-‘°v“" “ . .
v ’ ‘ ‘ ‘ A A. I

 

.0“

fig lllllllHillllllfllIHIIHIHNIUI{IIIHIIIIUIIHHIIllllHHl

_ O - . ‘ .- 7‘ I
- - -.—-.-.-—/ .9:m~réér\.m'm A-. -

--_ _~'ir‘*mK-e -

£>‘..1..o.'—- --

1 . - d‘
. .
. . .
—.'M‘~h_u q. cf-gw’

  

 

'p‘ I

-
~

 

 

I

  

HE BASIC DENSITY CF SCOTCH PINE

CRWAY SPRUCE, COMMON BIRCH AND

ITE BIRCH IN SOUTHERN FINLAND
Pentti Hakkila

 

 

dug—u... .

:_ .f i
.r‘o'L PIE . ‘
4; 33¢???-

.> x # v

' 0 . . -. '
. . n ‘ . _ .
—. H..-A— ....-.o 4.3.0 .04 ’.‘..b“.: 3*M‘
. . . .

a 1’43

0.. --"A0w. -.
: ; .l A

_» ‘4‘ 2. MAJ.

. h

'15:??-

a?
.431

3."
' 1
$1
4.
.1.
4

fr

4;.
,_..'

8‘..."

. £131

1?.
. u
_ ‘.-—.(“_—-—V~‘AIM

' a I.'
." ~\
A' "
,.

:' a"
‘O

' .r.\
. '0 I!"

-‘h“ «p- .4

. _..-..1,

 

:HE PASIC DEESIE' CF SCOTCH PINE,
NCRWAV SPRUCE, CCKKCN BIRCH AND

WHITE BIRCH IN EQUTHERN FINLAND

Pentti Hakkila

Submitted to
richigan State University
in partial fulfillment of the requirements

for the decree of

EASTER C7 SCIENCE

Department of Forest Products

1953

Preface

In spring 1962 the S m a l l - 3 i z e d T i m b e r
R e s e a r c h C o m m i s s i o n offered to finance
a study on the wood density of different timber assort-
ments if this could be done at the Forest Research Insti-
tute. T h e F o r e s t T e c h n o l o g y D e p a r t—
m e n t o f t h e F o r e s t R e s e a r c h
I n s t i t u t e considered it most eXpedient to estab-
lish first how and why wood density varies within the
stem and between stems. It was thought that the calculation
of the wood density of different timber assortments could
be done as an application of the basic data obtained.

This plan was accepted by the Small-Sized Timber Research
Commission, and the field work was started in the same
spring._The Small-Sized Timber Research Commission then
financed the investigation for three years.

Professor, Dr. P a a v o A r 0, Chief of the Forest
Technology Department, played a decisive role in the
successful outcome of the study, and participated closely
in all the phases of the work. He supported and speeded
up the performance of this study in many ways. Associate
Professor, Dr. V e i j o H e i s k a n e n, then Director
of Research of the Small-Sized Timber Research Commission,
offered valuable guidance, especially in drawing up the
research project and again in the final phases of the
study. In later stages, the course of the work was tutored
by Professor, Dr. K a l l e P u t k i s t o from the
D e p a r t m e n t o f L o g g i n g a n d U t i -

l i z a t i o n o f F o r e s t P r o d u c t s, U n i—
v e r s i t y o f H e l s i n k i.

When the field work of the study had already progressed
for two summers I was awarded through the agency of the
N a t i o n a l R e s e a r c h C o u n c i l f o r
A g r i c u 1 t u r e a n d F o r e s t r y a fellow-
ship of the W. K. K e l l o g g F o u n d a t i o n.

This enabled me to study in the D e p a r t m e n t o f
F o r e s t P r o d u c t s at M i c h i g a n S t a t e
U n i v e r s i t y under the direction of the Chairman
of the Department, Professor, Dr. A. J. P a n s h i n, and

II

Professor, Dr. A u b r e y E. W y l i e many aspects of
wood technology, including wood density. After my return
to Finland, Professor Panshin and Professor Wylie conti-
nued to provide assistance in conducting this study. The
valuable instruction at Michigan State University had an
appreciable effect on the final form of this study.

The statistical treatment of the material was possible
only with the assistance given by the C o m p u t e r
C e n t r e, U n i v e r s i.t y o f H e l s i n k i.
The computer programme was drawn up by Mr. P e k k a
K i l k k i, B.For. ’

The manuscript was checked and corrected by Professors
P a a v o A r 0, V e i j o H e i s k a n e n, A. J.

P a n s h i n and A u b r e y E. W y l i e.

The translation from Finnish into English was made by
Miss P a i v i k k i 0 j a n s u u, M. A., and
checked by Mr. L. A. K e y w o r t h, M. A. (Cantab.).

I extend my sincere thanks to all persons and institutes
mentioned in the foregoing.

Helsinki, October 1966

Pentti Hakkila

III

CONTENTS

Page

List of tables with page references .............. VI
List of figures with page references ............. VIII
List of symbols used in the text ................. IX
I. Introduction ................................ l

1. The need for a wood quality inventory and
the feasibility of making one ............

2. Density as a wood quality indicator ......_ 6

3. Arrangement and limitation of the task ... 11

II. Material and methods ........................ l3
1. Material ................................. l3

2. Methods used in investigation ............ 17

21. Collection of samples ................ 17

22. Determination of density ............. 19

23. Statistical procedures ............... 25

III. Results ..................................... 32
1. Density variation within the stem ........ 32

11. Variables causing density variation

within the stem ...................... 32
12. Density variation in the radial

direction of the stem ................ 4O
13. Density variation in the longitudinal

direction of the stem ................ 44
14. Estimating the average tree density .. 47
2. Density variation between the stems ...... 54

21. The external tree characteristics as
density indicators ................... 54
211. Regression of tree density on

tree age ......OOOOOOOOOOOOOOOOO. 54
212. Regression of tree density on
stem size and form .............. 57

213. Regression of tree density on
crown size ...................... 59
22. The influence of environmental factors
on tree density ...................... 61
221. The influence of stand structure
on tree density ................. 61

IV

Page
222. The influence of site quality on

tree density .................... 63
23. Stem—to-stem density variation
explained by external tree characteris-
tics and environmental factors ....... 68
24. Density as an inheritable prOperty ... 72
3. Density of the most important timber

assortments .............................. 76

31. Density of saw and veneer timber ..... 76

32. Density of cordwood .................. 79

IV. The reliability of the results .............. 84
V. Practical vieWpoints ........................ 90
VI. Summary ..................................... 93
Literature cited ................................. 96

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table
Table

Table

Table

Table

1.

10.

ll.

12.

13.
14.

15.

16.

17.

LIST OF

TABLES WITH PAGE REFERENCES

Information on the average weather
conditions of the investigation logali—
ties. (Padasjoki and Hattula represent
Vesijako and Evo).

Average measurements and properties
of sample trees.

The diameter distribution of sample
trees according to the diameter above
the bark at breast height. ............

Comparison by cores and stems of the
densities determined from increment

cores by the mercury immersion method (M)
and from disks by the water immersion
method (W).

Simple correlation coefficients for
some pairs of independent variables. ..

Percentage summerwood of pine and spruce
in annual rings of different ages.

Density of pine, spruce and birch wood
in different parts of the stem. .......

Density (kg/solid cu.m) at breast height
and average tree density.

Regression of the average tree density
on the density at breast height.1)

Density (kg/solid cu.m) at 25 per cent
height and the average tree density. ..

Regression of the average tree density
on the density at 25 per cent height. )

Coefficients of determination illus-
trating the correlation between tree
density and age.

Regression of tree density on age.

The interdependence between tree density
and stem form and size. Coefficients of

determination when the effect of age is _
not taken into consideration.

The interdependence between tree density
and stem form and size. The coefficients
of determination when the effect of age
is taken into consideration. ( ) For
spruce the age has been used instead of
the reciprocal of age).

....OOOOOOOOOOO

Wood density of trees of different tree
classes in a 50-year-old stand.

Density of pine, spruce and birch wood
on different forest site types.

VI

Page

14

l5

15

22
27
34
42
48
5o
51

53

54
55

57

58

63

65

Table

Table

Table

Table

Table

Table

Table

18.

19.

Per cent of the total tree density
variation explained by tree and stand
characteristics. (Values insignificant
at the five per cent level are given
in braCketS)O ......OOOOOOOIOOOOOOOOOOO

Per cent of the total tree density
variation explained by an increasing
number of external tree characteristics
as independent variables. (The thickened
number refers to the last significant
increase in the explained variance for
eBCh BpeCieS)o OOOOOOOOOOOOOOOOOOIOOOOO

Standard deviation and range of tree
denSity. ....0.........OOOOOOOOOOOOOOOO

Density of saw and veneer logs from
different forest site types. ..........

The effect of tOp diameter on the
density of saw and veneer logs. .......

Density of pulpwood obtained from
different forest site types. ..........

Density of different timber assortments.

VII

Page

70

71
72
76
77

8O
82

Figure

Figure

Figure
Figure

Figure
Figure

Figure
Figure

Figure

Figure
Figure
Figure
Figure

Figure

LIST OF FIGURES WITH PAGE REFERENCES

1.

IO.
11.
12.
13.

14.

The location of the investigation
places (Vilppula, Vesijako, Evo and
Punkaharju.

The sampling method in 10 and 20 m
trees.

Correction of tree density values
measured by the mercury immersion
method.

Correlation between wood density and
percentage summerwood of the increment
core.

Correlation between wood density and
core age.

Correlation between wood density and
ring width of the increment core.

Density of 10, 15 and 20 m pines,
spruces and birches at different stem
heights.

The variation pattern of wood density
in the longitudinal direction of the
stem.

Correlation between tree density and
tree age.

Tree density of pine against age. ....

Distribution of tree density of pine.
Distribution of tree

Distribution
birch.

Distribution of tree density of white
birch.

density of spruce.
of tree density of common

VIII

Page

14

18

23

33
37

38

44

44

55
84
85
86

86

86

LIST OF SYMBOLS USED IN THE TEXT

The following abbreviations are used in the text and in

the figure and table legends.

CB
cm
cu.m
dbh

birches combined

common birch

centimetre

solid cubic metre

breast height diameter above the bark

mean of the differences of paired observations
kilogram

metre

millimetre

number of observations

Scotch pine

coefficient of correlation in simple and multiple

linear regression

coefficient of determination correspondingly (This
value indicates the prOportion of the total varia-
tion in the dependent variable that is explained

by the
Norway spruce (in the figures)

regression)

sample standard deviation

sample variance
standard deviation of the differences of paired
observations (G u e n t h e r 1964)

standard deviation of the dependent variable asso—
ciated with the regression

white birch

see pages 29-31

sample mean
no observation

I. Introduction

1. THE NEED FOR A WOOD QUALITY INVENTORY AND THE FEASIBILITY
OF MAKING ONE

"Quality is the resultant of physical and chemical
characteristics possessed by a tree or a part of a tree
that enable it to meet the prOperty requirements for diffe-
rent end products" (M i t c h e 1 1 1961). The quality of
wood is thus a concept bound with the use of wood.

It is of primary importance in industrial planning to
know not only the volume but also the quality of the raw
material inventories. Only when the quality of the raw
wood obtainable is known can it be utilized to the best ad—
vantage, and only then is there a basis for forecasts of
raw material consumption and the quality of the end product.
When this goal has been achieved, it is possible to price
the raw wood equitably by taking the quality of the timber
into consideration separately in each individual case.

A fair price, again, is the only way of making the wood
producer conscious of quality considerations in growing fo—
rest. At the same time, the chances are increased of di-
recting forest tree improvement aCtivity more towards de—
ve10pment of the quality of wood side by side with the
present activity which is aimed chiefly at increasing the
volume of growth.

As there is no unit by which the quality of wood can
be expressed directly, the end must be attained by a
roundabout route. One property or several prOperties of
wood indicative of the suitability of wood for a certain
use are measured. These properties are called wood quality
indicators. To be serviceable, a wood quality indicator
should meet the following requirements.

- The indicator must give an idea of the suitability of

the wood for various uses.

- The indicator must be a sentitive measure of the varia—

tion in wood quality.

- The indicator must allow accurate measuring at reaso-

nable cost.

- The indicator must be easy to interpret and use,

not only in scientific research but also in practi—
cal forestry.

— The indicator must be applicable to all species of

wood.

Defects in wood affect essentially the quality of seve—
ral kinds of timber. The amount, distribution and size of
knots give an idea of the suitability — or unsuitability —
of wood as raw material for, say, lumber from a certain
tree species. The utilization value of wood is influenced
also by the occurrence of reaction wood, decay, blue stain
and other defects. The purpose of a general inventory of
wood quality, however, is to elicit just the basic starting
point, the quality of defect—free wood and its variation.

The symbol of the quality of wood employed in the case
at issue must thus be a characteristic of clear, defect—
free wood. It would be preferable if a property visible
to the naked eye could be used. Such prOperties are the
growth rate and uniformity of growth illustrated by the
ring width. The growth rate is expressed by the average
ring width, but the uniformity of growth is more difficult
to depict. The ring width can be determined by simple
methods and can often be seen clearly by the naked eye.
Its significance is easy to understand and it can be inf-
luenced by silvicultural treatment. In addition, the ring
width varies over a wide range.

There is a negative correlation between ring width and
pulp yield (Wye g e 1 i u s 1949). Ring width can be used
also for assessing the quality of sawlogs and lumber.
According to the stress—grading rules for Finnish
structural lumber, the average ring width of pine and
spruce lumber may be a maximum of 3 mm in the first and a
maximum of 5 mm in the second grade (S i i m e s 1952).
There are corresponding restrictions in several countries,
e.g. in the grading rules for southern yellow pine and
Douglas fir lumber in the U.S.A. (Standard Grading Rules
for Southern Pine Lumber, 1963 and Standard Grading and
Dressing Rules for Douglas Fir..., 1962). It has also been
shown that a positive correlation prevails between the

average ring width and the knottiness of the sawlog and
lumber (H e i s k a n e n 1954).

When the concept of juvenile wood became known, earlier
Opinions about the correlation between the growth rate,
i.e. the ring width, and the quality of wood had to be
revised. According to R e n d l e (1960), "juvenile wood
refers to the secondary xylem produced during the early
life of the part of the tree under consideration and
characterized anatomically by a progressive increase in
dimensions and corresponding changes in form, structure,
and disposition of the cells in successive growth layers".
Concurrently with these changes, thenegenerally occurs
also a narrowing of the ring from the pith towards the
cambium.

It was concluded earlier that there is a fixed corre-
lation between the quality of wood and the ring width. It
has been found later that this correlation is in many cases
smaller than had been assumed except in the zone of juveni-
le wood where some wood properties change concomitantly
with the ring width. The juvenile wood zone within which
the annual ring narrows from the pith towards the cambium
at the same time as certain other prOperties of the tree
change towards higher quality is, however, fairly narrow.
The time during which a tree grows juvenile wood varies
from species to species, but may be correlated with the
normal life span of the tree (R e n d 1 e 1959).

The correlation between ring width and certain other
properties of wood is considerably weaker outside than in
the juvenile wood zone itself (H i 1 e y 1955, R e n d l e
1959). Juvenile wood has been studied chiefly in pines,
and its occurrence in Norway spruce and birches is still
unclarified. However this may be, it is evident that ring
width alone is not a sufficiently accurate standard of
quality for pine, spruce and birch wood.

The prOportion of summerwood volume to springwood volume,
which is denoted as percentage summerwood, is another pro-
perty to be considered in selecting an indicator of wood
quality. In M o r k's (1928) definition, summerwood inclu—
des tracheids in which the common cell wall in tangential

direction between two cells is exactly half or over half of
the radial width of the lumen. If the joint width of two
cell walls is less than this, the tracheids are considered
to belong to springwood. Mork’s definition was intended
originally for spruce, but it is used commonly for other
softwoods as well.

As summerwood by definition consists of thicker-walled
cells, the unit volume of wood contains more wood substan-
ce the higher the percentage summerwood. When the percen-
tage summerwood increases, the strength prOperties of
wood are improved (J a l a v a 1933). Percentage summer—
wood is used in the grading of southern yellow pine and
Douglas fir lumber in the U.S.A. The requirement for lumber
of high grade strength prOperties, so-called dense lumber,
is that there are "on either one end or the other, not
less than 6 annual rings per inch and not less than 1/3
summerwood. Pieces averaging less than 6 annual rings per
inch and not less than 4 meet dense requirements if
averaging 1/2 or more summerwood" (Standard Grading Rules
for Southern Pine Lumber, 1963 and Standard Grading and
Dressing Rules for Douglas Fir..., 1962). The requirement
concerning the amount of summerwood is more important
than that applying to ring width (Short Course in Grading,
1963). The percentage summerwood of both pine and spruce
in Finnish structural lumber of strength class I must be
a minimum of 24. The corresponding figure for calss II
is 17, whereas for lumber of strength class III there is
no minimum percentage summerwood (S i i m e s 1952).

Both spring— and summerwood cells have their advantages
and disadvantages in paper making, and thus the presence of
both cell types in a certain ratio in pulp is desirable.
According to D a d s w e l 1 and W a r d r o p (1959), the
wood contained by Australian plantation—grown softwoods
should contain 15-50 per cent of summerwood if high—grade
pulp is desired.

The use of percentage summerwood as an indicator of
wood quality is limited above all by its unsuitability for
hardwoods, because the transition between spring— and
summerwood is indistinct in diffuse-porous hardwoods. In

softwoods, particularly in those that have pronounced summer-
wood bands, percentage summerwood affects not only the

amount and quality of the pulp yield and the strength of the
wood, but also its weight, appearance and workability. Per-
centage summerwood denotes, however, only the pr0portion

of the xylem that meets the minimum requirement mentioned

in the definition of summerwood, and it gives no indication
of the quality of springwood and summerwood which also va-
ries.

In addition to growth rate and percentage summerwood,
there are several other prOperties that may be considered
as wood quality indicators. They are properties of which no
idea can be obtained without laboratory apparatus. Fibre
length, cell wall thickness, fibril orientation and the
chemical composition of wood are indicators illustrating the
quality of wood used as raw material in pulp and paper ma;
king (D a d s w e l l and W a r d r o p 1959). Fibril
orientation also influences the properties of lumber
(Forest Products Laboratory 1960). The latter, however,
is so difficult to determine that the practical perfor—
mance of the work alone prevents its selection as an indi-
cator for inventories of wood quality.

None of the properties enumerated above meet fully the
demands to be made of a wood quality indicator. A more
serviceable indicator is the d e n s i t y o f w o o d
which denotes the dry—matter weight per unit volume of wood:
wood density is usually expressed. in grams per cubic cen-
timetre, in kilograms per solid cubic metre or in pounds
per cubic foot. If the volume of a piece of wood is
measured when the moisture content of the wood is above
the fibre saturation point, the term b a s i c d e n —

s i t y of wood is used for the density value obtained.
When the volume of absolutely dry wood is used, the oven-
dry density or d r y d e n s i t y of wood is obtained;
it is higher than the former value on account of the
shrinkage of wood.

When the wood density is divided by the water density
at a water temperature of + 4°C, the s p e c i f i c
g r a v i t y o f w o o d is obtained. This is a unit-

less quantity and consequently independent of the measuring

system. Specific gravity of wood must be distinguished

from s p e c i f i c g r a v i t y o f c e l 1 w a l l

s u b s t a n c e. The former depends on the volume of the

cell walls, the specific gravity of cell wall substance

and the amount of extractives, cell lumina and intercellu-

lar spaces. The latter is practically constant, about 1.53

(B r o w n, P a n s h i n and F o r s a i t h 1949).
The density and the specific gravity depend primarily

on the ratio between the volume of the cell walls and the

volumes of the intracellular and intercellular spaces, sin—

ce the specific gravity of the cell wall substance is

practically constant. A disturbing factor, however, is

the extractives in wood which increase wood density.

