cutting timber that is matured, everything is cut to the size of 5 or 6 inches
in diameter, and what remains is cut into firewood or burned at once, leaving a
dreary waste.

In conclusion the commissioners say:

While the statements given above prove beyond doubt the steady diminution of
our water su))ply, and show what is the commonly received explanation of this
state of things, a few of the towns heard from, and these mainly in the southern
l>art of the State, report no very marked variation in the amount of water in ponds
and streams for a considerable term of years, and an increase rather than a diminu-
tion in the amount of woodland. Much of this woodland, however, is the young
growth, brush wood only, which can not for years protect the ground from the dry-
ing effects of sun and wind, as did tlie older woods; and, besides, from its relatively
greater amount of foliage, evaporation proceeds all the more rapidly. Thai our
wooded districts here and there are on the increase can not, however, alter the force
of the facts which confirm the nu)re commonly received oi)inion as to the general
■condition of things within our limits. If in any instances the decrease in water
power has been checked or averted, it is all the more important to know how the
result has been reached, that the same means may be used elsewhere. In every case
this means has been connected with the jireservatitm or restoration of the forests.

On one point there is no division of opinion. It is not iu the open ground, but be-

168 FOREST INFLUENCES.

iieath the trees, that the moisture and the snow accnuinlate, and are slowly and surely
supplied to the springs and streams, which then have a perennial flow. Let the
same grouud lie dei^rived of its shade Aw\ this exposure to the sun hastens evapora-
tion, and the rain and melting snow rapidly pass oft' through the water courses he
fore any sufficient quantity can reach the permanent reservoirs under the surtace.
The snow on the exposed hillside may be swejtt oft' entirely by the wind; and even
when any considerable portion remains, much will evaporate, and after all he lost to
the soil and the sj)rings. The soil itself is often vi'ashed oft', and the exposed rocks
given over to perpetual barrenness.

In coufirmation of tbe conclusions of tlie New Hampshire forestry
commission may be appropriately cited the statement made in the New
Hampshire Geological Eej^ort (vol. i, p. 124), that when in the central
and southern j)ortions of the State the hay crop has been cut short by
drought, it has been known to be above the average in the northern
part, even with less rainfall, and as a reason it is claimed that the
forests in the northern section have secured a better distribution of the
results of rainfall and melted snow.

R. W. Piper, in his Trees of America, gives this illustration as com-
ing under his own observation:

Within about one-half mile of my residence there is a pond upon which mills have
been standing for a long time, dating back, I believe, to the first settlement of the
town. These have been kept in constant operation until about twenty or thirty
years ago, when the supply of Avater began to fail. The pond owed its existence to a
stream which has its source in the hills which stretch some miles to the south.
AVithiu the time mentioned these hills, which were clothed with a dense forest, have
been almost entirely stripped of trees; and to the wonder and loss of the mill-
owners, the water in the pond has failed, except in the season of freshets, and, what
was never heard of before,»the sti'eam itself has been entirely dry. Within the last
ten years a new growth of wood has sprung up on most of the land formerly occu-
pied by the old forest, and now the water runs through the year, notwithstanding
the drought of the last few years.

A gentleman in eastern Massachusetts makes the following state-
ment: Having made a contract to supply an extensive nail factory with
kegs in which to pack the nails made, he purchased a timber tract in
southern Vermont, throngh which ran a stream. Upon this stream he
erected a sawmill and began to cut the timber and make it into kegs.
It was not long before the amount of water in the stream was les-
sened to such a degree that he was obliged to erect another mill below
the first, and thus use the water a second time in order to maintain the
requisite X)()wer for carrying on his business.

It is a well-known fact that the flow of water in the Hoosick and
Housatonic ri^'ers, in western Massachusetts, has become so irregular
that the mill owners on those streams have been obliged to make stor-
age basins in which to hold the water of the spring floods for use in the
summer, or else to equip their mills with auxiliary steam engines. The
result is claimed to be due to extensive deforesting.

Mr. David Thomi)son, of Cincinnati, said to the American Associa-
tion for the Advancement of Science, in 1881;

LESSENED STUEAMS WITH LESSENED FORESTS. 169

It is nut uuusua] to liml iu many localities tlie beds of what were once importaut
mill streams, waterless, except when filled by sntlden freshets, and in Ohio certain
streams emptying into the lake, which were once declared navigable, will not now
float a canoe. Previous to 1S32 Capt. Delorac, of Hamilton, Ohio, annually sent a
fleet of flatboats down the Big Miami River at the spring rise; but with the de-
struction of the forests aloug that rii. er the rise became so uncertain that the enter-
prise was of necessity abandoned.

A tUrmer in LTlster Coauty, N^. Y., gives tlie following testimony ou
the snbject before us. He had cut an acre or two of trees on an ele-
vated portion of his farm. In giving the result he says:

My flrst loss was the drying up of a beautiful brook which had its source in my
grove, and which ran through a number of fields, furnishing water for cattle while
grazing. Five times tlie value of the wood I sold would have been refused for this
stream. In the vicinity of the place where the timber stood the ground became dry
during the summer. When rain fell it did not seem to be absorbed, the water ran
down the hillsides, making great gullies and doing much damage, while the fields
through which the brook flowed did not yield as good crops. I am now a strong-
believer in the value of woodlands on a farm.

A gentleman iu Oiujudaga, C'ouuty states that the streams iu that
county have visibly failed since li is boyhood, though he is not yet 40
years of age. There was at Coukling's Falls, he says, a grist and saw
mill which in his youth had a plentiful supply of water. Then it grad-
ually diminished. At first a spasmodic flow was marked; heavy fresh-
ets in spring, then low water iu summer, until the water failed and it
was necessary to run the mills by steam. So at Pratt's Falls, a few
years ago the flow of water was abundant. The story was repeated
there, violent freshets iu spring, followed by the usual failure, until
now, in summer, hardly a pailful runs over the falls. In this latter
case there was formerly a swamx), some 5 or 6 miles above the falls,
which has been reclaimed and all the trees and shrubs cut off. All
these chiiuges have occurred within fifteen years.

Ex-Goveruor Davis, of Maine, gives the following statement in re-
gard to the effect of forest removal on the flow of streams, in a case
with which he is well acquainted:

The Keiidii.skeag River empties into the I'enobscot at Bangor. The stream rises some
.30 miles from its mouth, one branch iu the town of Dexter, and another in the town
of Corinna. I am told that fifty or sixty years ago there was a continuous flow of
water the yt-ar round in this stieani, and at the town of Keudnskeag, 12 miles north-
east of Bangor, were sitnat»Ml large Inmbcr mills on botii sides of the stream. The
water-flow was sufficient to carry them the year round. But during thepast half
century the land along the shores of the stream has been cleared throughout the
greater part of its cour.se. The result is that we have heavy s]»riug freshets, also
heavy feshets in the fall, sometimes doing much damage. I ret^ollect, a dozen years
ago or more, when living in the town of Corinth, through which said stream flows,
almo.st every bridge ou the stream was carrit-d away in the month of Maicli. Now,
after the spring freshet subsides, the water falls rapidly unlil it dwimlb-s to a very
small stream, not one-half the amount flowing during the summer mouths that did
fifty years ago.

170 FOREST INFLUENCES.

Mr. Abbot Kinney, of California, au iutelligent observer, gives the
following recent testimony in regard to a particnlar Held on liis estate:

This field was cleared of a dense brush jjrowth, about 15 feet bigh, wbicb, in Cali-
fornia, is called chaparral. When first cleared the soil was quite dark in color and
full of vegetable detritus.

For two years no special care was required to prevent gnllies forming from the
rains. The rain-holding power of the field has constantly diminished, cross furrows
have now to be carefully prepared and maintained during the rainy season. A sharp
raiu now runs off without doing much good where it formerly soaked in. The dark
color has gone and the soil is now a plain red. It packs hard now, after every rain,
when formerly it did not pack at all, except in a pathway. On the edge of the
bluft's where the brush was left the old conditions prevail. I was doing some work
along these bluif edges and found that I could shovel the dirt easily after cutting off
the brush, but on the cleared land adjoining whore the plow had missed, near the
edge, the ground was so hard as to require a pick.

In 1885 the Oliio State Forestry Bureau issued circulars to its cor-
respoudeuts tlirougliout tlie State making inquiries in regard to tlie
observed influence of clearing the forests upon water supply. The re-
plies received are published in the first annual report of the bureau
and will be found quite in harmony with the testimony above given.

Such reports as these can be multiplied from every section of our
country, and, while there would always remain the onus of proof that
the change of Avaterflow and forest conditions were in causal relation,
it is difficult to conceive of any other causes for these experiences over
so widespread an area, than the change of surface conditions due to
deforestation, especially the burning of the forest floor. However
questionable the position of forest cover as a climatic factor, its rela-
tion to waterflow and soil conditions is attested by experiences la all
])tirts of the world.

I\r-NOTES ON THE SANITARY SIGNIFICANCE OF FORESTS.*

By B. E. Ferxow.

The subject of the sauitary significance of forests has been recently'
reviewed by Dr. E. Ebermayer, the well-known physicist at Muuich.
Otlier investigators have also contributed new material toward the
discussion of the subject. Especially the two Italians, Seratini and
Arata, investigated the intlueuce of forests on the quantity of micro-
organisms in the air, and Dr. Puchner investigated the contents of car-
bonic acid in the atmosphere under varying conditions.

Eberniayer shows that the oxygen exhalation of a forest in propor-
tion to the consumption by man is insignificant. He figures that a
family of four persons would require for respiration, in the burning of
the necessary fuel, the oxygen exhaled by 2^ acres of forest. The hy-
geuic significance of ozone he doubts. Puchner shows that the air in
the forest contains generally more carbonic acid than the open, due to
decomposition of litter. On the other hand, Eberniayer shows the air
in the forest soil to contain less carbonic acid than that in the field soil,,
three fourths times less in winter and five-sixths times in summer; this.
is explained by differences in moisture conditions. But, like sea and
mountain air, forest air is freer from injurious gases, dust particles, and
bacteria. Furtherinore, the shade and the processes of assimilation
and transpiration have a cooling effect in summer, a warming effect in
winter, hence extremes in temperature are checked. Protection against
winds and extremes of temperature which the forest offers is cited as
desirable for the location of sanitaria and finally a tribute is paid to
psychic inriuence, and the liygienic significance of the forest is pro-
nounced as scientifically established. But of much more importance
than the air is shown the forest soil, especially since cholera, typhus,

"These notes are based on the following publications:

Dr. E. Eberniayer, in WoUny, 1890: (1) Hygienic significance of forest air anrl
forest soil. (2) Experiments regarding the signiticanco of liiiinus as a soil constit-
uent and influence of forest, different soils and soil covers on composition of the air
in the soil.