2. DENSITY AS A WOOD QUALITY INDICATOR

Density denotes the weight of the wood contained in a unit
volume. Weight itself is a factor affecting wood properties.
Knowledge of the weight of timber is of especial importance
in connection with transport because the load size, the
transport charges and the buoyancy of wood in floating de-
pend on the weight of the timber. It should also be known
when the wood is used as building material. When the tradi-
tional method of volume scaling and weight scaling are
used concurrently, wood density is an essential link which
must be known for comparison of the volume and weight meas—

urementsr
Wood density also gives a picture of the mechanical pro—

perties of wood. There is an especially strong correlation
between density and the strength prOperties of wood such
as the tensile, bending, compression and shearing strength
and hardness (s i i m e s and L i i r i 1952). When the
density increases, certain strength prOperties increase,
too. Wood hardness is directly prOportional to density to
the power of 2 1/4 (M a r k w a r d t and W i l s o n
1935). The strength prOperties of high—density Scotch pine
are twice those of low-density pine (S i i m e s and
L i i r i 1952).

The ratio of the strength to the density oquood is ta-

ken into consideration in stress grading. When Finnish
structural lumber is divided into three strength classes,
the density 1) of pine must be a minimum of 500 kg/solid
cu.m in class I and a minimum of 450 kg/solid cu.m in
class II. The corresponding figures for spruce are 470 kg/
solid cu.m and 420 kg/solid cu.m (s 1 i m.e s 1952).
"Density can be used in the selection of high-grade piling,
transmission poles, and other products where high strength
is of major importance" (M i t c h e l 1 and W h e e —

1 e r 1959).

In the production of structural plywood, density is an
essential factor (H e s s 1965, O r t h 1965). It is
also important to select the outermost boards from strong
and stiff lumber in making laminated beams and arches. On
the other hand, it is possible to place weak lumber bet—
ween the outermost boards. By placing the pine boards in
the beam in the right way, beams are obtained that are
60 % stronger and 45 % stiffer than those in which the
low—density boards are placed in the outer parts of the
beam (S i i m e s 1965).

Density also affects the liquid absorption, swelling and
shrinkage prOperties of wood. Low—density wood absorbs more
water than heavy. The consumption of preservatives nihig-
hest in low-density wood with some preservation methods and
can be calculated when the density of wood is known. In
pine wood the tangential, radial and volumetric shrinkage
and swelling are directly prOportional, while the shrinkage
and swelling in the longitudinal direction of the tree are
inversely prOportional to the density of the wood (S i i —
m e s 1938).

The pulp yield per unit volume of wood depends on the
wood density: a unit volume of wood contains more fibre
the higher the density of wood. For instance, there is a
linear correlation between the sulphate pulp yield and the
density of Scotch pine (E r i c s o n 1962a) and southern
yellow pines (M i t c h e l 1 and W h e e l e r 1959).
Correspondingly the sulphite pulp yield in kg/solid cu.m
of Norway spruce wood can be expressed with the following

 

1) Volume measured at 15 per cent moisture content.

equation: yield = 20 + 433 ' specific gravity + 8 ° chlori-
ne number (K 1 e m 1949).

The correlation between the pulp yield of a given unit
weight of wood and the wood density is weak, in contrast.
It has been noted, however, that the cellulose yield per
unit weight of wood is smaller from juvenile wood than
from adult wood (Z 0 b e l and M c E l w e e 1958).

Wood density can also be used to predict the quality
of the paper obtained from wood. Density gives an indica—
tion of the cell wall thickness compared with lumen
diameter, for dense wood is composed on the average of
thickwalled cells. Thanks go their stiffness, thicker—
walled fibres retain their original rounded cross—section
form in paper-making, while thin-walled fibres collapse
and become ribbon-like in form, which facilitates the
bonding of the fibres (R u n k e l 1942).

This behaviour of the fibres affects the paper pro-
perties decisively. Paper made of thin—walled cells is
dense and opaque and has high tensile and bursting
strength. But the tearing strength, which is an important
prOperty especially in kraft paper, is highest in paper ma-
de of thick-walled cells (M i t c h e l 1 1961). Prolife-
ration of the number of thick-walled cells also kngthens
the beating time necessary in the pulp production and makes
sheet formation more difficult. Simultaneously, the paper
surface becomes rougher and the folding endurance of paper
weakens (D a d s w e 1 l and W a r d r o p 1959). Because
of their pale colour, thin-walled springwood fibres require
less bleaching to reach a desirable degree of whiteness
(E d 1 i n 1965).

The most suitable fibres for a given occasion depend on
the type of paper and its intended use. In fact one cannot
speak of a pulp that is univerally good for all paper
qualities. However, D a d s w e l l, W a t s o n and
N i c h o 1 s (1959) thought it desirable that the
thickness of the cell wall should be less than half of'
the lumen diameter.

If a great bursting strength is required of paper, a
low wood density is advantageous, but for a great tearing

strength dense wood is desirable (H i e t t, B e e r s

and Z a c h a r i a s e n 1960). The effect of the density
of the wood on the tearing strength is, however, conside—
rably less with Spruce sulphite pulp than with pine sulpha—
te pulp (S t o c k m a n 1962). If the smoothness and
closeness of a sheet of paper are the primary considerati-
ons, it helps if a relatively high prOportion of the fibres
are thin-walled (C u r r a n 1938). In the manufacture of
many types of papers, including newsprint and thin trans-
lucent papers with high resistance to folding and bursting
and with good tensile strength, low-density wood is often
preferred (E d 1 i n 1965).

In the viscose process low-summerwood pulps have certain
advantages. They have been found for example to show better
filterability of the viscose. This is not of great
importance in practice, however, as the percentage summer-
wood of spruce wood used as raw material is relatively
low (H e d i n, M a 1 m and W e n n e r b 1 o m 1963).

Wood of low density is the most suitable for the pro-
duction of groundwood pulp (S c h a f e r 1961). Some
types of paper made of groundwood pulp need such thin—~
walled fibres that low-density wood may be suitable despite
the smaller yield (B e s 1 e y 1962).

Both the fibre yield and the quality of paper thus
depend on wood density. High-density wood improves the
quality of some types, while low-density wood is good for
other types, but this does not detract from the usefulness
of density as an indicator of the quality of pulpwood. It
is important to know, on the one hand, what kind of wood
is suitable for a certain use and, on the other hand,
where wood of certain kind is to be had.

The quality of the raw material of particle boards
depends primarily on wood density, and the usefulness of
different tree species can be compared on that basis. For
a constant density of particle board, its tensile strength
perpendicular to surface and bending strength decrease and
the moisture absorption, linear swelling and thickness
swelling increase when the wood density increases (L i i -
r i 1964). Particle boards of the best quality are thus
obtained from low-density wood. A wood density of 300 kg/

10

solid cu.m is ideal for the strength and stiffness of the
board, its dimensional stability and weight (M i t c h e 1 1
1961). However, in manufacturing particle board of a

fixed weight the consumption of raw material is greater,

the lower the density of wood employed.

Wood density also indicates the quality of fuelwood, for
the heat value of wood for each tree species is directly
preportional to the wood density. The heat content of a
unit volume depends principally on its dry weight and
moisture and to some extent on the tree species (H e i s -
k a n e n and J o k i h a a r a 1960).

Wood density is expressed in most Finnish studies as
dry density. For this determination, the volume of the
test piece is measured when the wood is absolutely dry. In
closer correSpondence to real conditions is the basic
density of wood. To determine this, the volume of the test
piece is measured when the moisture of wood is above the
fiber saturation point. The density values were determi—
ned in the present study as basic densities, although the
term density is used for the sake of brevity in lieu of
basic density. When density values are taken from earlier
studies in which the dry density was used, mention will be
made of this. In accordance with the general practice the
term t r e e d e n s i t y or average tree density is
used for the average wood density of the stem in the
present study.

When extractives are removed from wood the density of
the extracted wood can be determined; this value may give
a better idea of, say, the paper properties than that
based on unextracted wood (B a r e f o o t, H i t -

c h i n g s and E 1 1 w o o d 1964). The most important
Finnish tree species, however, contain relatively small
amounts of extractives. The average acetone extract
content of Scotch pine is 2.6 per cent in southern and

4.9 per cent in northern Finland, of Norway spruce 1.5 and
1.9 per cent, respectively, and of birches 1.6 and 2.6 per
cent (Finnish Pulp and Paper Research Institute 1949 a and
1949 b). As the removal of such amounts of extractives

cannot appreciably increase the accuracy of using density

11

as an indicator of wood quality, no effort was made in the
present work, which is restricted to southern Finland only,
to remove the extractives from wood prior to the determi-
nation of density. In the heartwood of ScotCh pine the
amount of extractives may sometimes be considerable,
however (of. J a 1 a v a 1952).

3. ARRANGEMENT AND LIMITATION OF THE TASK

The ultimate object in planning the present study was a
national inventory of the wood density. However, it was
obvious from the outset that it would be more practical
to limit the first phase to a geographically small area in
southern Finland, extending it later to cover the country
as a whole. The purpose of the study was thus, first, to
clarify the variation in wood density and the influence of
different factors on it in a selected investigation area
and, second, to evolve a serviceable method for a wood
density inventory in which standing trees could be used.
The following investigation targets were set for the
realization of these aims.

l. The variability of density within the stem;

2. The variability of density between stems with special
attention to the effect of environment;

3. The density of the most important types of timber;

4. Evolving a method for the performance of a nation-
wide study of wood density using standing trees.

The tree species selected for study were Scotch pine
(Pinus_silvestri§_L.), Norway spruce (Pigeg ables (L.)
Karst.), and the two botanical species of birch, the com-
mon birch (Betulagverrugosa_Ehrh.) and white birch (Betula
cent of the total growing stock in Finland (I l v e s —

s a l o 1956).

The investigation was confined to clear stem wood. De—
fects such as knottiness, reaction wood and decay were
disregarded. The values given here for the density of pi-
ne, spruce and birch wood thus illustrate the basic den-
sity of completely clear, defect-free wood.

12

This phase of investigation was concentrated on a small
geographical area. An endeavour was made, however, to
collect the sample in such a way that conclusions could be
drawn on the basis of the results for the whole of southern
Finland, until the study of the whole country could be
performed. The representativeness of the sample will defini-
tely be checked in later studies of the geographical
variation of wood density.

It has become increasingly obvious since the present
study was started that ensuring future wood production in
Finland calls for a more intensive silvicultural programme
in which artificial regeneration and fertilization of
forests will be of decisive importance. It was therefore
tried to find out, partly by means of literature reviews,
the changes to be eXpected in the density of wood produced
by the future forests, compared with that of the existing
forests.

11. Material and methods
1. MATERIAL

The material for the study was collected in 1962—1964
from the experimental areas of the Forest Research Institu-
te of Finland at Vilppula, Vesijako and Punkaharju, from
private forests between the experimental areas of Vilppula
and Vesijako, and from the Evo forest district of the
State Board of Forestry. These places are included in a
geographically narrow area about 100 km long north-south
and about 250 km across in an east-west direction. The
location of the most important places appears from Figure
1 and from below.

Locality Location

Northern Eastern Height above
latitude longitude sea level, m

0

Vilppula ............ 62° 02’ 24 22’ 110
Vesijako ............ 61° 23’ 25° 03’ 110
Evo ................. 61° 12’ 25° 08’ 130
Punkaharju .......... 61° 47’ 29° 18’ 110

The climatic conditions of the places are representa-
tive of the important forest regions in southern Finland.
As the weather conditions of the growing season may be of
significance for wood quality, Table 1 gives information
on the temperature and precipitation in the localities
under investigation. The data are based on publications by
the Finnish Meteorological Office (1964) and by H e i —

k i n h e i m o.

The sample plots had to be restricted to stands in
which any tree whatsoever could be felled as a sample tree.
At Vesijako and Evo, stands recently flattened by storm
9 were also used. The distance between plots of similar tree
species, forest site type and age class was a minimum of
one kilometre. In the stands thus available for the
research task, the greatest possible number of sample plots
were chosen. The stand density, mean height, mean age and
composition of tree species were determined for each sample
plot.

13

 

14

 

 

 

Figure l. The location of the investigation places (Vilppu—

la, Vesijako, Evo and Punkaharju).

Table 1. Information on the average weather conditions of

the investigation localities. (Padasjoki and Hat-

tula represent Vesijako and Evo).

 

 

 

 

 

 

 

 

 

 

 

 

Locality Period Month Whole year
V VI VII VIII IX
Mean temperature, centigrades
Vilppula 1921-50 8.3 14.1 16.1 14.9 10.3 3.5
Hattula 1921-50 9.2 13.5 16.9 15.0 9.9 3.9
Punkaharju 1921—50 8.5 13.2 16.9 15.2 9.9 3.2
Precipitation, mm
Vilppula 1921-50 44 63 74 75 69 596
Padasjoki 1931-60 40 44‘ 65 71 65 542
Punkajarju 1931-60, 38 53 67 72 61 566

 

 

 

 

Sample trees were collected from mineral soils and

undrained swamps. The forest site types of mineral soils
which were included in the investigation are Qxalisf
Myrtillns_type (OMT), Myrtillus type (MT), and vaneihinn

15

type (VT) 1). The forest site type was determined by using
plant cover analyses. Undrained swamps irrespective of
their quality were accepted for swamp sample plots, but
for pine the selection was limited to pine swamps and for
spruce and birch to spruce swamps. Ten stems were taken
from each sample plot.

The material was grouped according to the tree age,
using the following classification: 21—40, 41—60, 81-100
and 101-120 years. For birch the last class was omitted
as unnecessary. .

The material was small for OMT pine and VT birch and
the youngest age class of the poorest sites. However, they
represent cases which are encountered fairly infrequently
in the practical harvesting of timber. The material also
included pine and spruce of over 120 years, especially
trees growing on swamps. Table 2 gives an idea of the na—
ture of the sample trees.

Table 2. Average measurements and prOperties of sample

 

 

 

 

 

 

trees.
PrOperty Pine Spruce Common White
birch birch
Dbh, CH1 0000000000 1309 14.5 1501 1105
Tree length, m . . 13.4 13.3 17.1 13.4
Crown prOportion ) 43.6 67.6 53.1 52.5
Age, years ....... 68 78 57 54

 

 

31') Crown prOportion = crown length as a percentage of
the total tree height (N y y s s o n e n 1954).

Table 3. The diameter distribution of sample trees
according to the diameter above the bark at breast

 

 

 

 

 

 

 

 

 

 

 

height. ,
ISpecies Dbh, cm ’Total

- 6+ 11- 16— 21- 26- ’31— 36-

5 10 15 20 25 3O 35 40
Number of stems, per cent of the total material

Pine 6 28 28 21 ll 5 1 0 100
Spruce 3 29 29 23 ll 3 l 1 100
Birches 6 34 31 17 9 2 l - 100

 

 

l) OMT is the moistest and most fertile of the mineral soil
forest site types mentioned, VT the driest and poorest.
According to the National Forest Inventory, the average
annual growth at various sites in the southern half of
Finland is as follows: OMT 4.5, MT 4.0, VT 2.9, undrained
spruce swamps 1.9, and undrained pine swamps 0.8 solid
cu.m/hectare (I 1 v e s s a 1 o 1963).

16

Average measurements alone do not give a sufficiently
accurate picture of the sample. Table 3 gives information
on the distribution of sample trees into diameter classes,
the main part of the material being 6-25 cm in diameter.

No attention was paid to the botanical species in
collecting the birch sample trees. Common birch and white
birch were thus included in the sample in the ratio in
which they occurred in the sample plots. The moister the
site in question, the higher was the proportion of white
birch. This affects the average density of birchwood

measured at various sites.

Birch species

Site Common White Total
Number of
trees
OMT ..................... 69 71 140
MT ...................... 92 67 159
VT ...................... 66 4 70
Swamp ................... 22 168 190
Total 249 310 ' 559

From six to ten increment cores were taken from each tree
at different stem heights. Additional cores were taken from
the saw timber part of the tree to establish the density
of sawmill waste from slabs. The primary sample thus
included the following quantities of trees and cores
extending from stem surface to pith.

Species Number of trees Number of
cores
Pine...................... 785 7 700
Spruce.................... 751 7 890
Common birch.............. 249 2 240
White birch............... 310 2 390
Total 2 095 20 220

In addition to the primary sample, different secondary
samples had to be collected to decide special questions
that arose in the course of the investigation. Only the
most important of these are mentioned here.

17

To establish the accuracy of the mercury immersion met—
hod used in determination of the density of the increment
cores, a sample was collected which contained a total of
652 paired observations from different species of trees.
Each pair consisted of an increment core and an adjacent
disk. Correction equations for the densities determined
from the increment cores were calculated on the basis of
this material.

A number of increment cores were selected by lot from
the primary sample to establish the dependence of wood
density on the percentage summerwood and the ring width.
The average prOperties of this sample were as follows.

Number Ring width Percentage

Species of summerwood
cores mm

Pine ........... 336 1.41 24.5

Spruce ......... 356 _ 1.59 17.3

Common birch ... 502 1.57 ..

For these cores, in addition to density, measurements
were taken of the ring width, and of the percentage
summerwood except for birch. To determine the effect of
the juvenile wood phenomenon, measurements were also taken
of the age of the core.

2. METHODS USED IN INVESTIGATION
21. Collection of samples

Ten healthy sample trees were felled in each sample plot.
Leaning trees were not taken because they contain an excep-
tional amount of reaction wood. Age, breast height diameter
above the bark, tree height, and crown prOportion were
determined for each stem. In addition, tree class was deter-
mined in accordance with the following classification: the
height of the trees in the second tree calss is 80—90 per
cent, of the third 70—80 per cent and of the fourth a
maximum of 60—70 per cent of the length of the dominant
trees of the stand, i.e. trees of the first tree class
(I 1 v e s s a l o 1928).

18

The felled sample trees were divided into timber assort-
ments according to the generally used grading rules (cf.
Tapion Taskukirja 1959). The following table shows the
minimum top diameters under the bark of the timber assort-
ments included in the study. The pine and spruce logs
were intended for the production of lumber (of. H e i s -

k a n e n and S i i m e s 1960). The birch logs were
destined for the preparation of veneer and plywood (cf.
Yleiset vanerikoivujen... 1961).

Pine Spruce Birch
Timber assortment Minimum tOp diameter
Logs ................... 5" 5" 6"
Pulpwood ............... 8 cm 8 cm 8 cm
Small-diameter pulpwood 5 cm 5 cm ..
Fuelwood ............... .. .. 5 cm

 

 

 

Distance from stump level, In

Figure 2. The sampling method in 10 and 20 m trees.

There were two types of samples to choose from: a disk
sawn from the stem and a core taken by an increment borer.
The disk method gives greater accuracy, because the disk
is many times the size of a core and in it the inner and
outer parts of the stem are weighted in exactly the correct
ratio (B r i c s o n 1959). In a core the inner parts of
the stem are over-represented.

One of the objects of the present study, however, was to
evolve a method for performing a study of wood density in
the country as a whole. Such a study is possible only when
standing trees are used. Moreover, as transport and
storage of the material limit the sample size in a
comprehensive study, the most practical procedure was to

19

collect the density samples as increment cores extending
from the stem surface to the pith. The inner diameter of the
bore was 5.5 mm.

An endeavour was made, by taking several increment cores
per stem, to establish the variability of density at diffe-
rent stem heights, the density of the timber assortments
prepared from the different parts of stem, the average den-
sity of the stem and the possibility of determining the
average tree density from a single core. The points in the
stem from which samples were taken were related to the
height of the tree. The first sample was taken at stump
height (cf. 1 l v e s s a 1 o 1948), i.e., at the relative
height 0 per cent. The following samples were always taken
from there at distances equal to 10 per cent of the stem
height (Figure 2). For tall trees, the last sample was
taken at a height of 90 per cent of the tree length. No
samples were taken from tOps with a diameter smaller than
3-4 cm. Thus, in the smallest trees, a core taken from a
height of 50 per cent was sometimes the last one.

The core length and the type of timber which the core
represented was noted for every core. From saw logs, two
cores were taken from every sampling point. The average
density of the cross-section surface was determined in
the usual way from one of the cores. The other core was
divided into two parts so that the length of the inner
part was equal to the tOp diameter of the log divided by
two. This part was supposed to represent lumber. The outer
part represented approximately the sawmill waste from
slabs that accumulates from the log.

22. Determination of density

The density of wood can be determined in many different
ways. Summaries of the methods employed have been published
in the literature (e.g. M a r k w a r d t and P a u 1
1946, J a 1 a v a 1952, E r i c s o n 1959 and 1966 a,

F o r e s t B i o l o g y S u b c o m m i t t e e 1963)
and it is unnecessary to repeat them here in detail. In
some methods wood density is determined without measuring
at all the weight or volume of the sample. Such methods

20

are the maximum moisture content method (S m i t h 1954
and 1955), the photometric measurement (M u 1 l e r -

S t o 1 1 1949, G r e e n and W o r r a l l 1964),
microscOpic measurement of the amount of’mitlwall substance
(T s o u m i s 1964, S m i t h 1965) or the use of beta
or gamma rays (B e r s e n e v and F o k i n a 1958,

P h i 1 1 i p s 1960, S a n d e r m a n n, S c h w e e r s
and H o h e i s e 1 1963, H o 1 1 o w a y 1965, etc.).

The mostcmnmonly applied method of determining wood
density is by measuring separately the weight and volume
of the sample. The weight is obtainable easily with the
help of alalance meeting the accuracy requirements. The
volume can be measured in many different ways.

The simplest method of determining the volume of the
sample is to calculate it from the dimensions. This
requires a sample which is regular in form like a
faultless increment core. The core volume is calculated
from the length and diameter. The diameter value may be
the bore diameter, as is the practice in the U.S. Forest
Products Laboratory (M i t c h e 1 1 1958). Measuring the
true diameter of the core by a micrometer instead of using
the bore diameter will, however, give a more accurate
result (W a 1 t e r s and B r u c k m a n n 1964).

Determination of the core volume from its length and
diameter presupposes that the core surface is unbroken and
even. It was noted when collecting the material that small
fragments tend to break off the cores and thus the true
volume of the core is smaller than the volume calculated
on the basis of the diameter measured at the unbroken point.
This was especially noticeable with rapidly growing spruce.
For this reason, a method was chosen which does not
necessitate an even core surface.

The volume of the increment cores was determined in the
present study by the mercury immersion method evolved by
Ericson. The volume of the core is determined by immersing
it with a needle in a bowl of mercury placed on a M e t t -
1 e r K 7 T balance and reading the weight of the mercury
displaced by the wood sample off the Optical scale of the
balance. E r i c s o n (1959) has described the method in
detail.

21

As the densities reported in the present study are basic
densities, the core volumes were measured when the wood was
wet. However, practical facilities were lacking for volume
determinations of green cores. They were therefore allowed
to dry to prevent decay. Before volume measurement, the
cores were soaked in water until they sank. Soaked cores
give practically the same accuracy as green cores (L a r -
s o n 1957, W a 1 t e r s and B r u c k m a n n 1964,

E r i c s o n 1966 a).

The accuracy of the mercury immersion method used for
the volume measurements was established with a control ma—
terial of pairs of samples consisting of a 4 cm thick disk
and an increment core. The disk densities were determined
by the method described by N y 1 i n d e r (1953 b) in
which the disk volume is obtained by measuring the amount
of water displaced by the wet disk. The densities thus
determined from the disks were regarded as correct values
and the core densities were compared with them. The volumes
obtained by this method differed from stereometrically
determined volumes in Nylinder’s studies by no more than
0.1 10.7 per cent.

The inner part of the stem is over—represented in
comparison.with the outer part in a core. If the density
around the pith is denoted by a and at the cambium by b,
it can be shown by simple calculus that the average density
of the part of the stem in question is (a + 2b)/3 provided
that density changes linearly from pith to cambium.
However, the core gives a density value of (a + b)/2,
weighting the inner part excessively. To improve the
accuracy of the result, the core may be divided into several
parts. The density can then be determined separately for
each part and the average density calculated by weighting
each part of the core in correspondence to the cross-secti-
on area represented by it. _

In the present study, cores over 4 cm long were divided
into three parts as recommended by E r i c s o n (1959).
The places at which the cores were cut, starting from the
pith, were core length/\fflIO and core length/\l—2. The
relative areas represented by the core parts were, corres-

22

pondingly, starting from the pith, 10, 40 and 50 per cent
of the total cross-section area. It was thus possible to
eliminate the greatest part of the above error. Using the
same symbols, we obtain the following expression as the
average density (D) of the core.

9 (b-a) 5 (b-a)
D = l a + b + -——~:——-+ ————s__ 20
[5 5 ’V2 ’Vm /
Table 4. Comparison by cores and stems of the densities
determined from increment cores by the mercury immersion

method (M) and from disks by the water immersion method

(W).

 

 

 

Species Difference W-M, kg/cu.m t-value
died
Comparison by cores
Pine ........... + 18 i 21 t222 = 12.50*
Spruce ........, + 14 i 23 t277 = 10.41:
Birhes ......... + 18 i 22 t150 = 12.29
Comparison by stems
Pine ........... + 18 i ll t21 = 7.41:
Spruce ......... + 16 i 12 t3O = 7.14
Birhes ......... + 19 i 8 t17 = 9.79".E

 

The magnitude of this particular systematic error
associated with the method can be calculated from the for—
mulas. If density increases linearly from pith towards
cambium the systematic error decreases to one seventh of
the original error when the core is divided into three
parts as described above.

Short cores (2.5-4.0 cm in length) were cut for density
calculations into two parts only, representing volumes of
30 and 70 per cent. Cores shorter than 2.5 cm were treated
whole. The difference between the pith and the outer part
in such short cores is small and consequently a small
source of error.

It is assumed in the formulas that density changes li—
nearly between the pith and the surface. In fact, the
increase or decrease in density is rapid at first but
slows down increasingly when moving from the pith towards

cambium. Hence, the formulas presupposing linear corre—

23

lation contaidmaximum errors which are bigger than the
expected errors. It can be calculated from the foregoing
formulas that this systematic error should not be greater
than 2 kg/solid cu.m. However, the control material
showed that the total error was considerably greater.

The magnitude of the systematic error of the method
employed can be seen in Table 4. The densities obtained by
the mercury immersion method from cores were on an average
4.1 per cent smaller than the values regarded as correct.
M a r k w a r d t and P a u 1 (1946), S p u r r and
H s i u n g (1954) and E r i c s o n (1959) also reported
that densities measured from cores were smaller than den-
sities measured from larger cubiform bodies or disks.

W a h l g r e n and F a s s n a c h t (1959) did not
find any significant differences between the average core
and disk values. They pointed out, however, that the
relationship was not in a 1:1 ratio at all levels of den-
sity.

Correted tree density, kg/cu.m.

500 _

400

 

300

 

300 400 500

Uncorrected tree density, kg/cu.m.
Figure 3. Correction of tree density values measured

by the mercury immersion method.

24

In the present study, the error was obviously caused
partly by the failure of the mercury to cover the core
surface completely because air bubbles tended to stay on it.
"The condition of the wood surface is certainly of great
importance for determining the occurrence of systematic
errors" (E r i c s o n 1966 a). This increased the amount
of mercury displaced by the core and the volume registered
was too big. The unevenness of the core surface was more
pronounced the lighter the weight of the wood. The mercury
therefore covered most poorly cores of low density. The
error in the core volume measurement was thus dependent on
the core density.

The magnitude of the correction can be arrived at from
regression equations which show the true density of the
sample as a function of the measured core density. The
correlation between the variables conforms fairly well with
the straight line, but when the samples in question are of
very light weight the error grows at an increasing rate
while the true density diminishes. A semi-logarithmic
equation resulted in fact in a significantlflbigger corre-
lation coefficient for pine and spruce than a straight line.
These equations were used to correct the density values
measured by the mercury immersion method (Figure 3).

All the per-stem densities were determined from uncor-
rected core values and then corrected by the below-mentio-
ned equations before the data processing. Core densities
were corrected by their own equations when used for other

purposes than to calculate per-stem densities.

Species Correction equation R2 Sy.x

Pine ..... Log Y = 2.21932 + 0.00100 X 0.9106 11.2
Spruce ... Log Y = 2.22260 + 0.00098 X 0.8368 11.2
33.1 + 0.970 X 0.9312 8.4

Birches .. Y

Y = corrected average tree density, kg/cu.m
X = average tree density measured by the mercury

immersion method, kg/cu.m

The unexplained variation connected with the formulas is
caused by several reasons. A part of the unexplained varia—
tion must be attributed to the inaccuracy of the density

25

values determined from disks, a greater prOportion to the
increment core method. Another factor causing unexplained
variation is the method of collecting the increment cores.
Cores from the same tree were always collected from the

same side of the stem. Although an endeavour was made to
avoid abnormal wood insofar as this could be judged with

the naked eye, the presence of compression and tension

wood and other defects was a source of error in the results.

23. Statistical procedures

The data processing was performed with an IBM 1620
computer at the Computer Centre, University of Helsinki.
The most important of the statistical procedures used in
this investigation is regression analysis. The regression
program computed among other things the following data.

- Mean of each variable '

- All possible sums of squares and products of up to

14 variables

- Simple correlation coefficients for all possible
pairs of variables (Table 5)

- Regression of the dependent variable on all possible
linear combinations of up to ten independent vari-
ables (Y =po +fl, X1 +52 X2 + ----------- ,Ichm )

- Standard deviation and t-value of each constant of
the regression equations

- Simple or multiple correlation coefficient for each
equation

— Standard deviation of the dependent variable associated
with each fitted equation

- F—value of each fitted equation

Tree density was always as the dependent variable in the
analysis (Y). The independent variables (Xi) included in
the regression analysis were as follows.

X1 = density at breast height, kg/solid cu.m
density at 25 per cent height, kg/solid cu.m

>4
m
H

1
height and tree length

X3 = X x X10 = interaction between density at breast

X4 2 age, years
X5 = age squared

26

X6 = logarithm of age

X7 = reciprocal of age

X8 = breast height diameter above the bark, centimet-
res

X9 = Xb/X4=X8 x X7 = interaction between breast height
diameter and tree age

X10 = tree length, metres

x11 = x8/X1O = breast height diameter divided by tree
length

X12 = crown proportion

X13 = tree class (1-4)

X14 = tree class squared

X15 = stand density (0.1—1.0)

X16 : if the site is OMT, X16: 1, otherwise X16:O site

X17 if the site is MT, x17: 1, otherwise x17=o $§i§

X18 if the site is VT, X18: 1, otherwise X18=O

X19 if the site is swame19= 1, otherwise X1920

X20 = X16 x X7 = X16/X4 interaction between tree age

X2, = X17 x x, = X17/X4 and site quality (pine and

X22 = x18 x X7 = X18/X4 birches)

X23 = x19 x X7 = X19/X4

X20 = X16 X X4 = X16/X7 interaction between tree age

X21 = X17 x X4 = X17/X7 and site quality (spruce)

X22 = X18 X X4 = X18/X7

X23 = X19 X X4 = X19/X7

The calculations

concerning density variation within the

stem, density of different timber assortments, and the

secondary samples were performed without computer help.

In all tests of this investigation the level of signifi-
cance is 0.05. An asterisk after a test value indicates
that the result of the test is significant at the five per

cent level.

sedwm m. mascwm oosHmHMaHos oomwwwoamsdm won mosm cmHHm ow abambmstbd

<mw~mvwmm.
<mswmawm Mn xm xw Ma Mm Mm N4 Mm Mm Nno Nnn Mnm
maswwm ooddmwmawos oomfifiaowmba

wwbm
xn H.oo .mm .sm ..mw .ss .mm I.mn .mm I.ws .wo I.o< I.wo
Mm I H.oo .ma .sq .wo .mm I.ms I.oo I.sm .o< I.HH I.mm
NW I I H.oo .wm .mm .ss I.s@ .mm .ww .mq I.om I.wm
Me I I I H.oo .mq .oq I.m© .wo I.mH .mm .ws I.ww
rm I I I I v.00 .mm I.<< .wm I.sm .Hm .so I.mw
No I I I I I v.00 I.©< .sw I.mo .wm .m< I.w<
x4 I I I I I I H.oo I.ss .sm I.sw I.Hm .sw
xm I I I I I I I H.oo .mo .mm .aw I.Om
we I I I I I I I I H.oo .sw .mo .mq .
who I I I I I I I I I H.oo I.o< I.ww
Nnn I I I I I I I I I I v.00 .sw
xnm I I I I I I I I I I I 4.00

Aao cm oobaabfimav

L8

Acosa.v

emawm m.

<msamswm

b<><$><><><1><><1
CD-fl O\\fi-b Care A

A «3
<3

MNNNN

_s
[\3

weapwm ooHHmHMaHos oommwwowmba wow moan Seven 0% abampmbambd

an

as

<mwwMGHmm.

Mm Mm XQ Mm

mascwm oowsmwmdaob oomwwaomea

Mm MW
.0H .0H
v.00 1.0m
I H.OO

.ss
.3
1:

Foo

mpwsom
.sw .ew I.so I.ma
.ww .wm I.wm I.mm
.wm .aq I.sm .mo
.mm .wq I.©o .wm
H.oo .mw I.mo .WH
I H.oo I.o< .wm
I I H.oo I.wm
I I I v.00

Mm who
I.mm I.mo
I.mm I.mw

.wm .04
I.mo .ws
I.sm .mm
I.mo .wq

.aq I.wm

.wm .wa
H.oo .mH

I H.oo

Ln

I.Hm
I.H<
.O¢
.Hm
.Hw
.00

.AO
.mq
.om
H.oo

mm

I.mm
I.wo
.oo
I.ww
I.m©
I.wm
.wm
.Hw
.wq
.Om
.mm
H.oo

88

Aoosdv
sedwm m. wasnwm oossmwmdaos oommwwowmbdm wow mosw meHm ow asamvmbambd

<mn~mswmm.

<mw~mawm N4 Mm NW Ms Mm Nm N4 Nm Nm Mno Nan Nnm

meanwm ooHHmHMaHob oomwwwowmba
oossob demos

an H.oo .m0 .m6 .sc .wn .se I.se .mm I.om .mm .Hm .os
an I H.oo .wm .sw .so .sm I.sa .mo I.Hm .me .Hm .06
an I I p.00 .sm .wm .sm I.me .mm .sn .ms .ss .es
as I I I H.oo .mm .mm I.mH .sw I.wm ..wh .so .om
as I I I I H.oo .mw I.ma .wa I.wm .mm .us .om
as I I I I I H.oo I.mm .ss I.ww .sw .so .om
as I I I I I I H.oo I.ss .wm I.sm I.wn I.Ho
am I I I I I I I H.oo .mw .mm .ms .ws
am I I I I I I I I H.oo .mw .mm .ae
ale I I I I I I I I I p.06 .sn .sm
all I I I I I I I I I I H.oo .Hs
arm I I I I I I I I I I I H.oo

68

Aoobd.v

abdwm m. meanwm ooadmwmdwo: oomwmwowmbsm wow mosm meHm ow abampmstba

<MHHMde

NNM

:4 x >4Ca

N >4 x >4
AJKOCDNONUT-P-WM—A
--‘O

>4
R)

N

<mwwmdwmm.

NM NW N» Mm Mm Nu Mm No

meanwm oossmwwawob oomwwaowmba
éfiwdm dwsofi
omw oWaN 0W0 0mm OWN IIOIWM cow I-OHA.

H.oo .mw .ws .wm .wm I.w> .Om I.mo
I H.00 .wm .mw .ww I.sm .mH (so
I I p.00 .mm .mm I.mm .wm I.ww
I I I H.00 .mm.1.ma .we I.wm

I I I I H.oo I.om .sH I.wm
I I I I I H.oo I.aw .wm
I I I I I I H.oo .00
I I I I I I I H.00

no

.Hm
.Om
.w4
.WH
.mm
.wm

I.wm

.mm
.mm

Ln

.OW
.Om
.wm
.wq
.wm
.wm

I.w©

.qq
.mo
.wm
H.oo

nm

loom
I.ow

06

III. Results
1. DENSITY VARIATION WITHIN THE STEM

11. Variables causing density variation within the stem

As mentioned before, percentage summerwood is one of the
most important factors affecting the density of softwoods,
for summerwood density is 2-3 times the density of spring—
wood. T r e n d e l e n b u r g and M a y e r - W e g e -
l i n (1955) reported a number of studies in which the
ratio of the dry densities for summerwood and springwood
was 2.3-3.0 for Scotch pine and 1.9-2.8 for Norway spruce.

Spring- and summerwood are not distinguishable with
equal clarity in diffuse—porous hardwoods, but the wood
forming in them at the beginning of the growing season
does differ from the wood that grows later in the summer.
The relative volume of vessels compared with other cells
decreases towards the outer margin of the annual ring,
especially in ring—porous but also in diffuse-porous
hardwoods. There is, on the other hand, a negative corre—
lation between birch wood density and the vessel area
(W a 1 1 d e n 1934).

In the present investigation, 53 per cent of the density
variation of the increment core in pine wood and 40 per
cent in spruce wood can be attributed to variation in per—
centage summerwood when the dependence between density and
percentage summerwood is denoted by a straight line. The
Change in percentage summerwood affects density more
sharply in Spruce wood than in pine wood (Figure 4), for
the slope of the regression line is significantly steeper
in the equation for spruce than in the equation for pine
(t690 = 4.46I).

Species Regression equation R2 Sy.x

Pine ........ Y = 307.3 + 3.49 X 0.5279 27.3

Spruce ...... Y = 296.9 + 4.61 X 0.4019 32.3
Y = wood density of increment core, kg/solid cu.m
X = percentage summerwood

33

Density, kg/cu.m

400 _,

 

300 .I
\
I I I I 1 I I 1
:30 30 40

10
Percentage summerwood

Figure 4. Correlation between wood density and percenta—
ge summerwood of the increment core.

In L a r s o n's (1957) study 60 per cent of the varia—
tion in the density of Pipus‘Elligttii could be attributed
to percentage summerwood, and S c h a f e r (1949)
succeeded in giving the corresponding explanation for 72
per cent of the density variation of pulpwood bolts made of
southern yellow pines. In P i 1 l o w ’s (1954) study on
Einus_ta§da the comparable figure was 38 per cent. Van
‘B u i j t e n e n (1964) determined the regression between
density and percentage summerwood in Pipus_Elliottii
separately for adult wood and juvenile wood. The value 0.79
was established for the coefficient of determination for
the former, while the value for juvenile wood was only 0.27.

In the present study, using the density variation of
the increment core, a smaller prOportion of the total
variation than in some other studies could be explained
by the variation in percentage summerwood. However, it is
evident that a considerable part of the wood density
variation is caused by the variation in percentage summer-
wood.

All wood density variation cannot be explained by per—.
centage summerwood for, in addition to the proportions
of springwood and summerwood, the density of springwood
and of summerwood also varies, depending on different
factors. "Density of summerwood in the annual ring is

related to summerwood tracheid characteristics. Negative

34

genotypic and phenotypic correlation exists between summer-

wood density and tracheid radial lumen diameter, radial

Table 6. Percentage summerwood of pine and spruce in annual

rings of different ages.

 

j“

 

 

 

 

 

Pine 4 Spruce
Ring age, years Percentage summerwood
X i S X i S
1 - 5 ........................ 16.3 1 5.1 114.7 i 6.9
6 - 10 ........................ 23.7 i 5.9 16.2 i 5.9
11 - 15 ........................ 26.7 i 6.3 18.1 i 5.7
16 - 2O ........................ 28.7 i 6.5 19.5 i 4.9
21 - 30 ........................ 28.6 i 6.0 21.5 i 4.5
31 — 4O ........................ 30.6 i 6.6 22.6 i 4.6
41 - 50 ........................ 29.5 i 9.8 22.7 1 4.9
51 - 60 ........................ 27.7 i 9.2 23.5 i 5.2

 

width, and tangential width. Double—wall thickness is
phenotypically and positively correlated with density but
no genotypic correlation is indicated. In springwood
radial lumen diameter, tracheid radial width, and tracheid
tangential width are phenotypically and genotypically
correlated with density in a negative direction. Double—
wall thickness is phenotypically and genotypically correla-
ted with density in a positive direction" (G o g g a n s
1962).

There is a general tendency for both the density of
summerwood and its percentage in the annual ring to
increase year by year up to a certain ring age (R e n d -

l e 1959). This causes variation in the wood density in
the radial direction of the stem. The change in percentage
summerwood in pine and spruce from pith to cambium can be
seen in Table 6.

The percentage summerwood of Scotch pine is exceptional-'
1y low in the vicinity of pith. It increases from the pith
with age (3 i r e n 1959, D i e t r i c h s o n 1964).
The prOportion of summerwood increases to 28—29 per cent
within the first 15-20 annual rings.

35

The percentage summerwood on pine in southern Finland
seems to remain at 28-29 per cent from the 20th year
until at least the 60th. But several investigations have
shown that the percentage summerwood of many tree species
decreases at a later tree age (L a r s o n 1957). The
percentage summerwood of pine in southern Finland is 28.2
according to J a 1 a v a (1934 b) and 21.9 according to
S i i m e s (1938).