Dr. H. Puchner: (1) Investigations of the carbonic acid contents of the atmos-
phere.

Serafini and Arata : (1) Intorno all 'aziono dei boschi sui mikro organismi trans-
portati dai venti.

171

172 FOREST INFLUENCES.

yellow fever, and malaria are, accorclinj? to Dr. Petteiikofer " soil dis-
eases of miasmatic orioin." In this connection a distribution must be
kept in view between those organisms which are disease producers,
pathogenic, and those which are more or less harmless parasites. The
latter, Saprophytic bacteria, it must be kept in mind, thrive on decom-
posing vegetable and animal matter 5 but pathogenic bacteria thrive
best on living organisms, although they occur also outside of them.

The conditions for the favorable development of the pathogenic
bacteria Ebermayer discusses at great length. The facts are stated or
established by him that the vegetable components of the forest soil
contain less nutritive matter (albuminoid, potash, and phosphates and
nitrates) for bacteria growth; that the temperature and moisture con-
ditions are less favorable; that the sour humus of the forest soil is an-
tagonistic to pathogenic bacteria; finally, that so far no pathogenic
microbes have ever been found in forest soil, hence this soil may be
called hygicnically pure.

Only when upper soil strata dry out and a wind, forming dust,
sweeps over them are microorganisms carried into the air; hence, with
less air movement in the forest, we would expect fewer microbes in the
forest air. This expectation is realized in the investigation of Sera-
flni and Arata, who tabulate their countings of bacteria, divided into
three classes— molds, liquefying, and nonliquefying bacteria for 40 suc-
cessive days, from May 6 to July 8 — and find, that with one excei»tional
day, one or two of these classes were always less numerous in the for-
est than on its outskirts and generally from twenty-three to twenty-
eight times less.

With these detail investigations are in accor<l the general observa-
tions in India, where villages surrounded by forests are never visited
by cholera, and troops are being withdrawn into forest stations in order
to arrest the disease which it has l)een found is invited by removal of
forests. Extensive moors, which exhibit similar unfavorable soil con-
ditions, have been observed to give the same immunity from the disease.
With regard to yellow fever the same observations have been made in
our own country.

If it is considered, says Ebermayer, that the coma bacillus which
produces cholera makes great demands in its nutrition and belongs to
the most sensitive bacteria, especially sensitive against free acid, being
destroyed by acid stomach juices, that when dry it dies quickly and is
readdy attacked by decay-producing bacteria, that it prospers best in a
temperature of 80° F. to 194° F., and ceases to grow when the tempera-
ture sinks below 54° F., the protective influence of the forest is readily
explained.

Malaria has been shown by Marchiafava and Celli not to be a
bacterial disease, but to be produced by parasitic protozoa, called by
these authors '' plasmodia," which are formed in the red blood corpus-
cles and when inoculated upon healthy people produce the disease.

SANITARY EFFECT. 173

Although their exterior existence is unknown they probably also
come from the soil, and probably, also, warmth or wet soil, periodically
changing to dry in tlie ui)per strata, are the best conditions for their
development. As long as the water covers the soil there is, therefore,
no danger, which only begins with the recession of the same and the
admission of air necessary for the development of the plasmodia, which
accumulate then near the soil. Since they can not rise very high
houses placed on hills within malarial regions or high above the grouDd
are free from the disease.

The same reasons which make the forest conditions unfavorable to
bacteria growth, namely reduced temperature and moister soil, with
more ready drainage usually, as well as the absence of dust and winds
raising the same and the filtration process shown by Serafiua by which
the bacteria coming from the outside are reduced in number, are also
reasons for the observed beneflcient efiect of forest plantations in ma-
larial districts, such as the Oampagna liomana.

Finally, a word on the value of larger parks in cities. While they do
not, according to these expositions, purify the air directly by the func-
tion of the leaves, as has been claimed, theycertainly establish air cur-
rents that may bring fresher air to the ground; but their hygienic
function consists mainly in reducing the temperature by their shade
and in furnishing better drainage conditions of the soil and i)urifying
the same by absorbing the results of decomposition from animal mat-
ter, and lastly by preventing' or reducing at least dust and with it baC'
teria in the air, keeping it ^aiev than it would otherwise be,

Appendix 1.

DETERMIN\TIONOF THE TRUE AMOUNT OF PRECIPITATION AND ITS
BEARING ON THEORIES OF FOREST INFLUENCES.

Ly Clkvklani) Anr.E.

The accurate raeasnrement of rainfaH, or the confidence to he placed
in the rainfall data that are 'oeing- accumulated in such great quanti-
ties, is a matter of profound importance in climatology, in engineering,
and even in storm predictions. This subject excites much considera-
tion in the public mind, since the comparison of the ancient and pres-
ent condition of Asia, Europe, and America leads to many forebodings
as to our own future; it is therefore eminently worthy of full treat-
ment and critical discussion. We must be satisfied as to the degree of
reliabihty of our data before Ave use them in attempting to elucidate
such Ojueries as are suggested by thoughtful men and such as were in
fact discussed at a recent symposium of the Philosophical Society at
Washington upon the question '<Do forests affect rainfall?"

On that occasion extensive statistics were presented by several mem-
bers of that society, but it seemed to me that certain other fundamen-
tal questions which are discussed in this paper nuist be considered if
we desire to get at the exact truth in regard to this matter.

The questioD, "Whether forests affect rainfall f" can apparently he attacked
from several .sides. Some are satisfied to appeal to deductive reasoning, therel)y
proving that the forests ought to affect raiulall, and their arguments have much
plausibility; others are satisfied to appeal to the historical evidence and quote the
dried-up streams of Europe, Asia and America, and the deserts that were once gar-
dens. But the instrumental mcteorok)gist and the observer who knows that rain
gauges have been carefully observed here and there for the past 300 years, or since
the first gauges were used by Leonardo da Vinci, still naturally look to the
records of the many stations that are accessible to us and expect these to give a
very definite statistical answer that can be relied on quantitatively as well as quali-
tatively.

Such raini'all records miist be the nllimafe test of the truth of any hyl)()th(^sis,
hut the (luestion is one of considerable difficulty since we have todetcniiine a small
quantity by means of observations that may be liable to large errors, and it is nec-
■Bssary to be on our guard against fallacious reasoning and against the liability to

173

176 FOREST INFLUENCES.

draw couclnsions fiaer than the data will warrant. To this end the law of errors
or the calculus of jjrobabilities must be appealed to. My present study relates prin-
cipally to subjects of fundamental importance, namely, the accuracy of the indi-
vidual statistics and the variability of distribution of rainfall chronologically and
geographically.

RELATION Oi" P.VTTERN AXD ALTITUDE OF GAUGES TO ACCURACY OF RAINFALL

MEASUREMENTS.

With regard to the accuracy of rainfall measurements viewed simply as com-
parable data, two matters have been studied experimentally, namely, the size and
style of the gauge and the altitude above ground. With regard to size it is satis-
factorily shown that no error of more than 1 per cent systematically attaches to
gauges of the ordinary forms and of diameters ang^here between 4 and 44 inches.
With regard to the altitude it must be conceded that for a hundred yeai's it has been
known in a general way that observations by gauges nt various heights above the
ground are not comparable with each other. The remarkable influence of altitude
was first brought to the attention of the learned world by Heberden, who, in a me-
moir in the transactions of the Royal Society of Loudon, in 1769, stated that a gauge
on Westminster Abbey over 1.50 feet above the ground caught less than half as much
as a gauge at the ground. Since his day numerous others have instituted similar
observations in their respective localities. Usually they have been satisfied with
observing only one or two elevated gauges, but t>f late years, in order to fully eluci-
date the subject, more elaborate measurements have l>eeu made; thus Phillips and
Gray, at York, England, have observed at eight difi'ereut altitudes including the
gauge on the tower of York Minster.

Rache, at Philadelphia, observed four gauges on top of a s(i[uare tower, and four
others on poles above them; Col. Ward, at Calne House, Wiltshire, observed ten
pairs of gauges at elevations of 20 feet or le.ss, each jtair consisting of an S-iuch
3,Dd ^ 5-inch gauge; Bates, at Castleton Moor, similarly observed ten pairs of
gauges ; Chi'imes, at Rotherham Reservoir, six gauges, at elevations of 25 feet or less ;

( ?) at Hawsker, four 3-inch gauges, at altitudes of 10 feet or less; Wild, at St.

Petersburg, six 10-inch gauges, at altitudes of 5 meters or less, and one at an altitude
pf 25 meters. A very laborious series of six or eight gauges at altitudes of 40 feet
pr less has, to ii^y knowledge, been carried on for some years by Fitzgerald, at Chest-
nut Hip, near Boston, but the results are not yet published.

J^ will be seen, therefore, that abundant observational data are ;^t haml for the
elucidation of the peculiarities of the rair^ gauge, and the results that can be deduced
from sucl^ data command our immediate attention. Whatever uiystery has hitherto
attached to the undoubted fact that elevated gauges catch less rain is now fully
explained away. This phenomenon is of the nature of an error in the rain gauge
dependiiig upoi^ the force of the wind that strikes it, and as will be seen, now that
^he knowledge of the source of error has Ijoen established, the njethod of correcting
Pf preventing it becomes simple.

Jt will be reuiembered that Benjau\iu Franklirt, upon reading Heberden's memoir,
at once, in 1771, in ^ letter to Percival explained his results by the hypothesis that
falling cold rain drops condense the moisture they meet with in the warmer lower
strata, and that Phillips, ii; 1831, independently revived this hypothesis as explaining
the increase of rainfall. A mnch truer explanation had been suggested by Meikle,
in the Annals of Philosophy for 1819, and by Boace (Annals of Philosophy, 1822),
to the effect that the deficiency is due to the velocity of the wind and to the fact
that the gauge stands as an obstacle to the wind; also Howard showed that the
strength of the wind affected the higher gauge. But these minor notices seem to
have produced but little effect among meteorologists, and it remains for W. B. .Tevons,
Phil. Mag., 1861, vol. xxii, to demonstrate that the Franklin-Phillips hypothesis

RAINFALL AND WIND. 177

was hifhly unsatisfactory, and in fact impossible, and that the true reason of dimin-
ution of apparent rainfall with the height of gauge is the influence of eddies of wind
around the building and the mouth of the gauge. This explanation had, however,
been also quite clearly pointed out by Prof. Bache, who had shown that eddies around
the top of the tower affected the distribution of the rainfall on the tower. Alexan-
der Dallas Bache and Joseph Henry were intimately associated in their scientific
work as early as 1835 (and especially after Henry came to Washington, in 1847,) and
the latter had adopted that which is now called Jevon's explanation, although as
we have seen it was first given by Meikle, 1819, and subsequently independently
arrived at by many others. This theory was definitely adopted and disseminateil
by Henry at least as early as 1853 in connection with his instructions to Smithson-
ian observers.