The increase in percentage summerwood from pith to
cambium is also evident in Spruce. The change is not
equally abrupt as in pine, but it seems to continue longer.

There is a causal relationship between the low percen—
tage summerwood in the vicinity of pith and juvenile wood
formation. Juvenile wood is characterized by thin cell
walls, less dense summerwood, a gradual transition from
spring- to summerwood and a small prOportion of summer—
wood (R e n d l e 1960). The density of juvenile wood is
therefore low. A progressive increase in cell dimensions
from pith to cambium occurs within the juvenile wood zone.
The consequence is an increase in density in the radial
direction.

In pines the gradual transition from juvenile wood to.
adult wood causes a quite distinct increase in density in
the area of the first annual rings. The juvenile wood zone,
however, is narrow, 5-8 annual rings in Pipus_Elligttii
and 8—11 in Pinus_ta§da (Z c b e l, W e b b and H e n -

s o n 1959). For Scotch pine, the width of the juvenile
wood zone has not been established, but as it is a longer-
1ived tree species its juvenile wood zone can be assumed

to contain slightly more numerous annual rings (cf.

R e n d 1 e 1959). According to an anonymous investigation
(1957), the annual ring of Scotch pine generally attains
the normal adult pattern between the tenth and the
twentieth years. .

Almost all studies of juvenile wood have been concerned
with pine species. Its significance may be smaller for
certain other tree species.

Wood density increases from pith to cambium also outside
the juvenile zone. The following regression equations Show

36

the dependence of the average density of the increment core
on the core age. The age of pine and spruce cores in the
sample was 4—100 years and that of common birch cores 4-75
years.

Species Regression equation R2 Sy.x.

344.8 + 2.220 x — 0.01477 x2 0.4049 27.2
358.8 + 0.828 x — 0.00376 x2 0.0704 40.1
424.5 + 2.999 X — 0.02400 x2 0.2715 34.2

Pine OOOOOOY
Spruce .... Y

Common birchY

Y 2 wood density of increment core, kg/solid cu.m
X = increment core age, years

The best coefficients of correlation were obtained from
the above equation models. Even then, in spruce wood it
was possible to relate only 7 per cent of the variation in
the average density of the increment core to age. In pine
wood the variation explainable by core age was 40 per cent
of the total variation.

The average density of a pine increment core is at its
maximum when the core is some 80 years of age, after which
the density begins to decrease. The corresponding point
of culmination is reached in spruce a few decades later,
in common birch a little earlier (Figure 5).

It can be assumed from the foregoing that the average
density of a whole tree also reaches its maximum at a
certain age. However, this age limit must be considerably
higher for the whole tree than for an individual increment
core.

Density, kg/cu.m

.. CB

 

500 -I
1
.I
f3
400.. S
q
‘k

I 1 I I

_
q
d
-
q
o.-

100

9'"
0

Core age, years

Figure 5. Correlation between wood density and core age.

The question of the effect of ring width on wood density
has aroused a great deal Of controversy. Knowledge of this
matter is of primary imputance for practical silviculture,
but Opinions on the subject are still contradictory. The
discovery of juvenile wood formation has given rise to new
points of view, but there is no unanimity.

Numerous investigators have concluded that there is a
negative correlation between density and ring width in pine
and spruce wood. For Scotch pine only, this view is
advocated by H a r t i g (1884), K i n n m a n (1923),

L u n d b e r g (1928), V o 1 k e r t (1941), B u r g e r
(1948), Y 1 i n e n (1951), M o 1 t e s e n (1957),

E c h o 1 s (1958) and N y l i n d e r (1961a), and for
Norway spruce by H a r t i g (1884), C i e s 1 a r and

J a n k a (1902), K i n n m a n (1923), V o l k e r t
(1941), K 1 e m (1944), w e g e 1 i u s (1949), N y -

1 i n d e r and H a g g l u n d (1954), A 1 d r i d g e

38

Density, kg/cu.m

 

j
: (CB
j
4
d F3
t s

 

Ring width, mm

Figure 6. Correlation between wood density and ring
width of the increment core.

and H u d s o n (1955 and 1958), M o 1 t e s e n (1957),

H a l e (1962) and B e r n h a r t (1964). On the other
hand, it has been observed in some studies that the den—
sity of Scotch pine is small when the annual rings are very
narrow. According to S i i m e s (1938), the density Of
pine is thus highest when the width of the annual ring is
1.0—1.2 mm, and according to S k r i p e n and R i a s o —
v a (1958) when it is 1.4 mm.

In the present work, the correlation between wood den—
sity and ring width was calculated from a material in
which the ring width was 0.3-3.5 mm and the increment co—
res were under 100 years old. The regression was denoted
by a straight line.

2

Species Regression equation ' R Sy.x
Pine 000000000. Y=434o6—‘2907X 0.2823 3307
Spruce ........ Y = 410.7 — 21.5 X 0.1743 37.8
Common birch Y = 507.1 — 19.6 X 0.0969 42.1
Y = wood density of increment core, kg /solid cu.m

H

X ring width, mm

39

The existence of the juvenile wood phenomenon was not
taken into consideration above, However, naked eye
examination of the cross section of the wood reveals that
the annual rings are wider than the average in the vicinity
of pith. The only exceptions are trees which, when young,
have grown slowly in the shade of large trees, such as
undergrowth spruces. As wide annual rings are often a
feature of juvenile wood, juvenile wood formation must
also be considered in analysing density and ring width.

According to T u r n b u 1 l (1947), the density of
Pings_patula increases outwards from the pith in proportion
to the age or number of rings from the pith. As ring width
nearly always decreases in the same direction, "there
appears to be a relation between narrow rings and high
density, but this relation is primarily accidental". "The
failure to recognize the existence and the properties of
juvenile wood has led to many of the conclusions where
rate of growth has been found to be strongly related to
wood density" (G o g g a n s 1961).

A surprising change occurs in the structure of regres-
sion equations when the effect of juvenile wood is elimina—
ted by excluding from the material the pith part of cores
cut into three (Figure 6). The correlation between wood
density and ring width is then smaller in pine and birch
wood but considerably greater in spruce wood than the above.

Species Regression equation R2 Sy.x
Pine ........... Y = 432.1 - 23.0 X 0.1098 37.7
Spruce ......... Y = 420.6 — 31.1 X 0.4242 27.3
Common birch ... Y = 508.2 — 14.7 X 0.0650 40.2

Y = wood density of increment core, kg/solid cu.m
X = ring width outside the juvenile wood zone, mm

Studies in which the juvenile wood phenomenon has been
taken into consideration have generally established that the
correlation between the wood density of pines and the annual
ring width is smaller than previously assumed. V o l k e r
(1941) assumed that this correlation may be only ostensible.

The correlation was later found to be fairly weak for

40

several pine Species (C o c k r e 1 1 1944, S p u r r
and H S i u n g 1954, B a n k s 1955, B a n k s and

S c h w e g m a n n 1957, L a r s o n 1957, C o o p e r
1960).

The present investigation shows, however, that no
conclusion regarding the prOperties of spruce wood can be
drawn on the basis of the juvenile wood formation of pines.
If the wood in the vicinity of the pith is disregarded,
the wood density of spruce increases as the ring width
narrows. This is supported by the regression equations
given by T a m m i n e n (1962 and 1964) which show the
correlation between sapwood density and percentage summer-
wood and the ring width. The coefficient of determination
obtained from the equations is 0.40 for spruce, but only
0.22 for pine.

The negative correlation between spruce wood density and
ring width is largely dependent on the negative correlation
between percentage summerwood and ring width (W e g e l i -
u s 1946, N y 1 i n d e r 1953 b, K 1 e m 1957). In
contrast, the corresponding correlation is weaker for pines
(L a r s o n 1957). However, R i s i and Z s l 1 e r
(1960) found no correlation between the wood density and
ring width of Pigea_mariana. It is worth of mentioning that‘
the average ring width Of spruce pulpwood in a study of the
19308 was 1.96 mm in the area in question here (V u o -

r i s t o 1937)-

12. Density variation in the radial direction of the
stem

The objective set was to find patterns in accordance with
which wood density varies within the tree. The problem was
divided into two parts: variation in density in the radial
direction of the stem, and variation in density in the
longitudinal direction of the stem. However, as these part.
tasks are interlinked, they are inevitably discussed con-
currently from time to time.

Table 7 shows the density variation in different parts
of the stem in pine, spruce and birches. The material was
divided into three age classes. These classes are not fully

41

'comparable because the youngest contains a higher proportion
of trees that have grown on good sites than the older age
classes. Three values are given for each relative height.
The first represents the inner part of the stem which con-
tains 10 per cent of the stem volume atthe heigth in
question. The third value represents the surface of the
stem, and correspondingly, 50 per cent of the stem volume.
The middle value denotes the density of the part of the
stem between other two parts.

The density of pinewood increases from pith to cambium.
This is seen most clearly in the middle age class of the
material and, again, at the relative height of 10 per cent.
An exception may be the butt part of old stems on account
of extractives that develop in heartwood (of. J a 1 a v a
1952).

A density increase from pith to cambium is in fact
characteristic of most softwoods (G o g gla n s 1961).

It has been established for Scotch pine by W i j k a n -

d e r (1897), H e 1 a n d e r (1918), T r e n d e 1 e n —
b u r g (1939), J a 1 a v a (1946), N y l i n d e r (1953 a)
and T a m m i n e n (1962). The change is similar in seve—
ral other pine species (0 r m a n 1952, S p u r r and

H s i u n g 1954, B a n k s 1955, Y a n d'l e 1956,

K e l s e y and S t e e l e 1956, J a y n e 1958, etc.).

Table 7. Density of pins,

42

different parts of the stem.

spruce and birch wood in

 

 

 

.Tzss_989._yearsr_lr_

 

 

 

 

 

 

 

 

Relative — 40 I 41 - 8o 81 +
stem height, ' ‘“"“”
per cent Part of_the core ,_3“1T113
Inner IMiddle Outer I Inner IIhdIe I oite___ I _lrrer__IIiidIe___,oiter
Densitylkg/solid cu.m
Pine

0 I 390 402 424 427 443 470 I 481 476 478
10 I 343 371 414 363 408 459 I 405 444 465
20 I 334 353 395 360 392 436 a 390 420 443
30 I 333 343 382 354 384 424 3 378 404 429
40 I 330 344 365 350 379 406 j 374 394 413
50 I 334 345 364 354 376 397 . 370 389 399
60 E 333 341 344 357 368 382 2 374 385 384
70 I .. .. .. i 345 363 364 § 374 378 376

I Spruce I

o 386 353 363 396 370 384 2 415 401 404
10 349 344 361 369 366 387 5 386 398 407
20 341 338 352 356 362 385 I 379 391 402
30 338 338 353 354 361 377 5 371 383 397
40 342 332 349 356 360 375 372 382 389
50 349 335 347 365 368 373 372 380 382
60 352 334 344 359 359 369 374 377 377
70 .. .. .. 361 358 366 375 376 369

Birches .

0 456 487 519 I 468 498 517 3 490 516 521
10 440 471 503 I 458 488 512 488 515 527
20 441 468 495 I 454 486 504 , 472 505 518
30 445 460 490 | 452 485 499 I 474- 505 520
40 439 465 485 5 454 483 497 I 471 502 517
50 429 454 471 462 486 499 I 473 501 508
60 438 436 464 I 453 476 503 I 487 503 494 .
70 .. .. .. I 449 473 467 469 494 499

 

 

 

 

43

The change in the density of Scotch pine is sharpest in
the area of juvenile wood, but continues outside this zone
as well. The density of forming wood depends appreciably
on the percentage summerwood and on the stem age at the
height in question, but the width of the annual ring is also
significant. The average density of the cross section of
pine seems to reach its maximum when the tree age at the
stem height in question is about 80 years. .

The wood density of spruce changes a little more irre-
gularly in the radial direction of the stem. It is rela-
tively high in the immediate vicinity of pith, higher
than in Scotch pine. The density decreases first towards
the cambium, contrary to the situation with the pine, but,
a few annual rings later, begins again to increase
towards the cambium. No endeavour was made in this
connection to establish the point where the change takes
place. Owing to the sampling method employed, the above
change in density can be established in Table 7 only in
the youngest age class. In a study by N y l i n d e r
.(1953 b), the density of wood in a spruce plantation was
at its minimum five annual rings from the pith. In Plcea
sltghgnpip the corresponding point was 10-15 rings from
the pith (B r y a n and P e a r s o n 1955).

The wood density in spruce does not apparently greatly
depend on the ring distance from the pith. 0n the other
hand, percentage summerwood and ring width appear to be of
great significance. It may be assumed from the last-~ I
mentioned factor that silvicultural treatments may affect
wood quality particularly strongly in spruce wood.

The wood density of birches increases regularly from
pith to cambium. This has been established also by

s t a u f f e r (1892), w a 1 1 d e n (1934), J a 1 a —
v a (1946) and K u j a l a (1946). However, the changes

in birches are smaller than those in Scotch pine and the

effect of ring width in particular on wood density is

weak.

44

13. Density variation in the longitudinal direction
of the stem

Density variation in the longitudinal direction of the
stem is important for the value ratios of timber assort-
ments. As the density variation within the stem is usually
greater than the inter-stem density variation, it is im-
portant for the timber user to be familiar with it.:g

Density, kg/cu.m
1

500 _,

.B
K—::::::::::::\\
‘ 10 x5 20

.I

 

—

-I

400

_q

i
I IO\I5\20
h

I T I T T I .' T f T I I I I T I I I
10 r-

Distance from stump level, m

 

 

Figure 7. Density of 10, 15 and 20 m pines, spruces and
birches at different stem heights.

Density, kg/cu.m

 

 

 

 

 

20 40 III) 80

 

Relative stem height, per cent

Figure 8. The variation pattern of wood density in the
longitudinal direction of the stem.

45

Density variation in the longitudinal direction of the
stem depends on the tree size. It is greatest in a long
stem, but sharpest in a short stem. Figure 7 shows that
the average wood density of pine and birches increases
with tree length. For spruce, the situation is reversed.
In pine and birches, the difference in the density of
stems of varying sizes in influenced more by positive
correlation between density and age, in spruce by the
negative correlation between density and rate of growth.
In harvesting spruce, small stems are obtained mostly
from the slow—growing trees of the suppressed tree classes.

The density variation in the longitudinal direction in
stems of varying sizes can be illustrated by a single cur-
ve when the absolute heights are converted to relative
heights. This gives variation patterns of general applica—
bility for each tree species, and these are shown in
Figure 8 fitted with the naked eye. It is possible with
the help of these patterns to compare the different tree
species.

Density of pinewood decreases from the butt towards
the crown, at first more sharply, but within the living
crown more slowly. S p u r r and)

{H s i u n g (1954) listed a number of studies showing
that this trend prevails not only in pines but also in
many other softwoods.

The variation pattern of density in the spruce stem
differs from most of the other tree species. The density
of the butt end at first decreases on moving towards the
crown, but from the middle of the stem it begins to
increase again. A similar conclusion was drawn by
B e r t o g (1895) according to whom the wood density of
Spruce decreases from the butt to the lower limit of the
living crown but increases again within the crown to-
wards the top of the tree. This result was arrived at by
B e r n h a r t (1964) as well. Some authors are of the
Opinion that density of spruce wood in the longitudinal
direction of the stem is relatively constant (T u r n -

b u 1 1 1942, J a 1 a v a 1946). The density of

Eigea_sitghen§i§ decreases towards the crown according

46

to one report (B r y a n and P e a r s o n 1955), while
a contrary trend was noted in Eicea_glauca in another
study (H a 1 e and F e n s o m 1931).

The density of birches decreases from the butt towards
the crown. This has been reported also by S t a u f f e r
(1892), J a 1 a v a (1946) and K u j a 1 a (1946). Wood
in the butt is 30 kg/solid cu.m heavier in common birch
than in white birch. However, the density of common birch
decreases more sharply towards the crown, and the wood'
density in these two birch species is the same at the
relative height of 90 per cent.

"Springwood formation is associated with active crown
growth, and on the main trunk the bulk of the springwood
is produced during the period of terminal elongation.
Proximity of the crown assures a high production of
springwood in the stem" (L a r s o n 1962). Therefore
the springwood portion of the annual ring tapers down-
ward and the summerwood portion tapers upward (T u r n —
b u 1 1 1937).

As the living crown of the tree recedes from the butt,
the effect of auxin that develops in the crown on the
butt end of the cambium, and via it on the prOperties of
the wood that develops in the butt end of the tree,
decreases. It is partly for this reason that the density
of the pine and birch wood which develops at a certain
height increases with age, as was shown by the regression
equations presented earlier.

The develOpment is different in the spruce stem. The
living crown does not contract equally rapidly as in the
pine and birches. The crown of spruce trees in the
investigation material comprised an average of 68 per
cent of the total stem length, while the comparable
figure for pine was 43 and for birches 50 per cent.

A considerable proportion of spruce wood develops
within the crown. The wood forming in the different parts
of the stem is thus fairly even in density. This concurs
well with the strength requirement made of the stem
because, since the crown reaches far down, the wind
pressure is distributed more evenly over the different

 

 

 

 

 

47

parts of the stem than in the pines. The different density
gradient of common birch and white birch may also be
caused partly by the different self—pruning of the stem.
The self-pruned part of the stem is generally longer in
common than in white birch (H e i s k a n e n 1957).

The variation of wood density in the longitudinal
direction of the stem shows that the densities of timber
assortments prepared from different parts of the stem
differ. This question, which is of especial interest to
the pulp and paper industry, is discussed in Chapter 3.
The density variation in the longitudinal direction of the
stem must be taken into consideration also in the
collection of density samples. The following chapter deals
with its effect on the sampling method.

14. Estimating the average tree density

When the present investigation was begun it was known
that several wood density studies would be needed in the
next few years. The tasks will be such that the samples
must be collected by increment borer from standing trees.
No such method has been used so far in Finnish conditions.
However, V u o k i l a (1960) proved that the average
tree density of Larix_sibiriqa can be measured most
accurately from samples taken at the height of 4-5 m.

N y l i n d e r (1961 a and 1961 b) established that at
10 per cent height the density of pine is 7 per cent and
of spruce 2 per cent higher than the average tree density.
The difference was not constant, but depended on, for
instance, the tree age.

The development of an effective sampling method for
the research tasks in question was one of the problems
of the present work. By the time the investigation was
commenced W a h 1 g r e n and F a s s n a c h t (1959),
U.S. Forest Products Laboratory, had worked out
regression equations for southern yellow pines. These
equations showed the average tree density as a function
of the reciprocal of the density of an increment core
taken at breast height. The best of the coefficients of
determination given by these equations was for Binus_

48

taeda, 0.53, and the poorest that for Binus Elliottii,
0.25. Multiple regression analyses did not increase appre—
ciably the prOportion of explained variation. It had also
been shown in young Einus taeda and Einus Elliottii trees
that there is a very close association between the density
of juvenile wood at breast height and that of the entire
zone of juvenile wood (Z c b e l, W e b b and H e n s o n
1959).

The density of Scotch pine, Norway spruce and birches
is higher at breast height than the average tree density
(Table 8). But, despite the fact that the density at
breast height tends to give too high average tree density,
it does not alter the density ranking of the trees
(B a r e f o o t, H i t c h i n g s and E l l w o o d 1964).

Table 8. Density (kg/solid cu.m) at breast height and
average tree density.

Density at Average tree

 

Species breast density Diffe- t-value
height rence
Y i S i4: S
Pine 432-4i44-3 408.6:32.8 23.8 t1568=ll.59*
Spruce 392.4:33.8 386.8:30.0 5.6 t1500= 3.36‘
Common *
birch 509.8:37.4 496.9:3l.2 12.9 t496= 4.18
White
birch 484.5:33.5 481.9:29.2 2.6 .t6,8= 1.03
Birches, &
combined495.8i37.4 488.6:3l.0 7.2 t1116: 3.50

A sample taken at breast height, thus, permits comparison
of the average tree density for a certain tree species.
When the average tree density is determined from a
sample taken at breast height, the result will be too
high unless corrected. The correction is best performed
by using a regression equation which gives the average
tree density as the function of density measured at
breast height. In addition to simple linear regression,
multiple linear regression was also tested for different

tree species.

49

The results of the regressionemalysis are given in
Table 9. Variables anumerated on p. 25 were tested as inde—
pendent variables. The table gives the best equations for
one, two and three independent variables and for
comparison with some earlier studies equations, in which
X1 and X8 are independent variables. The following
variables appear in the table.

Y = average tree density. It was calculated as the
weighted mean of increment cores taken from the
stem. The mean was corrected for the error caused
by the mercury immersion method.

X = density at breast height. As no increment core

was taken at breast height, this value was

obtained by interpolation from the densities of

the increment cores taken both sides of the 1.3—
metre height. The value obtained was corrected as
above.
- Y

X .10: See page 25.

3

Density at breast height was thus determined by using
two increment cores. This involved interpolation in which
the increment core taken at 10 per cent height generally
had more weight. If the stem height was exactly 13 m,
the value was obtained solely from an increment core taken
at 10 per cent height. The average height of all the tree
species was roughly 13 m.

50

 

 

 

 

Table 9. Regression of the average tree density on the
on the density at breast height.1)

Species Regression equation R2 Sy.x

Pine Y = 129.0+0.648X1 .7663 15.9
= 129.0+0.682X1-1.039X8 .8053 14.5
= l70.0+0.589X1-67.8X9 _ .8138 14.2
= 155.8+0.634X1-46.7X9—0.738X1O .8224 13.9

Spruce Y = 66.4+O.817X1 .8471 11.7
= 79.4+0.799X,-0.409X .8541 11.5
= 73.6+0.818X1-0.0014 X3 .8560 11.4
= 89.9+0.779X1-15.2X9-0.429X1O .8571 11.4

Common

birch Y = l22.8+0.734X1 .7716 15.0
= 122.2+0.737X,- 0.074X8 .7718 15.0
= l62.l+0.682X1—630X7 .7905 14.3
= 168.2+0.700X1-982X7-0.00028X3 .7922 14.3

White

birch Y = 95.1+0.798X1 .8414 11.6
= 95.1+0.798X1+0.000X8 .8414 11.6
= 98.5+0.777X1+0.128X4 .8477 11.4
= 36.1+0.907X1+1.249X4-0.00232X3 .8501 11.4

Birches,

combined Y = 118.5+0.746X1 .8134 13.4
= ll7.4+0.753X,- 0.146X8 .8141 13.4
= 126.3+0.746X1-30.5X9 .8248 13.0
= l43.4+0.722X1-—23.7X9—338X7 .8299 12.8

1) Density

both sides

at breast height (X ) was obtained by interpo-
lation from the densities of the increment cores taken

of the 1.3—metre height.

The coefficients of determination obtained from the

equations are high for all the tree species, especially

for spruce and white birch in which the wood density

changes fairly little in the longitudinal direction of

the stem. For comparison, mention may be made of the

equations employed in the southern and western wood

density surveys in the U.S.A. (U.S. Forest Service 1965a

and 1965b) which give the average tree density on the

basis of the density of an increment core taken at breast

height and the breast height diameter. In the former case

the best coefficient of determination, 0.59, pertains to

Einus longifplia, and the poorest, 0.44, to Einus

51

Elliottii (M i t c h e l 1 1965). The comparable figures
in the western wood density survey were 0.70 for Tsuga_
hetergphylla and 0.44 for Larix occidentalis (W a h 1 g -
r e n 1965). At least the following factors contributed

to the higher accuracy achieved by the equations presented

in this investigation.

Table 10. Density (kg/solid cu.m) at 25 per cent height

and the average tree density.

 

 

 

 

 

 

verage tree Density at
Species density 25 per cent Diffe- t-value
height rence
X': S X': S
Pine 408.6:32.8 408.5:36.l 0.1 t1568=0.02
Spruce 386.8:30.0 386.3:32.5 0.5 t1500=0.37
Common
birch 496.9131.2 495.9133.5 1.0 t496=0.36
White
birch 48l.9:29.2 481.9:32.5 0.0 t618=0.00
Birches,
combined 488.6:31.0 488.1:33.7 0.5 t -0.24

Present investigation

- Density at breast height
determined by interpola—
tion from densities meas-

ured from two increment

 

1116-

American wood density surveys

- Density at breast height
determined from one
increment core only.

cores.

- The increment core was - Increment core treated as
divided into three parts a whole.
for the density deter-
mination

- Increment core volume - Increment core volume mea-

measured by the mercury
immersion method. Sys-
tematic error was correc-

ted.

borer.

sured on the basis of the
core length and the dia-
meter of the increment

52

E r i c s o n (1966 b) developed the corresponding
equations for Scotch pine and Norway spruce in Sweden.

The coefficient of determination was 0.59*0.93 in diffe~
rent cases and for spruce higher than for pine.

The accuracy increases with the number of cores taken
from the tree (T a r a s and W a h l g r e n 1963,

K r a h m e r and S n o d g r a s s 1965). As it is

the intention to take only one increment core at breast
height in later studies, it is probable that the average
tree density cannot be calculated as accurately as in
Table 9. T a r a s and W a h 1 g r e n (1963) were able
to reduce the standard deviation for the equation of EiEQ§
lpngifglia from 21 kg/solid cu.m to 18 kg/solid cu.m when
instead of one increment core two cores were taken at
breast height and both were divided into three parts for
the density measurement.

It was shown by N y l i n d e r (1961 a) that the
density of Scotch pine at 25 per cent height is approxi-
mately the same as the average tree density. This is valid
for spruce, common birch and white birch as well (Table 10).

It appears from Table 10 that average tree density can
be determined from a sample taken at 25 per cent height
with even greater accuracy than from a sample taken at
breast height (of. Table 8). Although the practical impor-
tance of such a sampling method is insignificant, some
regression equations are presented here for the sake of
comparison. In these equations, the 25 per cent height
density is used. It was calculated as the arithmetic mean
of the densities of cores taken at 20 per cent and 30 per
cent heights, with the systematic error of the mercury
immersion method eliminated from the mean as described in
the chapter on the method. Table 11 shows that this
procedure really gives a higher accuracy. The use of the
method, however, is restricted to materials collected
from felled trees.

The results of this study show that the average tree
density can be established from an increment core taken
at breast height. The equations of Table 9 were based on

the assumption that the density of the increment core

53

Table 11. Regression of the average tree density on the

 

 

 

 

density at 25 per cent height. 1)

Species Regression equation RC Sy.x

Pine Y = 63.4+0.845X2 .8619 12.2
= 81.1+0.813X2—20.9X9 .8658 12.0
= 81.7+0.799X2—l5.0X9+O.059X4 .8681 11.9

Spruce Y = 48.2+0.877Y .9014 9.4
= 67.2+C.839X2—22.4Xq .9052 9.2
= 67.8+0.842X2-l6.2X§-0.217X1O .9061 9.2

Common

birch Y = 77.4+0.846X2 .8250 13.1
= 102.7+0.809X2—352X7 .8306 12.9
= 153.1+0.809Xa—841X7—10.3X6 .8314 12.9

White

birch Y = 83.8+0.826X6 .8484 11.4
= 78.1+0.823X2+0.535X1O .8529 11.2
= 84.7+0.811X2+0.803X1O-l8.7X9 .8553 11.2

Birches,

combined Y = 75.9+0.946Xn .8438 12.3
= 74.9+0.833Xi+0.482x1O .8484 12.1
= 82.1+0.8l9X§+0.747X1O—17.2X9 .8506 12.0

1) Density at 25 per cent height (X2) = the arithmetic
mean of the densities of cores taken at 20 per cent and
30 per cent heights.

taken at breast height has been correctly measured. The
serviceability of the equations is thus not necessarily
tied with a sample of certain shape or size. They can be
used whether the sample is an increment core taken at
breast height or in disc form. Also, the increment core
diameter does not have to be the 5.5 mm used here. What
is most important is that the systematic error of each
method is eliminated before the equations are applied.

As the core to tree density relationship within a
species does not vary significantly with geographic
location (M i t c h e 1 1 1964), it seems that the
problems set in the study can be solved by taking the
increment core at breast height from a standing tree. The
average tree density is obtained by using the relevant
equation from Table 9.

2. DENSITY VARIATION BETWEEN THE STEM
21. The external tree characteristics as density indicators

211. Regression of tree density on tree age

For pine, common birch and white birch, age was the most
important of the variables tested for effect on the average
tree density. The age range of the trees in the samples
was 25-150 years for softwoods and 25-100 years for birhes.
The density of pine and white birch within these age
ranges was illustrated best by a regression equation in
which both age and the square of age were independent
variables. For spruce, however, the square term proved
insignificant, and it was possible to illustrate the cor-
relation between average tree density and age by a
straight line. The density of common birch was explained
best by the reciprocal of age (Table 12).

Table 12. Coefficients of determination illustrating the
correlation between tree density and age.

 

Independent variable

Species Age Age2 Age+Age2 Logarithm l/age

of age

 

 

 

 

 

Coefficient of determination

 

Pine .2409 .1707 .3185 .2975 .3157
Spruce .1506 .1443 .1506 .1411 .1161
Common birch .1984 .1677 .2218 .2243 .2388
White birch .1208 .0990 .1383 .1320 .1284

Pine reaches its maximum density, according to the
present material, only after it is more than 100 years
old, white birch a little earlier. The density of spruce
and common birch stems did not reach the maximum value
in this material (Figure 9). However, these tree species
obviously show the same deve10pment later on, for a
decrease in wood density has been reported for the
narrow rings of over-mature specimens of practically
all the major species studied (L a r s o n 1957).

55

Tree density, kg/cu.m

 

 

- CB
.500 _ W8

‘ P
400 __

1

4

1 I I I I I I I I I I I I I I 1

Tree age, years

Figure 9. Correlation between tree density and tree age.

Table 13. Regression of tree density on age.

 

 

 

Species The best regression equation R2 Sy.x
Pine Y = 333.8+1.759X4—0.00806X5 .3185 27.1
Spruce Y = 356.5+0.390X4 .1506 27.6
Common

birch Y = 538.0—2 036X7 .2388 27.3
White

birch Y = 425.3+1.565X4-0.00850X5 .1383 27.2

 

The low density of wood produced by an old tree results
from the decreasing vigour of the tree.

The equations which illustrate best the regression
between tree density and age are given in Table 13 for
each tree species. The correlation is closest for pine,
but it is significant for the other tree species as well.
The correlation between tree density and age has been 2
correspondingly greater for pine than for other tree
species in foreign investigations, too. W h e e 1 e r
and M i t c h e 1 1 (1962) found that by far the most
important single variable tested to predict core density

56

was the reciprocal of age. The coefficient of determination
was 0.17 for Pinps_19ngifolia and 0.36 for Pinps_taeda, In
the western wood density survey (U.S. Forest Service 1965 b),
the significance of age was distinctly smaller for e.g.
Pseudptsuga and Abies, In K 1 e m’s study (1965) only 2

per cent of the density variation of spruce could be
attributed to age.

However, the part of the variation in tree density which
in the present work was seen to be related to tree age is
not caused in its entirety by age (of. K n i g g e and
S c h u 1 z 1966). Other factors, such as the growth rate,
are also involved. The youngest trees in the material are
from the best sites, as trees on poor forest site types do
not in 25 years reach the minimum stem size used in this
study. In spite of this, it is obvious that tree age is of
such great significance for tree density that attention
should be paid to the following points.

- An endeavour is made in Finland to shorten the
rotation time of the growing stock. In consequence,
the average density of the wood produced by the
forests of southern Finland will decrease.

— The future tree plantations will produce usable timber
at a fairly early age. The density of the timber to
be harvested from plantation stands will be exceptio-
nally low in the first thinnings owing to the young
age of the growing stock. Such timber should be
directed to the groundwood pulp, particle board or
other industry for which low-density wood is most
suitable.

— The growing stock of northern Finland is of conside—
rably older age than that of southern Finland, but
for a great part over-mature. This may be responsible
for the differences in the pulp and paper industry’s
raw material consumption which have been found to
exist in practice.

- Attention should always be paid to tree age in

studies of wood density.

57

212. Regression of tree density on stem size and form

The correlation between tree density and stem size and
form was studied with the aid of the following variables:
diameter at breast height above the bark (X8), stem length
(x10), and breast height diameter divided by the stem
length (X11). The last-mentioned variable was considered to
illustrate the taper of the stem.

The results of regression analysis are shown in Table 14
which gives the coefficient of determination obtained from
the equations. The values in brackets were obtained from
equations which the F—test did not prove significant at the
five per cent level.

Table 14. The interdependence between tree density and
stem form and size. Coefficients of determination when the
effect of age is not taken into consideration.

 

Variable
Spe01es X8 X10 X11
Coefficient of determination

 

 

 

 

 

 

Pine (.0000) (.0043) .0093
Spruce .0886 .0734 .0278
Common birch .0446 .0523 .0237
White birch .0073. .0132 (.0022)

The interdependence between average tree density and
the variables tested is poor. The correlation is negative
for spruce but positive for the other tree species.

It has already been shown that the average tree density
increases with age. It might be eXpected from this that
tree density also increases with stem size. This was the
finding for southern yellow pines by W h e e 1 e r and
M i t c h e 1 l (1962). As different results have been ,
arrived at in the present work, there must be some factor
which affects the density of stems of different sizes in
a way contrary to the influence of age.

This factor is the growth rate of the tree. Some small—

sized trees are young and grow rapidly, while others are

58

old and of slow growth. It is due to the contrary effect
of age and the growth rate that stem size does not seem
to have a significant effect on the density of the pine
and white birch stem. The density of common birch increa—
ses slightly with size, for age has a stronger effect than
growth rate on this species. The situation is reversed
for spruce in which the density decreases with increasing
stem size.

If age is taken into consideration as an additional
factor in the regression, the influence of stem size and
form on density in even-aged trees can be examined.

Table 15 shows the coefficients of determination for
equations in which the tree density is explained not only
‘by variables illustrating size and form but also by the
tree age or its reciprocal.

Table 15 differs considerably from Table 14. Stem size
now has a significant influence on the density of soft—
woods when stem age is constant. For birches, however, the
density does not seem to depend on stem size. When age
remains constant the correlation between density and these
variables is negative in softwoods, as shown by the
examples below.

2

Species Regression equation R Sy.x
Pine Y = 475.4-2 469X7-1.55X8 .3889 26.1
Spruce Y = 377.0+0.560X4-2.34X8 .3622 24.0

Table 15. The interdependence between tree density and
stem form and size. The coefficients of determination when
the effect of age is taken into consideration.(l) For
spruce the age has been used instead of the reciprocal

 

 

 

 

 

 

 

of age).
Variable
Species X7 X7+X8 X7+X10 X7+X11
Coefficient of determination
Pine .3157 .3889 .3496 .3582
Spruce 1) .1506 .3622 .3321 .1961

Common birch .2388 (.2394) (.2388) (.2398)
White birch .1284 (.1340) (.1290) .1383

 

The results are fairly consistent with the following
values introduced by N y l i n d e r (1959) in Sweden.

Species Age, Dbh, cm
years 10 20 30
Tree density, kg/solid cu.m
Pine 49 420 405 380
Spruce 45 397 352 338

The stem form also illustrates tree density (Table 15).
The more sharply a stem of a certain age tapers, the
smaller is the density of pine, spruce and white birch.
Taper does not appear to have an effect on common birch.
It has a particularly distinct influence on density in
the Norway spruce, as has been previously reported by
K 1 e m (1934), K 1 e m, L 8 s c h b r a n d t and
B a d e (1945) and P e c h m a n n (1954). In a recent
investigation the correlation between density and stem
form was negligible, however (K 1 e m 1965).

213. Regression of tree density on crown size

The relative length of the living crown of all sample
trees was measured, but the crown width, density and
vigour were not recorded. The picture of the crown was thus
inadequate. It was considered, however, that the effect of
the crown quality on tree density can be studied by means
of the correlation between tree density and crown pr0portion.

The tree density was higher in softwoods the greater
the proportion of the stem that was self-pruned. No
significant correlation was obtained for hardwoods.
However, as the proportion of living crown in the total
stem length of pine and spruce decreases with the age of
the tree, the correlation in question may be only osten-
sible. Hence, again, age must be a constant in the
regression. The equation receives the following form for
pine and birches,

Y = ,S’o +/‘>’1X7 +/=’2 X12

60

and for spruce the age (X4) is entered instead of the

reciprocal of age (X7). The estimate, b2, of ”32 now
shows the effect of crown proportion (X12) on tree density

(Y) when X1 is expressed in per cent.

2
Species R2) b2 t-value of b2
Pine .3235 —0.246 t782 - 3.02*
'Spruce .1807 -0.347 t748 = 5.24*
Common birch .2401 -0.094 t246 = 0.68
White birch .1329 -0.164 t307 = 1.27

The effect of the crown pr0p0rtion is still significant
for softwoods, but insignificant for birches. However, the
effect seems to be fairly small, for the density of pine
decreases 2.5 and of spruce 3.5 kg/solid cu.m when the
crown proportion increases by 10 per cent.

Obviously, the influence of the crown on the density
of wood that forms in the stem is much more important than
it has been possible to prove here. The vigour and earlier
deve10pment of the crown was disregarded in the present
study. 0n the other hand, it is to be eXpected that, given
the same growth rate, a spruce with few needles which
grows on a good site produces wood of similar density to
that produced by a spruce with profuse needles growing
on a poor site (S c h m i d t 1953).

"The crown of the tree is the regulating center for
all wood formation. The external factors of climate and
environment exert their influence directly on the growth
of the crown and only indirectly on the development of
wood" (L a r s o n 1963). It also seems reasonable to
relate the amount of juvenile wood to the magnitude of
the green crown (P a u 1 1960).

Severe and continued reduction of the lower part of
the crown of Pinus lpngifglia results in greatly reduced
ring width in the lower part of the trunk, but summer-
wood is reduced less than springwood (M a r t s 1951).
Green pruning an Open—grown tree converts it to a
simulated stand—grown tree, and the artificial crown
recession restricts crown-formed wood to the new crown
region (L a r s o n 1962).

61

22. The influence of environmental factors on tree density
221. The influence of stand structure on tree density

In order to study the correlation between stand density
and wood density, the density of every stand was determi—
ned by using a scale where the density of an Open space
is denoted by 0 and the density of a fully closed stand
by 1. These figures denoting stand density were tested in
regression analyses as variables influencing wood density.
As tree age was another explaining variable, the value of
stand density thus determined was insignificant in the
analysis for all tree species.

The result cannot be regarded as surprising. The
density in the stand at the time of investigation may
often be entirely different from what it has been during
the earlier development of the stand. Nor does the classi—
fication used take into account the site quality, and
consequently it does not illustrate the biological density
of the stand.

The size and vigour of the crown of an individual tree
and the tree growth rate are dependent on the stand
density. As these factors influence wood density, it fol-
lows that stand density must also affect wood density.
However, the material in the present study was such that
it cannot provide a direct answer.

The general trend in conifers is that trees growing at
close spacings produce heavy wood and Open-grown trees
with large crowns produce wood of relatively low density
(W i j k a n d e r 1897, C i e s l a r and J a n k a
1902, P a u l 1930, K 1 e m 1944, H i l d e b r a n d t
1954, L a r s o n 1957, E r i c s o n 1962 b).

B r a a t h e (1953) calculated that the wood density of
Norway spruce on fairly dense plots was 5-6 per cent

above that on more Open plots. There is also a positive
correlation between wood density and the stand basal area.
of southern yellow pines, but the correlation it not a
strong one (W h e e l e r and M i t c h e l l 1962).
However, it has been shown that the wood density of Scotch

pine (J a 1 a v a 1934 a) and Norway spruce (S i r é n

62

1952) seed trees increases in the butt end of the stem
immediately after thinning. A thorough study of the effect
of thinning on the density and percentage summerwood in
Scotch pine and Norway spruce was recently completed in
Sweden by E r i c s o n (1966 b).

The methods for producing dense wood in second-growth
Pseudotsuga taxifolia forests are similar to those
necessary for the natural elimination of side branches
in closely stocked stands (P a u 1 1932). The regulation
of growing space throughout the life of a stand is the
silvicultural tool most readily available to the forester
in controlling wood density (P a u 1 1963).