The essence of this explanation may be stated thus : In the case of ordinary rain-
falls we invariably have the air full of large and small drops, including the finer
particles that constitute a drizzling mist and the fragments of drops that are broken
up by spattering. All these are descending with various velocities which, according
to Stokes, depend on their size and density and the viscous resistance of the air; the
particles of hail descend even faster than drops of water and the flakes of snow de-
scend slower than ordinary drops. Now when the wind strikes an obstacle the de-
flected currents on all sides of the obstacle move past the latter more rapidly;
therefore, the open mouth of the rain gauge has above it an invisible layer of air
whose horizontal motion is more rapid than that of the wind a little distance higher
up. Of the falling raindrops the larger ones may descend with a rapidity sufficient
to penetrate this swiftly moving layer, but the slower falling drops will be carried
over to the leeward of the gauge, and failing to enter it will miss being counted as
rainfall, although they go on to the ground near by. Evidently the stronger the
wind the larger will be the proportion of small drops that are carried past the
gauge; or again, the larger the proportion of small drops and light flakes of snow
that constitute a given shower, the more a gauge will lose for a given velocity of
the wind. In brief, the loss will depend both upon the velocity of the wind and the
velocity of the descent of the precipitation ; therefore, a gauge will in general catch
less, in winter than in summer — less in a climate where light, fine rains occur than
where the rains are composed of larger, heavier drops ; less in a country or in a season
of strong winds than of feeble winds; less when exposed to the full force of the
wind by being elevated on a post than when exposed to the feebler winds near the
ground.

The action of the wind in blowing the precipitation over to the leeward of the
gauge depends on velocity rather than on the square of the velocity of the wind and
of the raindrop, and it is aggravated by the formation of whirls or eddies within the
gauge itself by reason of which light and dry snowflakes are even whirled out of the
gauge after being once caught in it. Similar remarks apply to the rainfall on the
top of a large square building with a flat or depressed roof; not only does the top as
a whole receive less than an equal area at the ground, but the distribution of rain-
fall an the roof is such that the least rain falls on the windward portion and the
most on the portion to leeward, while somewhere on the roof will be found a region
whose average rainfall coincides with that on the ground. But the location of this
region will vary with the direction and strength of the wind and the <|uality of the
precipitation, so that we have but little assurance that any single rain gauge on the
roof will rex)resent the rainfall on the ground.

An interesting illustration of this action of the wind has been noted by me in the
case of several gauges established in a cluster in a sandy region. Tlio gauges sat on
the ground; their mouths were 2 or 3 feet above the surface, aud being cylindrical
they offered considerable resistance to the wind. The windward gauges caught h-ss
rain than the leeward, but they also caught more sand, showing that t"he strong
winds wliich carried the light raindrops on beyond also stirred up the light surface
12444— No. 7 12

178

FOREST INFLUENCES.

sand and were just able to drop the sand into the windward gauges while carrying
the rain on to the leeward gauge. In accordance with the preceding explanation all
observations every where show that the higher gauges have the larger deficit in rain
catch and still larger deficits in the snow catch, and that both deficits increase with
the wind.

Mordecai states (Journal Franklin Institute, 1838, vol. xxir, p. 37) tliat he arranged
his observations at Frankfort Arsenal to show the rain catch at the ground and on
the tower 52 feet high according to the force of the wind as estimated by him on the
scale 0 to 10 as used by him, and found the deficit of- the tower gauge to be 0 per ceut
for calms and light winds, but increasing steadily up to 36 per cent for a wind of
force 8.

Bornstein (Met. Zeit. Oct. 1884) arranges the catch in protected and unprotected
gauges according to the velocity of the wind, and for seven mouths of observation
obtained the following deficits in millimeters, to which I add the same converted
into percentages on the assumption that the protected gauge is practically equiva-
lent to the pit gauge. This assumption, although it is not quite correct, will not

much change our results:

Taule I.

Wind

Num-

Protected

Unprotect-

Unvrotect-

Unprotect-

f'oice—

ber of

ganse

ed f;a"?o

ed gauge

ed gauge

0 to 12.

days.

catch.

catcli.

deticit.

dettcit.

JJfjn..

Mm.

3Iin.

Per cent.

0

5

3.30

3.00

0.24

7

1

37

88. 34

81.41

0.90

8

2

26

70.40

03.80

0.60

9

3

15

30. 62

29.89

6.73

18

4

15

43. 45

39. 57

3.88

8

5

1

2.50

1.50

1

40

6

1

1.76

1.38

0.38

22

The distinction between the effect of the winds in heavy rains and fine rains is
very clearly brought out by Bornsteiu's classification of the catch on twenty-six
days of fine rain aud forty-three days of heavier rains; the percentages are as shown
in the following table :

Table II.

Wind
force.

43 heavy rains.

26 fine rains.

No. of
day.s.

Deficit.

No. of
days.

Deficit.

0

1

O

3
4

Percent.

4

8

I
2

Per cent.
23
25

18
46
52

17
13

7
6

6
13
14
17

Mthough all these preceding data, both by Mordecai and Bornstein, are limited
in quantity, yet they conspire to show uniformly the same effect of the wind that is
shown in an exaggerated scale when the ordinary gauge is used to catch snowfall.
Similar results based on a somewhat larger series of observations are published by
Wild (Repertorium for Met., vol. ix), as shown in the following Table III, which
gives the percentage of catch during winter's snow and summer's rain separately for
several altitudes aud wind velocities :

RAINFALL AND ALTITUDE OF GAUGE.

179

Table III.

Altitude

Low wind velocities (2 to
5 meters per aecond) .

High velocities (6 to 9
meters per second).

Rainfall
(Apr.-Oct.).

Snowfall
(Nov.-Dec).

Rainfall
(Apr.-Oct.).

Snowfall
(Jan. -Mar.).

Meters.

?
2

25

Per cent.

100

95

92

81

Per cent.

100

89

86

20

Per cent.

100

94

84

50

Per cent.

100

80

82

16

The.se tables conclusively show the large influence of the wind on the catch of the
rain, to say nothing of its influence on the catch of the snow. It is therefore evi-
dent that the annual rain precipitation as shown by gauges at various heights,
although always dimiuishiug with the altitude, Avill diminish in ditl'erent ratios
according to the peculiarities of the precipitation and the wind in that locality.

Without attempting to go into spc ial refinements it will be sufficient for the
present to study the annual catch as recorded at numerous stations. I have there-
fore in the following Table IV arranged the results quoted by Wild (Repertorium
vol. IX, 1885) and some others that have been published elsewhere. This table
begins with the results of observations made at low altitudes and of these I have
taken the average of all observations made for four years at Calne, three years at
Castleton, eight years at Rotherham, and ten years at St. Petersburg with gauges
of from 5 to 10 inches diameter. I have omitted the observations for two years at
Hawsker with 3-inch gauges, because of the shortness of the series and the smallness
of the gauge, which latter point has slightly exaggerated the percentage of loss. The
combined result therefore for the 4 longer series is to show that for gauges of such
size and construction as are generally used in the weather bureaus of the present
day and for the average wind and snow or rain that occurs at these stations (which
are in fact fair representatives of the northern portion of the temperate zone) the
catch of rainfall diminishes with height of gauge, as shown by the percentage in the
last column and in which, of course, the catch of the normal pit gauge is adopted as

the standard.

Table IV.

Location.

Calne

Castleton ,

Rotlierliam

St. Petersburg .

London : "Westminister Abbey

Etnden

St. Petersburg, Central Observatory .

York : Museum

Calcutta : Alipore Observatory

Woodside : Walton on Thames

Philadelphia: Frankfort Arsenal

Sheerness : Waterworks

Whitehaven : St. James' Church

St. Petersburg, Central Observatory.
Paris: Astronomi<'al Observatory ...

Dublin : Monkstown

Oxford : Kadclitfo Observatory

Copenhagen : Observatory

London : Westminster Abbey .

Chester : Leadworks

Wolverhampton : Waterworks

York : Minster

Boston : St. Botolph Church

5inch and 8-inch gauges.
5-inch and 8-inch gauges.

.'"i-iiu^li gauges

10-iucb gauges

No. of
years.

Altitude.

Relative
catch.

Meters.

Per cent.

f 0

100

41
3l

• 8 '■

10 J

1

90

2

88

{ 3

86

4
5

85
85

I 6

84

1

9.1

77

2

11

72

1

13

68

3

13

80

7

15

87

1

15

73

3

16

95

3

21

52

10

24

66

10

25

59

40

27

81

6

27

64

8

34

59

4

36

67

1

46

52

2

49

61

3

55

69

3

65

60

2

79

47

180

FOREST INFLUENCES.

For gauges higher than 6 meters this table gives the results of the individual lo-
calities. If we consider the individual ligures in the latter part of this table it
would seem that the diminution of rainfall with elevation of gauge is decided, but
irregular; but it is proper to collect the data into a few mean values as shown in
the following table, in which the three higher groups may be considered to repre-
sent the average conditions of the precipitation in the temperate zone q^uite as
fairly as do those of the lower altitudes:

Table V.

Group.

IS'o. of
Btiitions.

Altitude.

Observed
deficit.

Square

rout of

altitiido.

Com-
putod def-
icit.

Obs.Comp.
deficit.

MetciH.

Per cent.

1

4

0

0

0.0

0

0

2

4

1

10

1.0

6

+4

3

4

2

12

1.41

8

+4

4

4

3

14

1.73

10

+4

5

4

4

15

2.00

12

+ 3

6

4

5

15

o o_^

14

+ 1

7

4

6

16

2.45

15

+1

g

7

13

21

3.61

22

—1

9

7

l-'S

36

5.29

32

+ 4

10

5

59

42

7.68

46

—4

If we may assume that on the average of the years and of the localities thus
grouped together there is a fairly uniform average quality of precipitation, we
should expect the deficiency at each altitude to have some definite relation to the
velocity of the wind, and it emphasizes our conviction that the wind is the principal
factor in bringing about this deficit when we find that these normal percentages are
fairly represented by the simple formula : Deficit=6 per cent of the square root of
the altitude expressed in meters. The adoption of the simple square root of the alti-
tude is of course suggested by the well-known studies of Stevenson and Archibald,
from which I infer that for these low altitudes the square root is a satisfactory ap-
proximation to the rate of increase of wind with altitude, while for much higher
altitudes the one-fourth or other power might be preferable. The constant factor,
6 per cent, that enters this formula will of course not be understood as applicable to
other gauges or velocities or qualities of precipitation than those included in the above
table, but the close agreement of the computed percentages of deficiency shows that
we appear to be on the right track, and that some method must be devised by which
to free rainfall measures from the influence of the wind at the mouth of the gauge.
We see, in fact, that the simple wind-gauge which we have trusted so long is liable
to systematic error, whose magnitude is really enormous as compared with the small
errors that we ordinarily investigate in connection with thermometers, barometers,
and anemometers.