The effect of stand structure on wood density can be
examined also by studying the differences between the
tree classes. Four tree classes were distinguished in
the present study in the manner described on p. 17, with
the tallest trees of the stand ascribed to class 1. The
correlation between wood density and tree age and tree
class can now be written

Y = [30+ /41X7+ /&X13
when X7 is the reciprocal of age and X13 represents the
tree class. Better results are again obtained for spruce
when age (X4) is entered instead of the reciprocal of age.
When the estimate of ,3 2 is denoted by b2, the following
results are obtained.

Species R2 b2 t-value of b2
Pine .3526 5.85 t783 = 6.68*
Spruce .2987 10.61 t749 = 12.57'*
Common birch .2392 0.63 t247 = 0.37

White birch .1431 2.29 t308 = 2.30 *

The wood density of trees in the lower tree classes is
higher than that of dominant trees. The differences are
esPeeially distinct for spruce, but the effect of tree class
is not significant for common birch. The coefficients of
determination cannot be improved by adding the square of
the tree class. Table 16 gives an example of the density

0f trees of different tree classes in a stand aged 50
years,

63

Table 16. Wood density of trees of different tree classes
in a 50—year-old stand.

 

Tree class

 

 

 

 

 

 

 

Species 1 2 3 4
Density, kg/cu.m
Pine 397 403 408 414
Spruce 359 369 380 391
Common birch 496 497 498 498
White birch 478 481 484 487

It has been proved that wood density increases from
dominant to suppressed trees in pine (B u r g e r 1948,
P i 1 1 o w 1954, L a r s o n 1957) and Norway spruce
(C i e s l a r and J a n k a 1902, K 1 e m 1934,

V o 1 k e r t 1941, B u r g e r 1952) stands. In
several cases, however, the intermediate tree classes
have been shown to have heavier wood than either the
dominant or suppressed trees (J a 1 a v a 1946, N y -
1 i n d e r 1953 b, H i 1 d e b r a n d t 1954).

222. The influence of site quality on tree density

As the quality of the site affects the volume growth
of trees, it may be assumed also to influence wood density.
W i 1 d e, P a u 1 and M i k 0 1 a (1951) reported
that site conditions exert a marked influence on wood
density. Most research Workers are of the opinion that
the wood density of softwoods is higher on poor than
on good sites. This applies to both Scotch pine (L a s —
s i l a 1929, S i i m e s 1938, J a 1 a v a 1946,

T a m m i n e n 1962) and Norway spruce (W i j k a n -
d e r 1897, K 1 e m 1934, W e g e 1 i u s 1946,

K 1 e m 1965). Several studies have reported that the
Percentage summerwood and density of softwoods depend
On the soil moisture supply during the growing season,
eSpecially towards the end of the summer (P a u 1 and
M a r t s 1931, C h a 1 k 1951, L a r s o-n 1957,

K r a u s and S p u r r 1961).

64

Investigation results concerning the correlation
between the wood density of diffuse—porous hardwoods and
site quality are contradictory. Populus_tremuloides and
Eppulus grandidentata_have generally a higher wood
density on the fertile than on the infertile sites, but
this relationship does not hold true in all instances
(W i 1 d e and P a u 1 1951). Wood density of
Liriodendron tulipifera is inversely related to site
index (B a r e f o o t 1963).

In studying the correlation between wood density and
site quality in the present study, use was made of the
forest site types which have such an important role in
Finland. The mineral soil forest site types covered by
the study are Qxalis-Myrtillus type (OMT), Myrtillus
type (MT) and Vaccinium type (VT), the moistest and
most fertile type given first. The fourth group consists
of undrained swamps (see p. 15).

These four different sites have been fitted into the
regression equations by assigning the forest site types
the value 1 or 0 as shown on p. 26. The independent
variables explaining tree density were the forest site
type, tree age, the reciprocal of tree age and the
interaction between age or its reciprocal and the forest
site type. When only the forest site type was considered,
the following correlations were obtained for the
different tree species:

Species R2 F-value of the equation
. . _ at
Pine .1313 F4’780 - 29.45*
Spruce .1104 F ; = 23.14
4 746 *
Common birch .1039 . F4;244 = 7.08
White birch .0191 F4;305 = 1.49

However, the question is not only of the direct effect
of site quality, for the tree age also exerts its
influence. The age of the growing stock on poor forest
site types, especially on undrained swamps, is higher on
an average than on good forest site types. Age must
therefore be taken into consideration in the regression

equations. The wood density of the different tree species

65
can the be predicted best from the following equations.

Pine and birches
Y :/30 +/41X16 +/42X17 +I£3X18 +/C4X19 +/€§X20 +/46X21 +

/37X22 +/48X23

Table 17. Density of pine, spruce and birch wood on
different forest site types.

 

 

 

 

Species Forest Age, years
site 25 I 50 I 75 I 10071125
type Density, kg/cu.m

Pine OMT 350 395 410 417 ..

MT 356 396 410 416 420
VT 383 416 426 432 435
Swamp 381 406 415 419 422

Spruce OMT 356 364 372 380 ..
MT 368 376 384 392 400
VT 376 383 391 399 407

Swamp 379 387 395 403 410

Common birch OMT 459 497 510 516 ..
MT 463 497 508 513 ..
Swamp 453 480 489 493 ..

White birch OMT 456 488 499 504 ..
MT 458 483 492 496 ..
Swamp 453 481 490 495 ..

 

 

Spruce

Y = £0 + fllxlé +/}2X17 + fl3X18 +fl4X19 +fl5X4

It is possible on the basis of the equations to
account for 37 per cent of the total tree density variation
in pine, but only 15 per cent in white birch, as shown by.
the coefficients of determination obtained from the
equations.

The equations can be used to predict the average tree
density when tree age and forest site type are known
(Table 17). However, great accuracy cannot be achieved in

66

Species R2 Sy.x
Pine .3671 26.2
Spruce .2098 26.7
Common birch .2778 27.0
White birch .1505 27.2

predicting the density of an individual tree. Young stands
on good forest site types yield considerably lighter—weight
timber than the old stands on poor forest site types. For
instance, the weight in solid cu.m in lOO—year-old VT
stands is 24 per cent higher for pine and 12 per cent
higher for spruce than in 25—year-old OMT stands.

The table warrants the following conclusions:

— When the fertility of mineral soil improves, the wood
density decreases. As far as undrained swamps are
concerned, spruce produces the heaviest wood just in
swamps, whereas the situation is the Opposite for
birch species. The density of pine wood on undrained
swamps is higher than on MT sites, but lower than on
VT sites.

- The density of pine wood is higher than that of spruce
wood, as is commonly known. An exception, however,
consists of young stands. The main reason for this is
juvenile wood formation in pinewood.

- The wood density of common birch is higher than that
of white birch except on undrained swamps where both

species have the same wood density.

As low-density wood forms on a good_site, it can be
expected that improving the fertility of forest soil by
fertilization will lead to a decrease in the density of
the wood produced by the stand. L a r s o n (1962) points
out that the anticipated response to fertilization would
be increased crown deve10pment. The increased crown
activity, in turn, would be expected to promote spring—
wood formation.

It was not possible in the present work to study the
effect of fertilization on wood density. The problem is,
nevertheless, highly t0pical as Finland’s forests will
probably be fertilized on a large scale in the near future.

67

For this reason, some investigation results which give
indications of the type of change to be anticipated in the
density of wood in consequence of fertilization are
presented here.

Application of N, P K20, and lime caused a decrease

0 .
of 8 per cent in the W808 density of young Pseudotsuga_
tagifglia (E r i c k s o n and L a m b e r t 1958).
Application of N, P205, and K20 significantly reduced the
wood density of Pinu§_taeda, and the wood density averaged
10 per cent less than in trees in the control plots
(Z 0 b e 1, G o g g a n s, M a k i and H e n‘s o n 1961).
W i 1 1 i a m s and H a m.i 1 t o n (1961) applied ferti—
lizer treatments of ammonium nitrate, superphosphate, and
these two together for Pinus_Elliottii. The addition of
fertilizers had a definite effect on wood prOperties, and
wood density was lowered on an average from 513 kg/cu.m to
479 kg/cu.m or 6.7 per cent. In E r i c s o n's (1962 a)
study nitrogen treatment caused a 4.9 and a 6.2 per cent
decrease in wood density of Scotch pine.

K e r r (1931) pointed out that fertilizing has less
influence on development of summerwood than on springwood
of southern yellow pines. In the study by K 1 e m (1964),
the amount of summerwood in Norway spruce did not
increase at all, but the quantity of springwood grew by
45 per cent as a result of fertilization. The result was
a 7 per cent decrease in wood density. It was established
by P e c h m a n n and W u t z (1960) that repeated
applications of nitrogen appeared to increase the percen-
tage springwood of Norway spruce. Wood density was
reduced correspondingly,hut the wood structure and density
were similar to those of faster-grown spruce on good sites.
E r i c s o n (1962 a) stated that with fertilization,
pine and spruce density generally decreases but the density
of "hunger wood" which has grown on a very poor soil may
even grow through the fertilization.

If forest fertilization becomes widespread in Finland,
it can be assumed that the average density of the timber
produced by the coniferous forests will decrease to some

extent and the raw material consumption of the pulp

54
’1 L1“

industry will grow correspondingly per production unit. Ho—
wever, fertilization can be used to increase wood production
without any fear that the change in the quality of the raw
material of the pulp and paper industry will be greater than
the present-day variation between fast—grown and slowgrown
wood (J e n s e n, V i r k 0 l a, H u i k a r i and

P a a r 1 a h t i 1964).

The studies performed so far on the effect of fertili-
zation on wood density have been concerned, however, with
young stands. It may therefore be premature to advance
hypotheses about the effects of fertilization of old forests
on the quality of timber, expecially saw logs. But on the
basis of Table 17 alone, it seems possible that the lowering
effect of fertilization on wood density is greatest in
young stands and it will decrease when the age of the stand
to be fertilized increases. Nor does it appear probable
that fertilization would result in a reduction in wood
density of birches.

23. Stem-to—stem density variation eXplained by external

tree characteristics and environmental factors

Several studies indicate that the stem—to—stem density
variation in a stand or locality is very high, but that
much of this variation has not been explained by environ—
mental factors (van B u i j t e n e n, Z 0 b e l and
J 0 r a n s o n 1961). The majority of recent writers
agree that it is very difficult to relate wood density to
environment (cf. Z 0 b e l and M c E l w e e 1958).

The variation of tree density has been attempted to
explain with the aid of different variables. It is easy to
observe that two of them, tree age and growth rate,
explain the variation in the tree density better than
other factors. The explanatory value of stem size, site
quality and other variables is largely based on the corre-'
lation between the variable in question and the tree age
or growth rate.

The combined effect of environmental factors can be
seen in the growth rate of the tree. A measure that

illustrates growth rate, on the other hand, is the average

69‘

ring width. In this study the quotient (X9) obtained by
dividingihe diameter in centimetres at breast height

above the bark by the age of the tree was used instead.
This is not.as accurate an indicator ofgrowth rate as ring
width, but it does make it possible to observe how the rate
of growth influences the wood density. By taking as a
second independent variable the reciprocal of age for pine
and birches, and the age for spruce, the following
regression equations are obtained.

Pine and birches Y 3 K30 + ’41X7 + ,czxg
Spruce Y = flo + ax, + pzxg

These equations take into consideration the influence
of both age and growth rate. The influence of the latter
variable on wood density can be seen from the following,
where the estimate of K32 is denoted as b2.

Species R2 b2 t-value of b2
Pine 03936 "' 9107 t782 3 10.03,k
Spruce .3429 - 154.0 t748 . 14.79‘
Common birch .2393 - 6.5 t246 = 0.41
White birch .1341 - 25.0 t307 = 1.43

Again, it can be seen that the density of softwoods is
significantly dependent upon the rate of growth. This
negative correlation is especially pronounced for Norway
spruce. For birches, on the other hand, the influence
of growth rate is not important.

In many recent investigations it has been concluded that
the relationship between growth rate and wood density is
of little or no importance (S p u r r and His 1 u n g
1954, L a r s o n 1957, R e n d l e 1958 and 1959,

2 o b e 1 and M.c E 1 w e e 1958, G o g g a n s 1961
and 1962). It is obvious, however, that this depends to a
great extent upon the tree species. Especially the density.
of Norway spruce seems to depend on the growth rate, but

a similar correlation can also be observed in Scotch pine.
This is the almost unanimous conclusion arrived at in
investigations in EurOpe concerning these tree species.

70

Table 18 shows what percentage of the variance of tree
density was explained by different variables. For pine and
birches the most important independent variable is the age
of the tree, in one way or another, but for spruce the
growth rate is even more important than age.

Table 19 shows the percentage of the total variance of
tree density explained by tree characteristics when the
number of independent variables increases from one to six.
It is not possible to improve the result for softwoods by

Table 18. Per cent of the total tree density variation
explained by tree and stand characteristics. (Values
insignificant at the five per cent level are given in
brackets).

 

 

 

 

Variable Symbol Pine Spruce Common White
birch birch
Explained variance, per cent
Age 14 24 15 20 12
Logarithm of age X6 30 14 22 13
l/age X7 32 12 24 13
Dbh x8 (0) 9 4 (1)
Dbh/age X9 25 33 3 3
Tree length X10 (0) 7 5 l
Dbh/length I1, 1 3 2 (0)
Crown prOportion X12 10 8 (O) (0)
Tree class X13 2 12 (1) (1)
Stand density X15 1 1 6 (0)
Forest site type X16-X19 13 ll 10 2
Site type x l/age 120AX23 37 .. 26 14
Site type x age x204X23 .. 2O .. ..

inclusion of the fourth tree characteristic, and for
birches the second or third variable even is insignificant.
If, however, the forest site type is included, the
proportion of the explained variation can be increased for
spruce and birches. The best equations then explain 41 per
cent of the total density variation of Scotch pine, 40 per
cent of Norway spruce, 28 per cent of common birch and

71

16 per cent of white birch. K 1 e m (1965) similarly
succeeded in explaining 43 per cent of the tree density
variation of Norway spruce with an equation that had 12
independent variables. In the western wood density survey
in the United States (U. S. Forest Service 1965 b) the
percentage of the explained variation fluctuated from

6 - 31 in different tree species.

Compared with existing forests, fast-growing timber
will be harvested in Finland in the future. The first
thinnings will take place when the growing stock is still
quite young. The rotation age of stands will also be
shorter.

Table 19. Per cent of the total tree density variation
explained by an increasing number of external tree
characteristics as independent variables. (The thickened

number refers to the last significant increase in the
explained variance for each species).

 

*X' Independgnt vagiablg_ 1 Explained
X I X X X X X variance,
4 6 7 8 9 10 11 12 per cent

>41

 

Pine
32
39
41
41

41
41

I I I I I I
I I I I I I
HNHNHN
NNI I I I
NNNNI I
NNHI II

HIIIII

Spruc
33
36
37
37
37
37

HI I I INm HNHNHI

IIIIII
NNNNII
NNNNII
NNNIII

NNNNHI
IIIIII
NHIIHI

Common birch

IIIIII

IIIIII

HNHNNN
I
I

NNNNI I
NNNI I l
NNI INI
N
4.

irch

 

NNNNHN
IIIIII
NNNIIIU‘NIIII
NNNNII
HIIINI
NNIIII

...:

\fl

72

Conclusions from the wood density of the future forests
can be drawn on the basis of the combined effect of
environmental factors or growth rate of the trees.
Attention has been paid previously to the possibility that
the anticipated extensive forest fertilization may in the
future forests result in a decrease in wood density. The
same effect may be caused by the more general use of
planting and seeding as methods of forest regeneration.

K 1 e m. (1944), 0 1 s o n, P o 1 e t i k a and
H i c o c k (1947), G i o r d a n o (1951), H i 1 d e -
b r a n d t (1954) and S i r e n (1959) have proved
that the forest-grown wood of Scotch pine and of Norway
spruce is higher in wood density than plantation-grown
material. With more planting space, wood density
decreases and the knot percentage increases. The production
of one ton of spruce sulphite pulp required 4.7 solid cubic
metres of wood from young plots with planting spaces of
1.25 x 1.25 metres, but 6.2 solid cubic metres from plots
of the same age with planting spaces of 3.5 x 3.5 metres
(I l e m 1944). When the stand gradually grew denser,
the differences began to decrease (K 1 e m. 1952).

24. Density as an inheritable property

The sample was collected from a geographically restricted
area to exclude this factor liable to cause variation. In
spite of this, the tree density varied widely. The standard
deviation of the tree density was 8 per cent for softwood
stems and 6 per cent for birches (Table 20).

Table 20. Standard deviation and range of tree density.

 

 

 

 

 

Density, kg/cu.mv

Species I .‘t S Range
Pine 409,: 32.8 311 - 521
Spruce 387 1 30.0 308 - 482
Common birch 497 1,31.2 407 - 571
white birch 482 .t 29.2 397 - 561

73

.Attempts were made to explain the variation in tree
density by using several tree and stand characteristics,
but even in the most favourable cases the unexplained
standard deviation for softwoods has been 6 per cent and
for birchesS per cent. To some extent this is due to
measuring errors, the possible omission of some
environmental factors and the fact that the tested linear
equation patterns have not always been the best possible.
In. spite of this, it is apparent that the main reason for
the unexplained variation is heritability, because wood I
density is the result of the interaction of environment
and genotype.

Heritability is the relationship of the heritable
jportion of the variability of wood density to the total
variability of wood density (2 o b e 1 1961). "When the
material is propagated by sexual means, the non-additive
effects of the individual genotype cannot be passed on to
their*progenies, so the term heritability must be employed
in the narrow sense. When vegetative propagation is used,
the effects of dominance and epistasis of individuals
are transferred unchanged. In this case the term herita-
bility may be employed in its broad sense" (T o d a 1958).

When investigating the possibilities of forest tree
breeding, the first point to clarify is the extent to
which the desirable wood properties are transmitted from
parent to progeny. For the present, there is not sufficient
information available on the heritability of wood density
in Scotch_pine, Norway spruce, common birch and white birch,
but some foreign investigations provide references and hints.

A close relationship exists between core wood density
and the density of the total tree bole (z o b e l, I e b b,
and H e n s o n 1959, T h o r b j o r n s e n 1961).
This speeds up the process of breeding, since it is possible
to predict from quite young trees, at least roughly, the .
density of the wood that will for: later. Data on one-year
and five-year old wood from.grafts of Einus_taeda_showed
significant differences in the wood density of the clones
(van B u i J t e n e n 1962). In another investigation
the inheritance of wood density was moderately strong for

74

twoeyear old seedlings of the same species (0 e c h,

S t c n e c y p h e r and Z c b e l 1962). A positive
correlation has also been found between the original tree
and the graft trees for Scotch pine and Norway spruce

(E r i c s o n 1960). Genetic variations and heritabilities,
however, will probably change as the trees grow older. This
is because the wood of the core is more sensitive to
environmental influences than adult wood (G o g g a n s
1962).

It has also been proved that there are differences in
wood density due to heritability between the geographical
races of Scotch pine (E c h o 1 s 1958) and Norway spruce
(K 1 e m 1957, D i e t r i c h s o n 1964). Still an
unsolved problem, however, is the extent to which the
variation in the wood density of these species can be
attributed to heritability.

Almost one-third of the tree density variation in
southern yellow pines may be due to genetic factors
(W h e e 1 e r and M i t c h e l 1 1962). F i e l d i n g
and B r o w n (1960) obtained broad-sense heritabilities
up to 0.7 and narrow-sense heritability 0.2 for Pinug
radiate. In another investigation broad-sense heritability
of this same species was found to be 0.74 in the Juvenile
wood produced during the first eight years of the trees
(D a d s w e 1 1, F i e l d i n g, N i c h o l l s and
B r o w n 1961). As to the conifers in general, narrow
sense heritabilities as high as 0.60 - 0.70 and broad-- ‘
sense heritabilities of up to 0.85 have been reported,
whereas the values for paplars are much lower (2 o b e l
1963).

One of the most important prOperties of wood, and one
which should be taken into consideration in forest tree
breeding, is wood density. This indeed is already being
done in many countries. "A good part of the initial work
of the U.S. Forest Products Laboratory in the field of
wood density surveys, was concerned with screening for
wood quality trees that had been selected and rated as
plustrees on the basis of form, vigour, growth rate and
the like. The purpose was to make certain that trees

75

destined for further study and use in tree improvement
programs were average or above, with respect to wood
quality" (M i t c h e 1 1 1958).

The problem of significant small-scale tests on wood in
living trees has occupied an important place in the
breeding programme of the U.S. Forest Service at the
Southern Institute of Forest Genetics. When the average
tree density, the diameter and the age are known, "the
wood quality index" can be determined (E c h o l s 1959).
In England, cores are extracted at breast height from
plustrees of Picea_§itchensis_to obtain evidence that the
timber is satisfactory as to its wood density and other
wood properties (P h inl l i p s 1963).

"The external appearance of a tree can be deceptive and
occasionally borings from promising-looking trees have
shown them to possess unsuitable timber qualities” (H o l -
1 o w a y 1965). One is not free to select for only
volume, or for density alone. There is a negative genetic
correlation between growth rate and wood density. This
means that one must evaluate the total wood production,
including both volume and wood density, for effective
breeding procedure (van B u i J t e n e n 1963). A strict
selection for wood density alone would result in a decrease
in the average diameter, while a selection for diameter
alone would result in a decrease in wood density (8 t o -
n e c y p h e r, 0 e c h and Z c b e 1 1964).

Thus, it seems apparent that wood density is a factor
worth major attention also in Finland when practising
forest tree breeding. The first step would be to screen
the plustrees and to remove from among them individual
trees that are not up to a minimum standard. Care should
also be taken in the new seed tree orchards now being
established not to include tree individuals that give a
sub-average wood density. By taking wood density into _
consideration in the forest tree breeding program one can
to some extent avoid the threat of decreasing wood density
inherent in future silvicultural methods.

76

3. DENSITY OF THE HOST IMPORTANT TIMBER ASSORTHENTS
31. Density of saw and veneer timber

It is possible to make use of certain stand characte-
ristics when seeking to acquire logs, poles or piles with
good strength prOperties. High-density logs are obtained
from slow-grown and old stands. Over-aged stands must be
avoided, however. The site quality permits certain
conclusions about the wood density of logs (Table 21).

Table 21. Density of saw and veneer logs from different
x forest site types.

 

 

 

 

 

 

 

Pine Spruce Birch
Forest site type Density, kg/cu.m

X13 T13 X18
OMT, HT 420 1 30 370 1 23 502 1 30
vr 433 1 37 382 1 22 523 1 34
Swamp 393 1 18 385 1 34 474 1 23
Average 427 1 36 373,1 26 501.1 32

Poor sites generally yield high-density logs, but the
differences between various forest site types are small.
The pine and birch logs of peat lands are an exception.
The former often come from over-mature trees and may
therefore be of exceptionally low density. For birch logs
the main issue is the composition of the tree species. The
proportion of white birch compared with common birch increa-
ses with the moisture of the site (cf.p.16). The following
figures show the difference in the wood density of common
birch and white birch logs.

Species Wood density of veneer logs,
kg/cu.m.
X71 8

Common birch 510 + 31

27

I+ I

White birch 478

77

One of the most important criteria for saw and veneer
timber pricing is the top diameter of the log. The way in
which the wood density varies with the log size is thus of
interest. Table 22 shows the density of logs of different
sizes. The table includes both butt and top logs.

Table 22. The effect of tap diameter on the density of
saw and veneer logs.

 

 

 

 

 

 

 

 

 

Pine Spruce Birch
Tap diameter, inches Density, kg/cu.m
I13 I18 T13
5. 5 1/2 413 1 33 375 1 22 ..
6, 6 1/2 424 1 32 379 1 28 472 1 25
7, 7 1/2 427 1 33 372 1 25 498 1 33
8, 8 1/2 424 1 32 372 1 23 502 1 3o
9. 9 1/2 426 1 29 367,1 25 505 1,31
10 + 457 1 31 369,1 23 515 1 30

The density of birch logs increases distinctly with the
log diameter. This is obviously true of pine as well, but
for spruce the differences are fairly small. In birch logs,
the proportion of common birch is highest in the thickest
logs.

Especially in pine the part of the stem from which the
logs derive affects the density of the logs. This can be
seen from the following.

Position of the log Pine Spruce Birch
Density, kg/cu.m

1 1,3 1'1,S X71 S

Butt logs 438 1,34 376,1 26 502 1 33

Other logs 402 1 22 368,1 23 501,1 22

The wood densities of logs of various grades were also
compared for pine. Although grading is based solely on
external log characteristics (of. H e i s k a n e n and
S i i m e s 1960), it is also an excellent guide to the
density of the logs. Wood density decreases as the grade

78

falls, being 10 per cent higher in the first than in the
third grade. The wood density of first grade pine saw logs
obtained from a VT site is as high as 454 kg/solid cu.m.

Log grade Density of pine saw logs,
kg/cu.m
2‘1,S
1 ' 450 1 34
2 433 1 32
3 408 1 28

Waste which originates when lumber is produced is
converted into chips for the pulp industry. These chips
are obtained chiefly from the outside portions of the logs,
where the wood density is highest. The wood density of
chips obtained from.pine and spruce slabs was determined as
described in Chapter II. The method is not entirely
suitable theoretically, but the following estimates give
an idea of the wood density of sawmill chips from slabs.

Species Density of sawmill chips
from slabs, kg/cu.m

Pine 441

Spruce 379

The wood density of sawmill chips obtained from pine
slabs is higher than given by any other timber assortment
in Finland utilized as raw material for pine sulphate
pulp. In addition, ”the chips coming for the most part
from slabs or the outside of the logs contain fibers
appreciably longer than the average of those in a whole
log" (G r e e n 1963).

On the other hand, the point that the wood density of
pine chips from slabs is 15 per cent higher than that of
spruce chips must also be borne in mind. This
notwithstanding, they are both sometimes used in a
mixture for the manufacture of sulphate pulp without
regard for the effect of this difference on the pricing
of the raw material or the production process.

79

32. Density of cordwood

The wood density of normal pine pulpwood is 7 per cent
higher than the wood density of pulpwood made from tops.
In addition, timber from tOps is known to contain a greater
prOportion of knot wood. According to N y 1 i n d e r
(1959), the percentage of knot wood is greatest at 80-90
per cent stem heights. Knot wood interferes in many ways
with the preparation of chemical and groundwood pulp
(cf. W e g e l i u s 1939). Moreover, when the smaller
solid content of a stack of timber from tOps is taken into
consideration, it is obvious that tOp timber is distinctly
poorer than normal pulpwood as raw material for pine pulp.
The differences in the wood density of spruce and birch
pulpwood are similar but smaller.

The minimum tOp diameter of pulpwood was in this
investigation 8 on under the bark. This was the practice
when the study was started. Since then the minimum top
diameter of pine and spruce pulpwood has been lowered to
7 cm, however.

In addition pulpwood of 5-8 cm diameter, so-called small--
diameter pulpwood, was made of pine and spruce. Two types
can also be distinguished in this small-diameter pulpwood:
pulpwood made from small—sized trees or from the tops of
taller trees. In pine, both are of lower density than the
corresponding normal-sized pulpwood, the difference being
2 per cent. The same applies to spruce as regards top
timber, but the difference is only 1 per cent. The density
of pulpwood made from small-sized spruces, on the other
hand, is 4 per cent higher than that of normal spruce
pulpwood, because this kind of timber primarily consists
of slow-grown undergrowth spruces. _

However, in practice the above-mentioned types of small-
diameter pulpwood are not kept separate. The relative
prOportion of both timber assortments varies with the
stand from which the timber is harvested. It is even
common that pine and spruce are not kept separate. In
addition, it appears to be characteristic that the wood
density of small-diameter pulpwood varies sharply between
stands.

80

Table 23. Density of pulpwood obtained from different
forest site types.

 

 

 

 

 

Pine Spruce (Eirch
Forest site type Density, kg/cu.m

Normal pulpwood f)

OMT, MT eeeeeeeeeeeeee 403 378 495
VT 00.00.000.00......0 427 394 520
Swamp 0.00000000000000 414 393 483
Average 0000000000000. 417 382 495

Pulpwood from teps

OMT, MT 0000000000000. 386 377 489
VT 00.000000000000000. 388 381 499
Swamp 0000000000000... 360 382 451
Average 000.000.0000.. 386 378 486

Small-diameter pulpwood 2)

OMT, MT eeeeeeeeeeeeee 382 389 ‘0
VT 00000000000000.0000 426 403 00
Swamp 0000000000000... 440 426 00
Average 0000000000000. 410 394 00

 

Small—diameter pulpwood from tops

 

OMT, MT .............. 376 370 ..
VT ................... 380 389 ..
Swamp ................ 380 387 ..
Average .............. 378 374 ..

1) Does not include pulpwood made from the tops of
log trees.

2) Does not include small-diameter pulpwood made from the
tops of saw log and pulpwood trees.

For these reasons, no endeavour was made to give a ,
figure of general acceptability concerning the density of
small-diameter pulpwood. But it is possible from the data
presented to calculate the wood density in different cases
when the proportion of pine and spruce, on the one hand,
and that of small-sized stems and tOps on the other hand
are known. '

 

81

It is possible to some degree to draw conclusions about
the density of pulpwood obtained from a stand on the basis
of the external stand characteristics. Table 23 shows the
effect of the forest site type on the density of pulpwood.

Poor sites yield dense pulpwood. 0f especially high
density is small-diameter pulpwood obtained from the small
sized stems of undrained swamps. Exceptionally low-density
pulpwood, again, is obtained from the first thinnings of
young, fast-growing forests.

Birch wood of poor quality is still utilized as
fuelwood on a large scale in Finland. The average wood
density of birch fuelwood is 492 kg/solid cu.m, according
to the present study. However, if both birch pulpwood and
birch fuelwood are prepared concurrently, the latter comes
chiefly from the tops of the trees. The wood density of
such fuelwood is 480 kg/solid cu.m.

It is possible that at least a part of the top now left
in the forest as logging residues will in the future be
used for the production of pulp. Because of this, the
density of this kind of wood was also determined. The figu-
res below show that the wood density of the tops left
behind in the forest as logging residues is low for pine
and birch but high for spruce compared with the means for
the tree species in question. A point worthy of consider-
ation is that the density of spruce logging residues is
higher than that of pine logging residues.

Species Density of logging
residues from tops,
kg/cu.m
Pine ....00000.00.00.0. 368
Spruce 0000000000000... 381
BirCh 0.000.000.0.00000 470

In Table 24 are assembled the average densities of
different timber assortments. It is prOposed that these
figures be used in southern Finland until the effect of
geographical variation can be assessed. It should be
emphasized, furthermore, that the figures denote the
density of clear, defect-free wood.

82

Table 24. Density of different timber assortments.
(l) and 2): see Table 23)

 

 

 

 

Timber assortment Pine [ Spruce I Eirch
Density, kg/cu.m

Saw and veneer logs 427 373 501
Normal pulpwood 1) 417 382 495
Pulpwood from tOps 386 378 486
Small-diameter pulpwood 2) 410 394 ..
Small-diameter pulpwood
from teps ' 378 374 ..
Normal fuelwood .. .. 492
Fuelwood from taps .. .. 480
Saw mill chips from slabs 441 379 ..
Logging residues from tOps 368 381 470

The dry weight of timber is also affected by defects in
wood. The most important is knottiness, for knot wood has
an above-average density. The density of knot wood may in
Norway spruce be three times that of normal stem wood
(W e g e 1 i u s 1939). B o u t e l J e (1966) reported
that the density of knot wood of Scotch pine is on the
average 2 times and that of Norway spruce 2.5 times as
high as the density of clear stem wood. The presence of
reaction wood in timber has the same effect, whereas
decay has a contrary effect. The combined effect of
defects is such that the true dry weight of timber is
probably slightly higher than the density of clear wood
indicates. H a r d y and W e i l a n d (1964) noted
that knots increased the dry weight of pulpwood made of
spruce species of Maine by 1.73 per cent.

The density of normal pulpwood in southern Finland is
417 kg/solid cu.m for pine, 385 for spruce and 495 for
birch according to the present study. J a l a v a (1959),
calculated the true dry weight of pulpwood to be 430, 390
and 500 kg/solid cu.m, respectively. As the latter numeri-
cal series includes the effect of knots, it can be
concluded that the figures reported earlier by Jalava for
pulpwood and those obtained here by another method are in

83

accordance. By way of comparison, mention may be made of
the studymadeby Braathe andOkstad (1964)
in Norway in which the average wood density of birch
pulpwood is about 502 kg/solid cu.m.

IV. The reliability of the results

The statistical procedure used in the investigation is
based primarily on regression analysis. The sample data are
then supposed to meet certain requirements. First of all,
it is assumed that the values of the independent variables
are measured with essentially no error.

Tree density, kg/cu.m

d

—J

500 _

 

 

Age, years

Figure 10. Tree density of pine against age.

The measurements were done as carefully as possible.
The magnitude of the error involved in the wood density
measurements and the correction of the error were
established in Chapter II.

Second, the sample must be drawn from a pepulation for-
which the variance is homogeneous. This means that.the
variance of the wood density values about the regression
surface must be the same at all points of a certain
combination of the independent variables. Graphic
presentation can be used to show that the sample meets

84

85

this requirement. Figure 10 serves as a pattern of that.

The third requirement id that the deviation of the wood
density values from the regression surface must not be
dependent on each other. The independence of errors can
usually be assumed if the Y values are randomly selected,
as is the case in this investigation. Purposive selection
of 1 values on the other hand is usually permissible.

Fitting a simple or multiple linear regression does not
necessarily assume the normality of the distribution of
the wood density values. But because t- and F-tests were
used in the tests of significance, the Y values should be
normally distributed. The distributions of the average
tree density values of pine, spruce, common birch and white
birch are shown in Figures 11-14.

Two main types of departure from the normal may occur.
In one the distribution of the data is asymmetrical or
skewed, the mean and median being different. The other,
‘kurtosis, occurs in symmetrical sets, characterized by

Observations, per cent

201

10—

 

i
_ -----_J]_--

 

 

 

 

 

 

*7’/’
l
1 300 400 7 7 509 , 600

Tree density, kg7cu.m

 

Figure 11. Distribution of tree density of pine.

either an excess or a deficit of observations concentrated
near the center of the range. When g1 is the measure of
skewness and g2 the measure of kurtosis (S n e d e c o r.
1957), the following test results are obtained.

If g2 is zero, there is no departure from normality so
far as this measure is concerned. In this investigation the
values of g2 seem to be positive for pine, spruce, and
birches. ”A positive value of g2 indicates an excess of

86

0b servations, per cent
20

d .

wt

11) ‘1

 

 

 

 

 

 

 

W; 3 1

l
300 400 500 600

Tree density, kg/cu.m

 

 

Figure 12. Distribution of tree density of spruce.

Observations, per cent
20 F'-

 

10 __

 

L

 

 

 

 

 

 

"'7/'/£;7 I "'J--I--[--

I
300 400

Tree deEEity, kg/cu.m

 

-

l
600

§

Figure 13. Distribution of tree density of common birch.

Observations, per cent

20 _‘

10 __

._ .. .... .... ........._ ....Z]

 

 

 

 

 

 

 

2,3 'n—l—F L

1* I I I ‘1

 

400 spa , 600

300
Tree density, kg7cu.m

Figure 14. Distribution of tree density of white birch.

87

Species t—value of g1 t-value of g2
Pine 000.000.000.000 + 2049* +1.060
Spruce ............. + 4.61* + 0.40
Common birch ....... - 0.79 + 0.56
White birCh eeeeeeee - 1045 + 0e32

items near the mean and far from it with a corresponding
depletion of the flanks of the distribution. This is the
manner in which the distribution of t departs from normal"
(S n e d e c o r 1957). The values of g2 are insignifi-
cant for every species, however.

If g1 is zero, the distribution of the sample is
symmetric. A positive value of g1 indicates an excess in
the number of observations smaller than the mean, whereas
a negative value indicates an opposite situation. For
birches, which have a relatively high wood density, the
value of g1 is negative but insignificant. For pine and
spruce, which are lower in wood density, g1 is positive
and significantly different from zero (of. K o l l m a n n
1951). '

As to the birches, no significant departure from
normality can be demonstrated. The excess of low tree
density values of pine and spruce, on the other hand,
indicates that the population from which they have been
drawn is asymmetric. I

As pine and spruce material is not perfectly normal,
the results of the t-test and the F-test must be treated
with a certain reserve. The tests were performed at the
5 per cent level of significance. If the difference was
then significant, this was denoted by an asterisk after
the test value.

It is advisable for the practical applications of the
results to know what the sample represents and how it
does so. As already mentioned, the study was confined to,
a limited geographical area. Collection of material was
possible in this region only in State forests for the
most part, although some private forests were also
available for the purpose. These forests must be
regarded as better than the average for the area.

88

The forest areas made available for the study also
included many stands in which it was not possible for one
reason or another to fell sample trees. Another limiting
factor was that artifically regenerated forests and
drained swamps were omitted from the material. An endeavour
was then made to collect a random sample from the remaining
stands of the study area, but in the final phase of the
field work the material was supplemented nevertheless in
some extreme cases such as OMT pines and VT birches.

For wood density, i.e. the Y values of the regression
equations, the material was selected at random as presuppo-
sed in regression analysis. On the other hand, purposive
selection was applied to certain stand characteristics
employed as X values. Although selection of X values is
permissible in regression analysis, it influences certain
applications of the results, such as calculation of the
average wood density of timber assortments.

The values reported for the wood density of timber
assortments are means weighted by bolt volume. These
means were further adjusted separately for each tree
species with the aid of the areas and the yield values
of different forest site types in order to provide better
conformity with average conditions in southern Finland.

The values thus corrected can be considered to represent
the average wood density in present-day conditions of
timber assortments harvested from the study area. These
figures may change if silvicultural treatments are intensi-
fied.

The pilot survey of the study material revealed that
the geographical variation did not seem to affect wood
density within the sample. Hence latitude, longitude and
altitude were ignored in selecting independent variables
of the regression equations.

The geographical variation of wood density is, however,
a matter of importance. mayr put forward the optimum
theory according to which each tree species forms the
highest density wood within the optimum of its range
(G a y e r and F a b r i c i u s 1921). J a l a v a
@946) holds this optimum range for Scotch pine to be

89

central Finland, and for Norway spruce south of Finland.
According to T r e n d e l e n b u r g (1939) the

Optimum range of Scotch pine reaches from southern

Finland to Germany, as to the wood density. The differences
in the pulp yield per unit volume of timber between the
pine sulphate pulp mills of southern and northern Finland,
and the differences in pulp properties, also indicate
differences in wood density of the raw material (of.

K a l 1 a 1966). In a recent study the difference in
wood density of pine and spruce pulpwood between southern
and northern Finland was not significant, however (U i t -
t o t e h o - - - 1966).

The values reported for the density of timber
assortments are most reliable in their application to the
study area. It is suggested however, that these figures
be used for the whole of southern Finland until the
geographical variation in wood density has been clarified
in later studies.

V. Practical viewpoints

It is necessary in the forest industries and in the
management of forests to pay increasing attention not only
to the volume but also to the quality of wood. As wood
quality is a concept bound with the purpose of use of the
wood, it is not possible to speak in general terms of good
wood or poor wood. In lieu of quality, wood quality
indicators which anable indirect determination of the
suitability of wood for various purposes must be used.
One of the most important wood quality indicators is wood
density.

High-density wood is advantageous for many uses, but in
some cases low-density wood gives a better end product.
For the wood user to obtain the most suitable timber for
his purposes and to be able to price if fairly, it is
important to know where to obtain timber of certain wood
density. This information is also needed in industrial
planning, sorting of raw material, and quality control of
the end product. 0n the other hand, the forest grower
should know which factors affect the density of the wood
produced by the forest in order to be able to influence
the situation when he so desires.

It is obvious that wood density is greatly dependent
on the inheritable prOperties of tree. The significance
of heritability lies in the possibility that forest tree
breeding may influence the wood density of the future
forests. It has appeared from some studies that
hereditarily fast-growing trees tend to produce wood
that is lighter than the average. Hence, if wood density
is disregarded, when choosing plustrees, the density of
the wood produced by the future forests may be lower
than it is at present. In selecting plustrees, their
growth rate should be determined as dry matter yield,
rather than as volume yield.

Another factor affecting wood density is the age of

the tree. The wood density of Scotch pine, Norway spruce
and hirches increases with tree age. Since the rotation

time of Finnish forests is to be reduced, the average

90

91

wood density will decrease at least in southern Finland.
Exceptionally low-density timber is obtained from young
stands in connection with the first thinnings.