ELIMINATION OF ERRORS OF THE RAIN GAUGE.

Two methods are open to us by which to eliminate this error of the rain gauge.
One is instrumental, the other observational.

Instrumental methods.— As before said, Profs. Bache and Henry seem, from their own
observations, to have clearly apprehended the nature of the error with which the gauge
is affected, and the latter was quick to suggest the remedy, namely, to so construct a
gauge that it shall closely imitate the conditions of the normal exposure, or that of a
gauge whose mouth is on a level with the ground, and which is, therefore, not cov-
ered over by the disturbing swift currents and eddies. The records of the Smith-
sonian show that Henry caused numerous experiments on this subject to be conducted
after he and Espy, in 18*48, inaugurated the Smithsonian system of meteorological
observers. In the second volume of Henry's collected writings, the reader can
easily consult his discussion of the erroneous explanations and his own correct ex-

ELIMINATION OF THE EFFECT OF THE WIND, 181

planation of the pheuomeuon. and at page 262 will be found Henry's suggestion of
" the shielded gauge." This shielded gauge was an ordinary small cylindrical
gauge; a few inches helow the mouth of this gauge a horizontal circular plate of tin
4 or 5 inches wide was soldered to it like the rim of an inverted hat; bj' this means
he hoped to ward off the disturbing eddies which would necessarily be formed almost
wholly beneath the flat rim and therefore harmless.

Although Henry's shielded gauge was described at least as early as 1853, yet I
have not found as yet any records or observations made with it, though such
probably exist, as Henry's suggestion was widely distributed among the Smithso-
nian observers.

In 1878 Prof. Nipher, of St. Louis, published the first results of his observations
with his own shielded. gauge, as independently invented by him. He surrounds the
upper portion of the gauge by an umbelliform screen made of wire gauze; the fall-
ingrain strikes on this and breaks up, and falls down to the ground without spatter-
ino- into the mouth of the gauge at the center, while the ganze sufficiently breaks
up the wind currents to maintain a normal condition of the air at the mouth of the
oauffe. Nipher's own experiments with this gauge showed that its catch at a height
of 118 feet above the ground was nearly the same as that of the ground gauge itself.

The invention of the shielded gauge gives us the required instrumental solution
of our problem. Of late years Bcirnstein in Berlin and Wild in St. Petersburg have
experimented very largely with Nipher's shielded gauge and have reported in its
favor. Hellraann, during 1887, also observed with a Nipher gauge, and finds the
effect of the shielding to be very favorable, but not so much so as to make it quite
equal to the ground gauge.

The good accomplished by the shields adopted by Henry, or by Niiiher, can also
be largely attained by a simple system of protection from wind. By " a protected
gauge" I mean an ordinary gauge whose mouth is a few feet above the ground, and
which is surrounded at a distance of a few feet by a fence or screen separate from
the gauge, and whose top is a little above the mouth of the gauge. The protecting
fence is therefore so arranged that it diminishes the wind at the mouth of the gauge
without itself introducing new and violent injurious eddies. Both Bcirnstein, Wild,
and Hellniann have experimented with such protected gauges, the protecting fence
being so constructed that the angular altitude of the top of the fence as seen from
the mouth of the gauge is between 20 and 30 degrees. The catch of the gauge thus
protected always exceeds that of the frfie gauge, so that the correction to reduce it
to the ground gauge is comparatively quite small, the deficit being reduced from 25
per cent down to 3 or 4,

Hellmann has also made the following interesting experiment: The roof of the
Academy of Architecture in Berlin, where the Royal Prussian Meteorological Insti-
tute is temporarily domiciled, covers about 50 meters square, and is not merely flat,
but depressed considerably below the rampart walls of the building; it, therefore,
constitutes a grand protection to any gauge placed near tlie center of the roof, and
accordingly Hellmann finds that in this location gauges catch more than anywhere
else on the roof or the ramparts, and but little less than agaiige on the ground. His
conclusion is that the Nipher, or similar protection, can nearly, but still only partly,
annul the injurious influences of strong winds on the catch of the gauge.

The reduction or correction of rainfall for altitude, as it has hitherto been called,
is therefore really a correction or reduction of the readings of the rain gauge for an
instrumental error due to the wind.

Observational methods. — As an observational method of obtaining the true rainfall
from the gauge reading, aud if it is impracticable to establish a normal pit gauge in
a good location, or if it be desired to determine approximately the correction to be
api)lied to past records obtained from a gauge that still remains in the former place,
the following arrangement offers a fair api)rt)xiniation.

Kthe present gauge has been standing in an open field at a few foet elevation,

182

FOREST INFLUENCES.

place two or more similar qaufies near it, and similarly located as far as obstacles
are concerned, except only that one of these is to be decidedly lower than the old
one and the other decidedly higher. From a comparison of the simnltaueous rec-
ords of any two gauges and their altitudes, we should for each separate rainfall,
rather than for the monthly and annual sums, deduce the normal rainfall by the so-
lution of two or more equations of the form:

Observed catch of gauge = (1 — x "/altitude) X (desired catch of normal pit gauge).
Where x is the unknown special coefficient of deficiency due to wind at that al-
titude— that is to say, having two gauge catches, ci and Cjfor the two altitudes Hi
and flj— we obtain the true rainfall {R) by the formulai:

c, = (l — x^/5^) jj
« Ci={l—xVH.i)R

whence

VUz — VH^

^k

(Ci— C )=C|-f H {Cx—Ci).

If Ci and H] relate to the lower gauge we shall generally have Ci > Cx and fl^i <^ Hi
and the coefficient n will be a positive fraction, whose value is given in the follow-
ing table for such combinations of units as may easily occur in practice.

Table VI.

Altitndu

of njiinr

gauge.

Values of Ji for .iltltudes of lower gauge

1

2

3

4

5

2
.3
4
5
6

2.414
1.366
1.000
0.828
0.689

4. 450
2.414
1.721
].:566

6. 469
D. 4:!8
2, 414

8.474
4.450

10.485

If the present gauge is located upon the top of a building, perhaps the best that
can be done to study the accuracy of its records is to locate other similar gauges so
as to get the average rainfall over the whole roof at the same uniform altitude ; the
next best would be to establish a standard protected or shielded gauge as high as
practicable above the roof.

If a new observing station is to be started then a single shielded or protected
gauge isbetterthana single unprotected one; buttwo more shielded gauges at differ-
ent altitudes afford the means of calculating the correction for wind which will, of
course, be quite small for this style of gauge.

VARIATIONS IN GEOGRAPHICAL DISTRIBUTION OF RAINFALL,

By the combination of records from widely separated rainfall stations we ordina-
rily seek to determine the uniformity or irregularity of rainfall as to its geographi-
cal distribution. The study of horizontal distribution of rain should be first made
by means of simultaneous observations at many stations within a small region. The
most instructive work of this kind that I know of is that just now being carried on
by Hellmann in the " experimental rainfall field" of the Royal Prussian Meteorolog-
ical Institution. This institution was iu 1884 officially transferred from the bureau
of statistics, where it bad been organized by Mahlmann and Dove, over to the bu-
reau of religion, education, and medicine, where it is now intimately connected
with all the scientific and educational work in Berlin and is under the directorship

GEOGRAPHICAL DISTRIBUTION OF RAIN AND WIND.

183

of Prof. W. von Bezold. The expcrimeutal rain field really cousists of tlie city of
Berlin iiiul the country arouud especially to the westward, embracing a region of
about 15 kilometers s<iuare, within which are located Berlin, Spandau, and Potsdam.
The forests on the westward, the intermediate gardens and fields, the valley of the
river Spree, and the citv of Berlin, otter a great variety of surfaces but without any
mountains or hills. The average height of the ground al)Ove sea level is scarcely 50
meters, the average distance of the gauges from each otlier, namely, the mean of all
possible combinations is about 4.5 kilometers, the maximum distance being 11 and
the minimum 0.5.

Within this area Hellmann has twenty-one stations, some of which rejiresent sev-
eral gauges. His work began in 1884, and he adopted as the standard height of the
mouth of the gauge 1.07 meters above the ground. I have selected for study the eleven
stations for which complete records for 1886-'87 are given in Hellmann's Reports. The
accompanying Table VII shows the rainfall for each station for each year and the de-
partures of each station from the annual mean of the eleven. From these departures
we get the probable error of any one annual rainfall as plus or minus 6 per cent of
its own value. That is to say, assuming that the same quantity of rain and snow
fell uniformly over the whole of this small region and that the gauges, if unaftectedby
any error, should therefore agree among themselves perfectly, then their failure to
do so is such that it is an even chance that any given rainfall is discordant from the
average by plus or minus 6 per cent. At first Hellmann suggested that the records
of stations 1, 2, 3, and 4, which Avere in the open land east of the forests showed that
less precipitation fell there than over the forests, attbrdiug an argument for the idea
that the forest attracted an extra amount of rain ; but of the other stations there
were also some that were jirotected by the forests, and next year all of these reported
large rainfalls. Now all of th>sse gauges were at a standard height in open regions
sucli that only the variations in wind proper or in the currents induced by neigh-
boring obstacles could conceivably aflect the catch of the gauge; moreover the diff"er-
ences between the stations were greatest in the winter and least in the summer
months and all the study of the configuration of the ground around the stations tends
to show that the diff"erences in the catches of the gauges were due to the irregular-
ities of horizontal distribution of the strength of the wind as influenced by the sur-
roundings. In other words instead of studying the geographical or horizontal distri-
bution of the total annual rainfall it was safe to assume that that had been uniform for
each year over this small area, and that we are studying simply the horizontal distri-
bution of a deficiency in catch or a rain-gauge error due to very local winds at the
mouths of the gauges.

Tap.lk VII.

Station, Helliiiann'.s niimber.

Observed
precipitation.

Departures.

1886.

1887.

1886.

1887.

1886.

1887.

1

mm.
394
363
387
388
435
438
402
462
473
444
422

mm.
514
507
524
505
492
536
549
509
536
516
516

mm.
30
61
37
36
11
14
38
38
49
20
2

mm .

4
11

6
13
26
18
31

9
18

2

2

Per ct.

1

14
9
9
3
3
9
9

12
5
0

Per ct.
1

2

2

3

2

4

2

6

1

7 ;

1

8

2

9

2

10 . .

2

12

14

1
0

Mean

424

518

336

140

This conclusion is confirmed by examining the records in the summer months
separately from those iu the winter. Local showers are frefiueut during the summer

184 FOREST iNFLUENCEg.

and the irregularities iu liorizoutal distribution are presumptively greatest at that
time. During the winter the extended layers of clouds give us no a priori reason
to expect large irregularities in the geographical distribution of snow fall and rain.
Hellmann's records show that the geographical irregularities in the catch of his
gauges is really least in summer and greatest in winter, thus confirming our convic-
tions that on the average of the year the precipitation is uniformly distributed and
the variations in catch depend on the geograi)hical distribution of the wind at the
gauges during the fall of rain and snow.

The eleven gauges here selected from Hellmann's data were unprotected and uni-
formly 1.07 meters above ground, and it is evident that they would not have neces-
sarily shown a similar discrepancy of 6 per cent among themselves had they been
placed at some other altitude. As the absolute deficits of each gauge increase like
the wind with the square root of the altitude, so also should the apparent irregu-
larities in geographical distribution. But this rule should not be so far stretched
as to assume that gauges at the ground surface would therefore show uo irregulari-
ties in the horizontal distribution of rain, the fact being that there is even for them
an outstanding uncertainty of 2 per cent, which is the total combined effect of all
the irregularities of measurement and the drifting of snow or rain.

In general, then, we conclude that in the case of a number of gauges placed within
a few miles of each other, and of which we know nothing as to the height and exposure,
except that in general the observers have placed them in fairly open situations, there
is no reason to give a preference to the reports of one gauge rather than that of
another, since if the observers are equally reliable the irregularities of catch are
likely to far exceed the errors of careful observers. Again, the probable error of 6
per cent, due to unobserved and uncontrollable irregularities in the action of the wind
on these ordinary cylindrical gauges located 1.07 meters above the ground, indicates
the utmost limit to which any attempt at refinement in drawing annual isohyetal
lines should be carried at present, at least in the climates such as thatof Berlin, and
until the data are corrected for wind effects. Finally, any attempt to deduce from
such gauges the relative rainfall over the forest, the cleared land, the hill and the
valley, can only be successful in so far as we make due allowance for the influence
of the wind and the character of the precipitation.

CHRONOLOGICAL VARIATIONS OF RAINFALL.

What has just been said with regard to geographical distribution holds good
equally with regard to the chronological variations in rainfall. Undoubtedly there
are years of large and of small precipitation, but if we analyze these years we shall
see that they differ, not only in the quantity, but at the same time in the quality of
the precipitation and in the forces of the winds. Until we are able to correct the
measured rain or snow for the wind eff^ect we must include this large source of uncer-
tainty in the catalogue of errors to which our measurements are subject ; thus, in some
years, there may be a heavy snowfall of very light snow flakes falling during strong
wind, and in spite of all our eff"orts to estimate we get too small a record. Again, if
we confine ourselves to the summer rains only, namely, those that directly affect the
growth of plants, we shall find that in almost every long-continued series of observa-
tions at any locality trees, houses, and other obstacles have gradually grown up in
the neighborhood so that the average wind force at the gauge has undergone a steady
progressive diminution and the gauge, therefore, catches a larger percentage at the
close of the series than at the beginning, unless the obstacles were always so near as
to shelter the gauges. I have computed the departure of each annual total precipi-
tation (rain and snow) from the mean of forty-six years at Fort Leavenworth, Kans.
(using post-surgeon's record only) ; of twenty-two years at Spiceland, Ind. (observa-
tions by H. R. Dawson), and forty-two years at Washington, D. C. (observations at
the Naval Observatory). From the mean of these departures it is easy to compute

CHRONOLOGirAL VARIATIONS OF RAIN AND WIND.

185

the 80-«;allecl probable error or departure for auy one year, or the index of variability
of annual precipitation. The results are given in the accompanying Table VIII, and
are interpreted in the following paragraph :

Table VIII.

S(;iUoiis.

Num-
ber of
years.

Le.iveiiwortli

Spicelaiid

AYiisliirigtou .

46
22
42

Average
total an-
nual precip
itation.

Ti)c]ies.
32. 48
39.40
39.48

Probable error
of annual precipi-
tation.

Inches.
6.02
5.47
5.20

Percent.
18
14
13

Probable error of
mean of 49 years.

Inches.
0.86
0.78
0.74

Per cent.
2.6
2.0
1.9

The mean annual catch at the Fort Leavenworth gauge is 32.48 inches, as given by
46 years of observations, which, however, differ among themselves from year to year
so much that it is an even chance that the catch of any one year will differ from this
mean by more or less than 18 per cent of its value or by 6.02 inches ; this 18 per cent
is in part due to actual irregularities in rainfall and in part to the variable effect of
the wind and the irregular proportions of snow and rain ; the actual rainfall is larger
than this catch by an unknown amount depending on the character of the precipi-
tation and the strength of the wind at mouth of the gauge.

It is therefore evident that any conclusion as to a change of climate during these
years involving quantities less than the probable errors of the mean rainfall must be
entirely illusory.

EECOM.MENDATIONS.

Our study of the rain gauge .(section 6) and its errors would have a melancholy
conclusion did it not afford us some suggestion as to the proper methods of determin-
ing and allowing for these errors. In view of our present knowledge we now see
that in establishing new stations better methods of exposure should be adopted and
such as are in fact very different from those that have hitherto been considered
allowable. We must closely Imitate the conditions prevailing at the average surface
of the ground, that is to say, in the order of preference the exposure would be: (1)
the pit gauge ; (2) the protected or the shielded gauge near the ground ; (3) several
protected or shielded gauges distributed over a flat roof; (4) the shielded gauge on
posts considerably elevated above slanting roofs. Moreover, in no case should a
single gauge be relied upon, but in all cases at least two similar gauges at very dif-
ferent heights should be observed. From the records of these two gauges we can
compute the catch of the normal pit gauge by the formula previously given.

As this formula is also applicable to the ordinary and in fact any form of gauge, we
furthermore see that an approximate "correction, needed to reduce valuable past rec-
ords to the normal gauge, may now be determined, if these old gauges are still being
recorded, by at once estaVdishing near them two or more similar gauges at consid-
eraldy different heights; from the records of all these gauges for the next few years
wc may determine, at least approximately, a correction applicable to the past years
of historical records. Finally, we are warned against attempting to draw from past
records conclusions that are finer than the data will justify.

APPLICATION TO FORESTRY.

To this presentation of the rain-gauge question I will add that although the ideas
here given may not be altogether new to those especially interested in rainfall, yet
their application to the special precaution of establishing rain gauges in pairs at two
different altitudes has only been carried out by Ebermayer in Bavaria, and more
recently by Brandis and Blauford in the forests of the central provinces of India, but

186 FOREST INFLUENCES.

1 do not know that tlie formula for correction has as yet been used by them. Blau-
ford phiced his upper gauges at the height of 60 feet, i. e., above the tree tops, and the
lower ones at 1 foot. The result of his first year's observations showed tliat the high
gauges in the forests gave 4 per cent more rain than the high ones in the open fields,
a result entirely iu accordance with the facts and views we have already presented;
the gauges in the forests were at a height above the average toi)s of the foliage of the
trees, decidedly less than the 90 fectliy which the gauges in the open lands were ele-
vated above the ground, therefore the high forest gauges should experience less wind
and consequently catch more rain than the high open land gauges, while the average
rainfall for the whole country averages the same. Again Blauford's low gauge gave

2 per cent greater catch in the forest than in the open land, a result also perfectly
accordant with our views (Met. Zeit., 1888, V, p. 23(3) and serving to di8sii>ate
the last argument in favor of the idea that forests appreciably increase the rainfall
as distinguished from the catch of the gauge.

The preceding study assumes that of the local pairs of gauges one at least is high
enough to escape spattering from the ground, and this is easily attained in ordinary
rains, but iu the case of drifting snow this is difticult, and I williiresent my coiiclu-
eious as to the correction for drift and spatter at some future time.

Appendix 2.

ANALYSIS OF THE CAUSES OF RAINFALL WITH SPECIAL RELATION

TO SURFACE CONDITIONS.

By George E. Curtiss.

Tlie possibility of clianging the amount of raiufall by bnraan agency has in late
years received a great deal of popular attention, and an undercurrent of hypothesis
with regard to it has gradually set in, Avhich itself has stimulated scientific inquiry.
The question has been directed most frequently to the effect of forests and to the
special inquiry as to whether forestatiou increases and deforestation decreases the

rainfall.

In this inquiry the statistical method has been extensively employed, and all the
existing rainfall data bearing upon the question and some that are entirely irrele-
vant have been discussed. Special raiufall observations have also been inaugurated
for the purpose of determining the relation of forests to rainfall. Without giving a
resume of these various investigations it is sufficient to state that no definitive con-
clusion has as yet been derived from them. The one conclusion which the statistical
results seem to yield is that, if forests affect the rainfall the amount of effect has, iu
most cases, not been greater than the amount of probable error in the observations
themselves, and, therefore, the statistics give no assurance that the effect is not an
error of observation. This, however, is a result of importance, for it serves to de-
limit, for the regions to which the statistics apply, a maximum value which the sup-
posed effect of forests has not exceeded. Tliis maxinmm value is, iu most cases, but
a small fraction of the total rainfall, an amount too small to be of any considerable
hydrographic or economic importance. Another reason for the unsatisfactory re-
sults of the statistical investigations is, that they have seldom been com1)inod with
a rational explanation of the process by which a change in the raiufall may be
brought about, and, consequently, they have not helped to clarify the misty meteoro-
logical conceptions which are current thereon. To do this is the object of this
paper.

If in any region the rainfall is increased by a forest cover it must be brought about
either (1) l)y an increase in evaporation, which increase must be precipitated over
the same region, or (2) by a diversion to the forest area of rain that would other-
wise fall in some other locality.

Let us now analyse the causes and conditions of ra in fall, so far as they arc now
understood, and see if the results will not materially aid us in determining how far
and in what cases each of these methods of increasing the rainfall can be operative
and efficient. I say iu what cases, for one of the most important objects of this an-
alysis will be to make clear that an effect that obtains in one climate may be entirely
absent iu another. Ideas as to the causes of rain have been greatly simplified by re-

187

188 FORERT INFLUENCES.

cent studies, and many different conditions tliat are generally ennmerated by the
books as prodnctive of rain have been set aside.

The following quotation, from a recent paper by Blandford, sets forth with great
clearness the opinions that meteorologists are now adopting:

"As a result of a long study of the rainfall of India, and perhaps no country affords
greater advantages for the purpose, I have become convinced that dynamic cooling,
if not the sole caiise of rain, is at all events the only cause of any importance, and
that all the other causes so frequently appealed to in popular literature on the sub-
ject, such as the intermingling of warm and cold air, contact with cold mountain
slopes, etc., are either inoperative or relatively insignificant." (Nature, xxxix, 583).

The ascensional movement of the air requisite to dynamic cooling maybe brought
about by three different processes: (1) by couvective currents; (2) by hills and
mountains; (3) by cyclonic circulation; and I classify rainfall as convective, oro-
graphic, or cyclonic, according as it is due to the first, second, or third of these
causes of ascensional movement. Of course in some cases all of these may be opera-
tive at the same time, but for the purposes of analysis they will be separately con-
sidered.

In a region of purely convective rainfall, the circulation is primarilj^ vertical, and
the moisture evaporated is largely precipitated before being carried away by hori-
zontal currents. Therefore an increased evaporation will be followed by an in-
creased rainfall. Consequently anj- change in surface condition which increases or
diminishes the evaporation will, in such a climate, be followed by a corresponding
increase or decrease in precipitation.

In the case of orographic rainfall, currents that are essentially horizontal are
forced to become locally ascending. The moisture evaporated in the region lying in
the patli of the current is partially or entirely precipitated over the region where
the ascensional movement is developed. Consequently an increased evaporation
will, to a greater or less extent, be restored to the basin by an increased precipita-
tion. The extent to which this will take place, that is, the proportion of the in-
crease of moisture that will be returned to the basin, will depend on the extent and
height of the mountains and the relative frequency of the orographic rain-bearing
winds. For example, if these winds prevail only half of the time, only half of an
increased evaporation can, in general, l)e precipitated as orographic rainfall. It
seems quite possible, therefore, with observations showing the amount of change in
the evaporation and with observations of the relative frequency of wind direction,
to compute with considerable closeness a maximum value which the resulting
change in rainfall can not exceed.

The second point to be considered with respect to orographic rainfall is whether a
forested hill or mountain can originate or divert to itself moisture-laden currents
which, without a forest cover, would not exist or would give their rain to other
localities. Mountains, by means of their heated surfaces, develop upward currents,
but when they are forest covered the observations presented by Mr. Fernow in this
bulletin show that the air is less heated than when the surface is bare. Forests
will, therefore, tend to diminish rather thaji to augment the diurnal currents which
set in upwards toward the summit, and which, by dynamic cooling, precipi-
tate their moisture. Likewise, there is lacking any sufficient reason to supjiose that
lateral currents moving in some other direction will more likely be deflected from
their course and diverted toward this mountain summit because it is covered with
forest and hence relatively cool.

The general class called cyclonic rainfall includes a great variety of rain types
related to a cyclonic circulation, some of which are as yet by no means well under-
stood.

In the ordinary progressive area of low pressure the cyclonic circulation is largely
horizontal, but with an upward component. This upward component produces the
usual rainfall of our cvclonic storms. In these storms the horizontal component of

CLIMATIC ILLUSTRATIONS. 189

the circulation is so large thatthe moisture evaporated over one region is precipitated
over another. Consequently in regions where rainfall is of this type, an increased
evaporation in any region will not be followed by an increased rainfall in that same
locality.

In local thunder storms we have a type of rain related to a cyclonic circulation in
which the vertical component often becomes very large, as compared with the lior-
izontal component. This predominating vertical component is due to convection and
the accompanying rainfall is to be considered as partly or largely convective. Con-
vection induces and imitates a cyclonic circulation which may continue after the
direct convective action has ceased. It becomes desirable, therefore, to separate the
directly convective rainfall, whose amount can be locally increased by increase of
evaporation, from the cyclonic rainfall whose amount can not be so increased. An
approximate separation can be made when the diurnal periodicity of rain is known ;
for an excess of rainfall during the afternoon hours may be classed as convective.

Secondly, we have to consider the question as to whether cyclonic rain storms can
bedetiectedfromone course to anotherby diiference in surface conditions. Evidently
great areas of low pressure can not be aftected and so the question is restricted to
local cyclonic storms of small area. With respect to these the question is to be
answered in the affirmative, for it is entirely in accord with physical principles to
suppose that small cyclonic storms, areas of clouds and disturbance and other un-
settled masses of air with a progressive motion, will tend to follow paths of high
humidity and high temperature, lines of atmospheric weakness which all disturb-
ances will seek, just as seismic disturbances follow lines of weakness in the earth's

crust.

Whether the air over a forested area as compared with a neighboring unforested
area constitutes such a preferable storm path is a question to which a dogmatic an-
swer can not now be given. As shown by observation the absolute humidity is in
general higher over the forest and the temperature lower than over adjacent open
spaces. These arc apparently conflicting conditions— one decreasing, the other in-
creasing the relative density and stability of the forest air. It seems hardly proba-
ble that the resultant Avill have any .quantitative importance. However, the inves-
tigation of the question is a fruitful direction of meteorological inquiry.

The practical point here to be emphasized is that it is not enough to say that the
air over a forest is more humid and that therefore there will probably be more rain-
fall. The increased humidity will be quite ineffective unless this moist air can be
cooled dynamically over the region itself, and thus condensed and precipitated.
This is a necessary additional condition, and the evidence that this will occur in
any relative high degree, either by convection currents or by deflected storm paths
is yet lacking.

CLIMATIC ILLUSTRATIONS.

The equatorial rain belt is the most prominent region with almost exclusively
convective rainfall. The Brazilian forest region, the Aruwhini district of central
Africa, the Malaysian Archipelago, and the valley of Upper Assam in India are in or
near this belt. They have light winds, and the moisture evaporated from the sur-
face is precipitated before being carried to any considerable distance by horizontal
currents. Under these conditions an increase or decrease in the evaporation will
be followed by an increase or decrease in rainfall. But they are the very regions
Avhere any nuiterial change in the evaporation appears to be most difficult to effect.

The evaporation is large. Blanford estimated that for the Aruwhini district
probably over half of the rainfall is duo to the direct restoration of the moisture
evaporated. The surface is Tnaintaincd in a continual state of saturation and evap-
oration proceeds uninterruptedly at a nearly uniforni rate. How will deforestation
affect this evaporation? In the first place, a cleared surface will have a higher tem-
perature, and the winds will have freer play, both of which results will conduce to

190 FOREST INFLUENCES.

increase the evaporation, and, consequently, also to increase the rainfall. The wind
effect, however, will be small because the regions under consideration are in the belt
of calms. On the other hand, if deforestation should initially induce a materially
increased run- oft", so that there should be left much less water to evaporate, the
CA'aporation would be diminished, and conseiiuently also the rainfall would bo
diminished ; and when the rainfall had fallen off the streams would fall to their
previous quantity of discharge. We have, therefore, a possibility of two opposite
tendencies acting at the same time — one to increase the rainfall, the other to decrease
it. The question as to which will preponderate, and to what extent, is apparently
a question to be ascertained for each separate district, for the result in one region
may T)e quite different from what it will be in another. Furthermore, this analysis
suggests that a very promising method of investigating the effect of forestation and
deforestation in districts of convective rainfall appears to be not by observations of
rainfall, but by comparative observations of the discharge of the streams. Finally,
in the equatorial belt, which is the only portion of the world having an almost exclu-
sively vertical circulation, the rainfall is so much in excess of the needs of vegeta-
tion that its possible modification has not the same economic importance as in
higher latitudes. There is no motive to increase it, and if diminished no detriment
would result.

Bordering on each side of the equatorial belt are the regions of the trades, which,
over the ocean, are almost rainless ; but over intercepting land areas, such as Central
America and the Antilles, considerable rainfall occurs. This is frequently difficult
to analyze, but it is largely convective and in hilly regions partly orographic. The
seasonal distribution shows that the rainfall is intimately related to the annual
oscillation of the limits of the trade wind, and that the rainy season requires a special
explanation. With the exception of the well-known tropical cyclones of the seas,
the distribution of pressure over the trade region is unfavorable to the develop-
ment of a cyclonic circulation, and consequently cyclonic rainfall is seldom pre-
sented. It is easily seen that the application of this very general statement to
the (juestion of the effect of surface conditions requires us to consider some special
individual locality. Let us take the island of Barbadoes. This island is 21 miles
long, 14 miles across its widest part, and lies in latitude 13" N., longitude 59° 37' W.
Its interior is hilly and rises at points to over a thousand feet in height. From an ex-
tended series of rainfall records carried on by Governor Rawson the average rain-
fall on the coast is found to be 50 inches, and rising 64 inches on the windward side
and in the central highlands. During three-fourths of the year the northeast trade
wind prevails and the rain comes from that quarter. In October, when the southern
limit of the trade reaches the island, the wind turns to the west and the heaviest
rains occur, making it the wet season. The distribution of rain over the island,
both with the trade wind and the west wind, shows that the rainfall is partly oro-"
graphic, but probably the largest part of it must be considered as convective. On
ascount of the smallness of the island and the prevailing fresh winds, practically all
of the moisture thus precipitated on the island comes from the ocean, and the
moisture evaporated from the island itself is carried away to sea. When the
island was covered with forests the convective action of the island could not have
been greater than at present, because its temperature Avould have been lower, and if
a greater evaporation took place there no appreciable amount of the additional
vapor could be precipitated on the island itself We have every reason, therefore,
to conclude that destruction of forests or any other service change in Barbados is
powerless to sensibly increase or decrease its rainfall.

Passing from the region of trades, we reach latitudes favorable to cyclonic de-
velopment. Here convective and cyclonic action are frequently combined. In the
warmer latitudes and in the summer season, the equilibrium of the atmosphere be-
comes unstable and convection currents are set up which induce an incipient cy-
clonic circulation. Then there is a combined convective and cyclonic rainfall. A3

CLIMATIC ILLUSTRATIONS. 191

we fi;o northward the direct convective feature of the rainfall becomes less pronii-
ueut, and the purely cyclonic rainfall predominates. Over the whole region
orographic rainfall is added to tlie other two classes when hills or mountains are
situated in tlie ])ath of moisture-laden currents. With this general statement tlie
further examination of the effect of surface conditions on the rainfiill in mi<ldle
latitudes must be applied to concrete cases.

In the United States the Great Plains region of Kansas, Nebraska, and the Dakotas
is tliat for wliicli a variation in rainfall has most often been claimed or anticipated.
Many settlers believe that the so-called "rain belt" is moving rapidly westward
with the extension of cultivation and settlement.

The single condition favorable to such an increase of rainfall consists in the
steadier and larger evaporation which no doubt takes place over the cultivated area,
but the other conditions necessary to condense and precipitate this moisture over
the same region are largely absent. There are no hills or mountains to produce
orographic rainfall, and without these barriers the high winds constantly carry
away the moisture evaj^orated from the surface and preciiiitate it far to the east-
ward. Hence, not only is the increased moisture in the air but a small fi-action of
the total rainfall, but also only a small portion of this is gathered into convective
or cyclonic currents and restored again by them to the prairie soil. Evidently a
small amount must be thus restored in every summer rain and the total rainfall in-
creased thereby, but quantitatively this must be very small.

There is one region deticient in rainfall and water sujiply for which claims in be-
half of an actual or possible increase of rainfall due to human agency are less often
made, but which the preceding analysis leads me to believe would not be unreason-
able to anticipate. I refer to the San Joaquin Valley of California. This valley is
flanked by the Coast Range on the west and by the Sierras on the east. The mois-
ture evaporated trom the surface can not escape from the basin, but will be largely
precipitated either over the valley or on the sides of tlie adjacent mountains which
constitute its watershed. If, therefore, the increase in irrigation and in the extent
of cultivated area produces a material increase in the evaporation, it seems reason-
able to expect that this moisture will be restored by an increased rainfall in the
valley and in the adjacent mountain sides. One consideration only wotild appear
to retard and diminish this effect. The inclosure of the valley prevents that rapid
indraft of air which renders possible a rapid vertical circulation. Thus the ac-
tivity of the whole process is rendered sluggish and the total amount of moisture
passing through the cycle from evaporation to rainfall is smaller than with a more
rapid circulation.

In the manner here outlined the possibility of a variation of rainfall may be
investigated for any region, andr with sufficient meteorological data even quanti-
tative values may be computed. The general result indicated by the analysis of
the physical processes involved as well as by the statistical data so far collected is
that the margin of such possible change is very small. In most cases it can not
exceed a small per cent, and in other cases it can not occur at all. It appears, tliere-
fore, that in the estimation of the importance and value of forests, a disjjropor-
tiouate and undue amount of emiihasis has been i)opularly given to such an inllueuce.

INDEX.

Abbe, Prof. Clevpland, detcrtnination of the true
amount of ])reci])itati(>u aiitl its bearing ou
theories of foi'est influences, 175.
Air currents, effects of forests on, 18, 120.
Air humidity:
per cent of, 16.

in woods, glades, and plains, compared, 102.
relative, 103.

Fautrat's observations, 104.
Libiirnau's ronchisions, 104.
Air temptra lures :

in interior of forests, 14, 51.

difference between forests and-open ticlds, 51.

aunual range, 52.

monthly range, 53.

deciduous and evergreen fctrests, 58.

effect of elevation, 60.

effect of young forests, 00.

temperatures in tree crown and above trees,

15, 61,
temperature gradients, 68.
Fautrat's observations above trees, 70.
vegetation, effects on air temperatures, 76.
effect of evaporation, 81.
temperatures in glades, forests, and tields, 83.
Swedish observations, 85.
conclusions, 88.
Alps, French, effect of denudation, torrents, area

needing to be reforested, cost, 150.
Austria, systematic observations in forest mete-
orology, 33.
Becquerel, M., meteorological observations, 27,

49, 73, 103, 111, 119, 121.
Bohemia, observations regarding influences of

forests ujiou precipitation, 126.
Brazil, observations regarding influences of for-
ests upon preLii)itation, 128.
Carbon, annual production by plants, heat ab-
sorbed by assimilation of carbon, 82.
Chinook, effect on temperature, 133.
Climate:

effect of forests on, 9, 17, 23, 24.
general and local, 10.

literature of forest influence on, m(*tJiods of
determining such intlueuce, imperfect in-
struments and observations, 25.
no certainty that a given change of climate is

due to Giiange in forest conditions, 26.
secular changes, 26.
Dr. Liburnau's conclusions, 34.
(See also Meteorology and Water supply.)
Climati<; illustrations, 189.
Climatology, forest, two problems, 23.
methods of investigation, 23.

12144— No. 7-

-13

Curtiss, George E., analy-sis of the causes of rain-
fall, with special relation to surface condi-
tions, 187.
Currents, air. (See Air currents.)
Congress :

of German foresters and farmers at Vienna,

1868, 35.
international meteorological, at Vienna, 1880,
36.
Denudation, effect in French Alps, 159.
Drainage:

influence of forests upon, 20, 141, 157.
retardation of watertlow by forests, 158.
experience in the French Alps, 159.
examples of the influence of forests on water-
flow, 162.

{See also "Water supply.)
Ebermayer, Dr., meteorological observations, 30,

75, 99, 146, 148, 150, 171.
Evaporation :

per cent in field and forest, 16.
influence of forests upon, 20, 96, 132.
annual, in Holds and woods compared with

precipitation, 96, 100, 134.
influenced by wind, 133.
reduced bj' forests, 138.
modified by trees of different kinds and .ages,

98.
amount from bare soil and water surface, 99.
effect of litter on, 99.
Fautrat, M., observations, meteorological, 29, 69,

103, 108, 125.
Fernow, B. E.:

article — introduction and summary of con-
clusions, 9.
relation of forests to water supplies, 123.
notes on the sanitary significance of forests, 171.
Filtration:

process, 139.

Wollny's experiments, 148.
relation to floods, 150.
influence of various soils upon, 150.
influence of grass cover, 150.
influence of snow, 152.
influence of the forest floor, 152.
Forest influences. {See also Meteorology, forest.)
influence upon climate and waterflow, 9.
reasoning from insufHcieiit premises, need of

further exact observations, 10.
water management depend<mt upon forest

maiiagemcut, 12.
summary of conclusions, 13.
ditl'eience of meteorological conditions within
and without the forests, 14.

193

194

INDEX.

Forest inflnences — Continued.

influence of forests upon climate of the sur-
rounding country. 17.

influence upon upper and lower air currents,
18.

influence upon comparative temperature in
fore.sts and open fields. 18.

phenomena of glades inclosed by forests. 19.

influences on water and soil conditions, 20.

sanitary influence. 21. 171.

forests a special kind of surface covering, 23.

methods of investigation. 25.

systematic meteorological observations, 26.

French, 27.

Geiman, 30, 39.

Swcdisli. 31.

Austriiin, 33.

efifei'tson temperature of soil, 40-50.

comparison of eflect of deciduous and ever-
green trees, 47.

temper:itnres of air in forests, 51-69.

temperatures above forests, 09-72.

tree temperatures, 72-70.

relation of Ibrests to evaporation, 90-102.

rehition lo humidity. 1(12-10.').

rainfall in. above, and near forests, 100-118."

windbreak eflect of, 118.

relation to water supplies, 123-170.

intlneiiceon hailstorms, 129.

infliieneeon surface drainage, 20, 157.

sanitary siguiticance of, 21. 171.

relaticm of forest soil to cholera, typhus, yel-
low fever and niala»a — investigations in
India. 172.

value of city parks, 173.

influence oii wind and storms, 118.

influence on evaporation, 138.

efi'ect on rainfall, 175.
France :

observations of Beequerel on meteorological
Iiroblenis. ■_'7. 49, 73, 103, 111, 110, 121.

observations of Krut/.scli and liivoli, 28.

government investigations of forest influences
on climate, 29.

Fautrat's observations above and below tree
tops, 29.

observations near Nancy on rainfall, 126.
Frost, eflects of vegetation upon. 76.
Gauges, rain:

relation of pattern and altitude to accuracy
of measurements, 170.
- why elevated gauges catch less rain than
those near the ground, 176.

action of wind upon, 177,

observations of ^Mordecai and Biirnstein, 178.

observations of Wild and others, 179.

elimination of errors, 180.
Glades :

in forests, special i>henonK'iia of, 19.

temperatures in, 19, 83.
Geneva, observations on tree temperatures at, 72.
Germany :

results of Ur. Ebermaycr's observations on
interior forest climate, 30.

meteorological obsi-rvatiou stations, 14.

instruments used, 39.

Germany — Continued.

observations on soil temperatures, 40.
observations on temperatures in interior ot

forests, 51,
temperature in tree crowns and above for-
ests, 61.
relation of evaporation to forests, 96, 134.
observations on precipitation, 107,113,134.
water capa<-ity of soils and soil covers, 146.
water conductivity of soils, 149.
Gradients, temperature, 68.
Hailstonus, supposed influence of foresta upon,

121,129.
Hamberg, Dr., observations on temperature in

woods, glades, and plains, 85,
Harrington, M. W., review of forest meteorolog-
ical observations: a study preliunnary to
the discussion of the relations of forests to
climate, 23-121.
Heat:

Hamberg's observations on temperature and

humidity, 31, 85.
comidicated relation to vegetation, 76.
amount absorbed in transpiration of plants, 80,
degree necessary to produce evaiior.itiun in

water, 81.
amount absorbed by fiu'csts, 81.
amount produced by vertical sun's rays, 82.
percentage of lieat of vertical sun's rays ab-
sorbed in passing throngli the ai]-, 82.
absorjitioii by assimilation of carlxui. 82.
variations in degrees during tlie ilay and at

night, 83.
degrees in forests and fields, (Sweden,) 85.
{.S'ee also Temperatures.)
Hohnel, investig:itions on transpiration of for-
ests, 78.
Hough, F. B., report on forest climate, 30.
Hnmidity, forests and, 16, 102.
India:

temperatures in wooded and treeless terri-
tories of. 01.
inflnence of forests upon precipitation, 128.
• sanitary influence of forests, 172.
Influence of forests. (See Forests.)
Intereeptiim. {See Water supply.)
Leaves :

time of ai)pearaiice and disappearance in de-
ciduous trees, eflect on temperature in spring
and summer, 60.
Liburnau, Dr. Lorenz, meteorological observa-
tions, 33,71.104.
Liuht, influence on transpiration, 77.
Litter eflects:

on soil temi)erature, 49.
on Water capacity, 147.
on conductivity of soils, 150.
or drainage, 152.

nuiiaria, cholera, I'tc., relaticm of forests to, 172.
Meteorological conference, international, Tienna,

36.
Meteorology, forest :

difference of meteorological conditions within

and without the forest, 14.
influence of forests upon the climate of the
surrounding country, 17.

INDEX.

195

Meteorology, forest— Coiit iiuiod.

iutiiifiice uiMiii water and soil iMiiKliliouj, 20.
sanitary iiitlui-iue ol I'urest.s, -1.
systenialic observations on, 'J(i.

literature of forest meteorology, 24.
inetliDils of solving problems otV^f).

early observations on— Hunter, ScbaeiiH', 2G;
Pietet and Mauriee, 'JG; Kaemtz, Jiravais,
and Tliomas, Borgeau, 27.

later ol'servatious, l'"rauee, Messrs. T5ecf|iierel,
27; Contegril, Bellot, 28; Germany, Hof-
man. I'.aur, Krntzscb, Rivoli, 28.

observations of Matbieu, at Nancy, and of
Fautrat and Sartiaux, 29.

observations of Eberiiiayer in Bavaria, 30.

Cei-man stations for ol)servatiiuis, descrip-
tion of, 30.

(diservations in Sweden, 31.

in -\nstria, deseription of stations. Dr. Lo-
reuz-Liburnau, 33.

results of observations, 33.

eonelnsions of Dr. I,iburnau as to in"ieuce
of forests npou climate, 34.

l>rogramme of observations i)ropii8ed by con-
gress of (iernian foresteis and farmers, 35.

system of Prussian service, 35.

eonfereneo at Vienna, 36.

<d>ieets and methods, pSralle! stations, 37.

I'iilie's cvaiiorometer, 38.

instruments and methods of the German serv-
ice, 31).

Dr. Liburnan's method, 39.

soil temperatures, 40.

mean annual ditlerences of tempei-atnre for
lields and woods at the surface and depths
of C inches and 4 feet, 41.

ettects of forest litter, 43, 49.

differences in the influence of deciduous and
of eveigreen frees upon soil tenii)erature, 40.

etl'ect of elevation above sea lev id, 48.

differences in yonng trees and tdd, 48.

differences of soil temperature under sod and
bare surface, 50.

temperatures of soil in.side and outside of a
forest, 50.

air temperatures in forests and open fields,
51.

annual range, 51.

monthly range, 53.

difference in the influence of deciduous and
of evergreen forests, -58.

effect of elevation above sea level, 60.

effect of young forests, 60.

temperatures in the tree crown and above
trees, 01.

temperature gradients, 68.

temperature above forests, 69.

Fautrat's observation, 70.

tree temperatures, 72.

Geneva observations, 73.

tree trunk a poor conductor of heat, 74.

ditlerences between tree and air tempera-
tures, 74.

vegetation and air temperatures, 76.

heat used in processes of vegetation, 76.

transpiration, 77.

Meteorologj', forest— Continued.
Hohnel's investigations, 78.
theoretical cali'ulation of lieat absorption, 80.
temperatures in woods, glades, and plains, 83.
Swedish observations, 84.
conclusions regarding temperatures within

and witliout forests, 88.
tem])erature8 in wooded and treeless terri-

t(u-ie8, 89.
Woeikofl's observ.ations, 91.
temperatures in relation to parallels of lati-
tude, 92.
relation of evaporation to forests, 96, 132.
annual evajioration in fields and woods com-
pared with pr<'<i|iitation, 90, 100, 134.
percentage of evaporation iu woods compared

with that in open fields, 97. 99, 132, 138.
evaporation as modified by trees of different

kinds or ditierent ages. 98.
evaporation from baie soil and from water

surface, 99.
effect of litter on evaporation, 99.
forests and humidity, 102.
absolute Inunidily in woods, glades, auJ

plains coni]);ired, 102.
relative humidity. lO.t.
Fautrat's observations, 104.
Dr. Liburnan's conclusions, 104.
carbonic-acid gas and o/.one tiikeii uii by for-
ests, 105.
rainfall in,- above, and near forests, lOG.
rain caught under trees and in the open fields,

106.
ditierence in ca.se of evergreen and deciduous

trees, 107.
Fautrat's observations, precipitation over

woods and open fields, 109, 125.
German observations, 110.
inrtuenco of forests on rainfall, 1 11-118.
French obseivations, 112.

precipitation over wooiled and treeless dis-
tricts, 113, 126.
influence of altitudes, 127.
Java and ("elebes, 115.
the Ania/.on, 129.
j ob.servatJ(ms in India, 91, 117, 128.

presumption in favor of forests, 117, 125.

wintl and storms, lis.

windbreak i-IVect of forests. 118, 133.

forest circulation, 120.

influenceon fogs, frost, and ice, on hailstoriiis,

121, 129.
water supply, fixed and circulating. 123.
distribulitui of circulating, how influenced,

124, 130.
relation of forests to wafer supply, 123.
elements of dissipation, interception, 130.
evaporation, 132.

evaporation and precipitation compiled, 134.
transpiration, 130.
relation of forests to snow, 137.
conservative and dissipative influences of

forests on water supply, 140.
distribution of terrestial waters, superficial
and underground drainage, brooks and
rivers, 141.

19G

INDEX.

Meteorology, forest — Coiifinned.

run-off intiuenced by topograiiliy, 113,
water capacity of soils, 144.
of litters, 147.

water conductivity of soils, filtration, 148.
influence of grass cover upon tiltratiou, IjO.
of otlier covers, J 51.
of snow, roots of trees, 152.
of utMilogiiiil stratiflration, formation of
sinings, conditions atfectinu; tlieir flow, 153.
influence of forests upon surface drainage, 157.
retardation of water flow. l.")8,
J'lencb Alps experience, 159.
Mis.si.ssippi bad lands. I.'i9.
influence of forests on water How, examples

of 1C2.
cft'cct of removal of forests in tbe Adirondac

region, 165.
in tlie basin of tbe Schuylkill, IGC.
in New Hampsliire, 1G7.
in other States, 108.
Moisture, quantity of:

returned into the atmosphere from a forest by
transiiiration and fvapoiation from trees
and soil, 16. (See Humidity.)
Miittrich, Dr., influence of growing forests upon

rainfall, 113.
Nancy, France, results of observations on influ-
ence of forests upon precipitation, 108, 112. I'JG.
Observations on forest meteorology : France, 27,
108,111.126; Germany, 30, 40, 51, 61, 96, 107, 113,
134.146,149; Sweden, 31; Austria, 33.
Parks, city, value ot, 173.
Plains, temperature in, 83.

Precipitation :

over wooded and treeless districts. 111.

Dr. I.orey's table, 113.

influence of forests upon, 125.

results of observations in Bohemia, 126.

results of observations near Nancy, France,
126.

results of observations in India and Brazil,
128.

su pi)Osed influence of forests upon hailstorms,
121, 129.

determination of the true amount of precipi-
tation and its bearing on theories of forest
influences, 175.
Hain, not the most satisfactory form of watcravail-
able for human use, 12.

Rainfall :

in, above, and near forests, IOC.

Becquerel's observations. 111.

measurements, relation of pattern and alti-
tude of gauge to accuracy of, 176.

variations in geographical distiibution of 182.

chronological variations of, 184.

analysis of the causes of, with special relation
to surface conditions, 187.

Eaingauge:

elimination of errors of the, 135.
instrumental methods, 180.
observational methods, 181.

Sanitary influence of forests :

in production of oxygen and ozone, 21.

Sanit.ary influence of forests— Continued.

by protection against sun and wind, 21.

by unfavorable effect upon microbes, 21.

notes on, 171.
Snow :

influence. of forests on melting of, 20, 152.

influence on climate and weather, 23.

influence on percolation and drainage, 152.
Soil conuilioiis, inflmnie of forests ujwn, 20, 4J.
Soil temperatures, 14,40.

ditterences of, 87. {See Temporalure.)
Springs, their formation and conditions aflectiug

their flow, 153.
Storms, forests, wind, and, 118.
Summary of conclusions as to forest influeucca,

13.
Sweden :

results of observations on forest meteorology
by Dr. H. E. Hauiberg, 31, 84.
Temperatures, air:

under, within, and abovetree crowns, 14, 61.

in woods, glade.*, and plains, 15, 83.

in neighboring forest and open field stations,
19,51.

mont hly range, German observations, 53.

in deciduous and evergreen forests, 58.

in tree crowns and above trees compared with
open fields, 01. "

gradients, 68.

above forests and near ground, ditt'erciice be-
tween evergreen and detiduous trees, 70.

ditterences between tree and air, 74.

relation to transpiration, 76.

theoretical calculation of heat absorption, 80.

conclusions regarding temperatures within
and without forests, 88.

temperatures in wooded and treeless territo-
ries, 89.

Dr. Woeikoff's conclusions, 91.

temperatures near parallels of latitude, 92.
Temperatures, soil:

in forests and open field.^, 40.

mean annual dift'erences, tables of, 41, 44, 45.

comparison of effect of evergreen and decidu-
ous ti'ees on, 47.

influence of elevation, 48.

ellectsoflitTer, 43,49.

diiference under sod and bare surface, 50.

differences inside and outside of a forest, 50.

temperatures, tree, Geneva observations, 72.
Torrents in the French Alps, 159.
Transpiration:

moisture returned to atmosphere by, 16, 19.

its relation to temperature, 76.

to evaporation, 77.

iurtne-ice of light on, 77.

amount of, with diflerent plants, 78.

.imount transpired by evergreen and decidu-
ous trees and various species compared, 79.

variation from day to day and with the sea-
sons, 80.

beat absorbed in, 81.

amount of water in trees, 136.
Trees :

temperature in the tree crown and above
trees, 61.

INDEX.

197

Trees— Continued.

teiuiicrature of Genova observations, 72.
ditiereuce of tt'niix'iatiire between trees and
air, 74.

(See also Temperature, Transpiration, and
P>aporation.)
condensers of aqueous vapor, dew, and hoar

frost, 121, 124.
interception of rainfall by, 130.
amount of water in, V,i6.
Vegetation :

relation t^ and eti'eets on lieat and water, 76.
amount of heat and water absorlied by, 7C.
etiects of absorption on elimate, 76, 77, 78.
efieets on air temi)eratui-e, 76.
table showing estimated amount of trans-
piration, evaporation, and precipitation of
various i>laiits, 78.
theoretical calculation of lieat absorbed by,
80.
Vines, Dr., table showing effects of sunlight on

transpiration of plants, 77.
"Water:

amount caught by rain gan^es, 17.
evaporation affecting air temperatures, 81.
Waters, terrestrial, distribution of, 141.
Water conditions, inlluence of forests upon, 20.
"Water capacity and eomluctivily of soils and soil-
covers, 140, 144.
Water flow, influence of forests upon, examples,

158.
Water management, its connection with forest
management, 12.

Water supply :
the total, 123.

influence of forests upon the disposition of, 130.
interception, 130.
evaporation, 1S2.
transiuration, 136.
elements of conservation. 137.
efieets of forest in the case of snow, 137.
gener.ll eti'ect of forests in reducing evapora-
tion, 138.
summation of the conservative and dissipa-

tive inrtnenees, 140.
formiition of sprin;^s and condiiions afifecting
their flow, 153.

{See alio Evaporation, Eainfall, etc.)
Weather Bureau, memorial pf the A. A. A. S. to
the Secretary of Agricultiirein reuard to, 11.
Weather, clear and cloudy, etiects on tempL-rature

in glades and plains, 87.
Wind and storms, protection afforded by forests,

118.
Windbrealvs, effects of, upon crops, 118, 134.
Woeikofl", Dr.:

on influence of snow-covering on climate, 23.
temperatiue, in wooded and treeless terri-
tories, 91.
on parallels of latitude, 92.
influence of forests on rainfalls, llj.
Wood, as a conductor of heat, 74.
Woods, temperatures in, 83.
(See also Temperature.)
Wooded districts, iireeipit.ition in, 111.
Wooded territories, temperatures in, 87

New York Botanical Garden Librar

00259 8595