The third important factor affecting wood density is
the growth rate of the tree. Density decreases in both
Scotch pine and Norway spruce with acceleration of the
growth rate, whereas the role of this factor appears to
be insignificant in birches. It is to be expected that the
wood.of the fastgrowing plantation forests of the future
will be of expecially low density. The fertilization of
forests has a similar influence. It must be noted, howe-
ver, that when the age of the stand increases the effect
of the growth rate on wood density decreases.

It may be assumed that wood density varies with both
race and climatic factors in different parts of Finland.
This is suggested by the differences known to exist in
the raw material consumption and the properties of the
end product of the pulp industries in southern and
northern Finland. To establish these differences, a
study of the variation in wood density in different parts
of Finland must be undertaken, for instance with the
method evolved in the present work.

It has also been shown in the present study that there
are considerable differences in wood density between tim-
ber assortments. It is thus possible for industry in some
cases to take consideration wood density in its pricing
procedure without incurring special costs. It is also
possible in some degree to screen raw material, for
instance by keeping the low-density pine pulpwood from
the tops of saw timber trees separate in the timber yard.
Most of this pulpwood is gathered into separate stacks in
the harvesting stage already (of. M u r t o 1949).

Let us take as an example a mill producing kraft paper
from pine. Its pulp yield per solid cubic metre of wood
is directly prOportional to the density of pulpwood. The'
quality of the product, too, improves with the increase in
wood density. Hence, it is possible to place the various
timber assortments utilized by the mill in an order of
avdantage on the basis of the dry matter content per solid

92

cubic metre of wood. If the value 100 is assigned to the
density of normal pine pulpwood, the following values will
be obtained for other assortments. For the sake of compari-
son, the corresponding values for spruce are also presented.

Timber assortment Pine Spruce
Ratio
Saw mill chips from slabs ............ 106 91
Saw logs (butt, middle and top logs) 102 89
Normal pulpwood ...................... 199 91
Small-diameter pulpwood .............. 98 95
Top saw logs ......................... 96 88
Pulpwood from teps ................... 93 90
_Small-diameter pulpwood from taps .... 91 90
Logging residues from tops ........... 88 91

A list of this type naturally does not give a final of
the real value ratios within a tree species. It is
necessary to consider also the variation of knottiness,
heartwood content and other factors in the different
timber assortments. Wood density, however, constitutes
the basis of the calculations.

Wood density varies in many cases more than the solid
cubic content of a stack. In practice, however, it is the
latter that is considered and wood density that is
disregarded.

VI. Summary

Knowledge of wood density may be of great significance
in determining the quality of timber used for a certain
purpose. The continuous enquiries addressed to the Forest
Research Institute, especially by industry, have shown
that data on wood density are needed for many purposes in
practice. Therefore, in spring 1962, the Forest Technolygy
Department of the Forest Research Institute started a
study to ascertain the influence of different factors on
the variation of wood density in Scotch pine (Pinus_
sflygggLLNwmywmw(hgg@giflJmeJ,
common birch (Betula_gerrucgsa_Ehrh.) and white birch
(Betu1a_pubescens Ehrh.).

The object was to find out the wood density variation
within the stem and between stems, and the wood density
of the most important timber assortments. The investigation
was confined to a limited geographical area in southern
Finland. In addition, a method was to be evolved for a
wood density inventory using standing trees.

The sample consisted of 2 095 felled trees from which
a total of 20 220 increment cores were taken. The wood
density of the cores was determined by the mercury
immersion method (E r i c s o n 1959) as basic density.
The unit of measurement employed for density was
kilogramme per solid cubic metre. The statistical
procedure was based chiefly on regression analysis.

Some of the most important results of the study are
reported in the following.

The density of pine increases from pith to cambium
especially sharply in the region of Juvenile wood. The
same tendency continues outside the juvenile wood zone.
The density of the wood that forms depends above all on
the percentage summerwood (Figure 4) and the tree age at
the respective stem height (Figure 5). Ring width
(Figure 6) is of significance, too, even outside the
Juvenile wood zone.

The density of spruce wood decreases within the first
annual rings towards the cambium, but begins to increase

93

94

again in a few years. The percentage summerwood and the
ring width appear to be the most important factors.

The density of birches also increases from pith to
cambium. The importance of the age of the annual ring is
also great, but the ring width does not influence wood
density of birches.

Density decreases in the longitudinal direction of
the stem from butt to t0p especially sharply in Scotch
pine (Figures 7 and 8). The change is similar in common
birch and white birch, though smaller in the latter. The
density of Norway spruce decreases from the butt to the
middle of the stem, but then begins to increase again
towards the top. Density variations in the longitudinal
direction of the stem are decisive for the density of the
different timber assortments obtained from the stem.

The density variation between stems has been explained
with the help of various independent variables and their
combinations (p. 25). The best multiple linear regression
equations explain 41 per cent of the tree density for
Scotch pine, 40 per cent for Norway spruce, 28 per cent
for common birch and 16 per cent for white birch (p. 62).
Regarding the individual independent variables, the tree
density variation in the pine and the birches is explained
best by the reciprocal of age (Tables 12 and 13), in the
spruce by the quotient of breast height diameter and age,
which illustrates the rate of growth (p. 61). The site
quality, i.e. the forest site type, also affects wood
density (Table 17). It has been concluded on the basis of
the magnitude of the unexplained density variation that
hereditary factors have a great influence on wood density
and its variation.

It can be assumed on the basis of the results that the
density of the softwood timber produced by the future
forests of Finland will decrease a little under the
influence of the intensification of silviculture. To
offset this, forest tree improvement activity should take
into account, in addition to volume growth, also wood
density and other wood properties.

By way of application of the investigation results,

95

the density values of different timber assortments have
been calculated. The differences between the timber
assortments are considerable, especially in Scotch pine.
The weight of a solid cubic metre of pine wood is highest
in saw mill chips from slabs, then in saw logs, normal
pulpwood, small-diameter pulpwood, pulpwood from t0ps,
small-diameter pulpwood from tops and lowest in logging
residues from tOps. The differences between the timber
assortments are smaller in spruce and birch wood.
Comparison of corresponding assortments shows that
Spruce is of lower density than pine except for the
logging residues from tops.

It has been noted in many connections that an inven-
tory should be made, as soon as possible, of the wood
density variation in different parts of Finland. This
can be done from increment cores collected from standing
trees on the basis of the method evolved in the present
study (Table 9). Samples taken at breast height can be
used to predict the average tree density with sufficient
accuracy so that the method can be applied also to deter—
mining the wood density of plustrees.

A l d r

Literature cited

1 d g e, F. and H u d s o n, R.H. 1955. Growing
quality softwoods. Quarterly Journal of Forestry
49: 109-114.

1958. Growing quality softwoods: Variation in
taken from a commercial consignment. Quarterly
Journal of Forestry 52: 107-114.

y m o u s. 1957. How should we grow conifers?
Forestry conference at Darlington Hall. Quarterly
Journal of Forestry 51: 322-331.

s, C.H. 1955. The mechanical prOperties of the
mature wood of Pinus longifolia, Roxb. grown in
the Union of South Africa. The Journal of the
South African Forestry Association 26: 18-31.

a, C.H. and S c h w e g m a n n, L.M. 1957. The
physical prOperties of fast- and slow-grown Pinus
patu1a and P£_taeda from South African sources.
The Journal of the South African Forestry
Association 30: 44-59. .

f o o t, A.C. 1963. Selected wood characteristics
of young yellow—poplar (Liriodendron tulipifera
L.). Part I. Specific gravity and toughness.
Forest Products Journal 13: 233-239.

f o o t, A.C., H i t c h‘i n g s, R.G. and

E l l w o o d, E.L. 1964. Wood characteristics
and kraft paper prOperties of four selected
loblolly pines. I. Effect of fiber morphology
under identical cooking conditions. Tappi 47: 343-
356.

h a r t, A 1 b r e c ht. 1964. fiber die Rohdichte
von Fichtenholz. 0n the specific gravity of the
wood of Norway spruce (Picea_abies_ (1..) Karet.)..
Holz als Roh— und Werkstoff 22: 215-228.

e n e v, A.P. and F o k i n a, A.G. 1958. Opyt
ispol’zovanija radioaktivnyh izotOpov dlja
issledovanija drevesiny. An eXperiment on the use

96

van

97

of radioactive isotOpes for investigations on
wood. Derev. Prom. 7(8). (Cited from Forestry
Abstracts 20: 2282. 1959).

o g, H. 1895. Untersuchungen fiber den Wuchs und
das Holz der Weisstanne und Fichte. Forstlich-
naturwissenschaftliche Zeitschrift 4: 97-112,
177-216.

e y, L o w e l l. 1962. Quality control could
start in the woods. Tappi 45: 20 A - 32 A.

e 1 j e, J u l i u s B. 1966. On the anatomical
structure, moisture content, density, shrinkage,
and resin content of the wood in and around
knots in Swedish pine (zines silvegtgig L.), and
in Swedish spruce (Bicea ables Karat.). Svensk
Papperstidning 69: 1-10. .

t h e, P e d e r. 1953. Undersflkelser over
utviklingen av glissen gjenvekst av gran.
Investigations concerning the development of
Norway spruce regeneration which is irregularly
spaced and of varying density. Meddelelser fra
Det norske Skogforsoksvesen. Bind 12: 209-301.

t h e, P e d e r and 0 k s t a d, T o r b j 8 r n.
1964. Omsetning av trevirke basert pa veiing og
torrstoffbestemmelser. Trade of pulpwood based on
weighing and dry-matter samples. Meddelelser fra
Det norske Skogforsoksvesen. Bind 20: 1-64.

n, H.P., P a n s h i n, A.J. and F o r s a i t h,
0.0. 1949. Textbook of wood technology. Volume I.
New York.

n, J. and P e a r s o n, F.G.0. 1955. The quality
of Sitka spruce grown in Great Britain. Empire
Forestry Review 34: 144-159.

1 j t e n e n, J.P. 1962. Heritability estimates
of wood density in loblolly pines. Tappi 45: 602-
605.

1963. Heritability of wood properties and their
relation to growth rate in Pinus taeda. World
consultation on forest genetics and tree impro-
vement. FAO/FORGEN 63—7/2. Stockholm.

van B u

van B u

B u r g

N

C e c h,

98

i j t e n e n, J.P. 1964. Anatomical factors
influencing wood specific gravity of Slash pines
and the implications for the development of a
high-quality pulpwood. Tappi 47: 401-404.

1 j t e n.e n, J.Pu, Z 0 b e l, B.J. and J o -
r a n s o n, P.N. 1961. Variation of some wood
and pulp prOperties in an evenaged loblolly pine
stand. Tappi 44: 141-144.

e r, H. 1948. Holz, Blattmenge und Zuwachs. IX.
Mitteilung. Die Fahre. Mitteilungen der
Schweizerischen Anstalt fur das Forstliche
Versuchswesen 25: 435-493.
1952. Holz, Blattmenge und Zuwachs. XII. Mitteil-
ung. Fichten im Plenterwald. Mitteilungen der
Schweizerischen Anstalt ffir das Forstliche
Versuchswesen 28: 109-156.

F.C., S t o n e c y p h e r, R o y and

Z c b e l, B r u c e. 1962. Early results from
the co-operative loblolly pine heritability study.
Proceedings of a forest genetics workshOp, pp.
64—68. Macon, Georgia.
k, L. 1951. Water and the growth of wood of
Douglas-fir. Quarterly Journal of Forestry 45:
237-242.

1 a r, A. and J a n k a, G. 1902. Studien fiber
die Qualitat rasch erwachsenen Fichtenholzes.
Centralblatt fur das gesammte Forstwesen 28: 337-
403.

r e 1 l, R o b e r t A. 1944. Shrinkage and
density of ponderosa pine wood. Journal of
Forestry 42: 288-290.

e r, G.A. 1960. Specific gravity of red pine as
related to stem and crownformed wood. Iowa State
Coll. J. Sci. 34: 693-707. (Cited from Tappi Mono-
graph Series - No. 24. 1962). '
a n, C.E. 1938. Relation of growth characteristics
of southern pine to its use in pulping. Journal
of Forestry 36: 576-581.
w e l l, H.E., F i e l d i n g, J.M., N i c h o 1 l s,

D a.d s

E c h o

E d 1 i

E r i c

E r i c

99

J.W.P. and B r o w n, A.G. 1961. Tree-to-tree
variations and the gross heritability of wood
characteristics of Pinus radiate. Tappi 44: 174-
179.

w e l l, H.E. and W a r d r o p, A.B. 1959.
Growing trees with wood prOperties desirable for
paper manufacture. Reprint, Journal of Australian
Pulp and Paper Industry Technical Association.

w e l 1, H.E. W a t s o n, A.J. and N i c h o 1 s,
J.W.P. 1959. What are the wood properties requi-
red by the paper industry in the trees of the
future? Tappi 42: 521-526.

r i c h s o n, J o n. 1964. Proveniensproblemet
belyst ved studier av vekstrytme og klima. The
provenance problem illustrated by studies of
growth—rhythm and climate. Meddelelser fra Det
norske Skogforsoksvesen. Bind 19: 499—656.

1 s, R o b e r t M. 1958. Variation in tracheid
length and wood density in geographic races of
Scotch pine. Yale University, School of Forestry,
Bulletin No. 64. ‘

M. 1959. Estimation of pulp yield and quality of
living trees from paired-core samples. Tappi 42:
875-877.

n, H.L. 1965. The summerwood: springwood ratio in
conifers. The Journal of the Society of Foresters
of Great Britain 38: 91-112.

k s o n, H a r v e y D. and L a m b e r t,

G r e g o r y M. 1958. Effect of fertilization
and thinning on chemical composition, growth, and
specific gravity of young Douglas-fir, Forest
Science 4: 307-315.

a o n, B 8 r j e. 1959. A mercury immersion method
for determining the wood density of increment core
sections. Statens Skogsforskningsinstitut. Avdel-
ningen fdr skogsproduktion. Rapporter nr 1.

1960. Studies of the genetical wood density
variation in Scots pine and Norway spruce. Studier
aver den arftliga volymviktsvariationen hos tall

E r i c

F i e 1

Finnish

Finnish

100

och gran. Statens-Skogsforskningsinstitut.
Avdelningen f8r skogsproduktion. Rapporter nr 4.
s o n, B 8 r j e. 1962 a. M8jligheterna att 8ka
skogsvaxten genom markf8rbattrande Atgarder.
Svenska Skogsvardsf8reningens Tidskrift nr 2:
195-202.

1962 b. Gallringens inverkan pa vedens h8stvedhalt
och torr-ravolymvikt. En orienterande unders8kning
av tre gallringsf8rs8k med gran i s8dra Sverige.
Skogsh8gskolan. Institutionen f8r skogsproduktion.
1966 a. Determination of basic density in small
wood samples. An examination of some sources of
error. Skogsh8gskolan. Institutionen f8r skogs-
produktion. Rapporter och Uppsatser nr 9.

1966 b. Gallringens inverkan pa vedens torr--
rdvolymvikt, h8stvedha1t och karnvedhalt hos

tall och gran. Effect of thinning on the basic
density and content of latewood and heartwood in
Scots pine and Norway spruce. Skogsh8gskolan.
Institutionen f8r skogsproduktion. Rapporter och
Uppsatser nr 10.

d i n g, J.M. and B_r o w n, A.G. 1960. Variation
in the density of the wood of Monterey pine from
tree to tree. Commonwealth of Australia, Forestry
and Timber Bureau. Leaflet No. 77: 1-28.
Meteorological Office. 1964. Ilmatieteellinen
keskuslaitos. Kuukausikatsaus Suomen saaoloihin.
58: 5-9, 13.

Pulp and Paper Research Institute. 1949 a. Under-
s8kningar betraffande cellulosavedens kvalitet,
utf8rda under aren 1945—48. Granved och sulfit-
cellulosa. Centrallaboratorium AB. Kokf8rs8ks-
avdelningens unders8kningsrapport nr 37.

1949 b. Sulfatcellulosavedens kvalitet. Central-
laboratorium AB. Kokf8rs8ksavdelningens under-
s8kningsrapport nr 38.

Forest Biology Subcommittee. 1963. Existing methods of

value for small sample measurement of wood and
fiber properties. Tappi 46: 150 A-156 A.

101

Forest Products Laboratory. 1960. Longitudinal shrinkage

G a y e

G i o r

G o g g

of wood. Forest Products Laboratory Report No.
1093.

r, K a r 1 and F a b r i c i u s, L u d w i g.
1921. Die Forstbenutzung. Berlin.

d a n o, G. 1951. Studio comparativo sul legno
del larice, dell’abete rosso e del pino silvestre
proveniente dalle foreste alpine e dai rimbos—
chimenti dell’Appennino. Comparative study of
the wood of larch, Norway spruce and Scots pine
from alpine forests and from plantations in the
Apennines. Italia Forestale a Montana. (Cited
from Forestry Abstracts 12: 3305. 1952).

a n 3, J a m e s F. 1961. The interplay of
environment and heredity as factors controlling
wood prOperties in conifers with special
emphasis on their effects on specific gravity.
North Carolina State College. School of Forestry.
Technical Report No. 11.

1962. The correlation, variation and inheritance
of wood prOperties in loblolly pine (Pinus_taeda
L.). North Carolina State College. School of
Forestry. Technical Report No. 14.

n, H a r o l d V. and W o r r a 1 1, J o h n.
1964. Wood quality studies I. A scanning micro-
photometer for automatically measuring recording
certain wood characteristics. Tappi 47: 419-427.
n, R o b e r t P. 1963. The control of pulp
quality in a bleached kraft pulp mill. Tappi 46:
14 A - 26 A.

t h e r, W i 1 l i a m C. 1964. Analysis of
variance. Englewood Cliffs, N.J.> ‘

H a l e, J.D. 1962. Minimum requirements for defining

species norms for quality of variable woods.
Tappi 45: 538-542.

H a 1 e, J. D. and F e n s o m, K. G. 1931. The rate of

growth and density of the wood of white spruce.
Canada Department of the Interior. Forest
Service. Circular 30.

102

H a r d y, S t e v e n S. and W e i l a n d, G e o r g e
W. III. 1964. Weight as a basis for the purchase
of pulpwood in Maine. University of Maine. Maine
Agricultural Experiment Station. Bulletin 14.

H a r t i g, R o b e r t. 1884. Der Einfluss des
Baumalters und der Jahrringbreite auf die
Beschaffenheit des Holzes. Allgemeine Forst- und
Jagd-Zeitung 60: 128-137.

H e d i n, T o r e, M a l m, E r i k and W e n n e r -

b 1 o m, A x e 1. 1963. Comparison of springwood
and summerwood enriched dissolving pulps. Tappi
46: 548-550.

H e i k i n h e i m o, 0 1 1 i. Metsantutkimuslaitoksen

kokeilualueita 1, Vesijako. Helsinki. Undated.

- Metsantutkimuslaitoksen kokeilualueita 3, Punka-

harju. Helsinki. Undated.

- " - Metsantutkimuslaitoksen kokeilualueita 5, Vilp-
pula. Helsinki. Undated.

H e i s k a n e n, V e i j o. 1954. Vuosiluston paksuuden
ja sahatukin 1aadun valisesta riippuvuudesta. 0n
the interdependence of annual ring width and
sawlog quality. Communicationes Instituti Fores-
talis Fenniae 44.5.

- ” - 1957. Raudus- ja hieskoivun laatu eri kasvupai-
koilla. Quality of the common birch and the white
birch on different sites. Communicationes Insti-
tuti Forestalis Fenniae 48.6.

H e i s k a n e n, V e i j o and J o k i h a a r a,

L a u r i. 1960. Puupolttoaineiden lamp8arvoista.
Pienpuualan Toimikunnan tiedotus N:o 33.

H e i s k a n e n, V e i j o and S i i m e s, F.E. 1960.
Ehdotus manty- ja kuusisahatukkien laatuluoki-
tukseksi. Suomen Puutalous 42: 385-389.

H e 1 a n d e r, A. B e n j. 1918. Metsankaytt80ppi.
Porvoo.

H e s s, J o h n M. 1965. A relationship of wood density
surveys to plywood strength values. Proceedings
of the symposium on density - - - a key to wood
quality: 162-168. Madison, Wisconsin.

J a 1 a

103

t, L o u i s A., B e e r s, W a 1 t e r L. Jr.
and Z a c h a r i a s e n, K a r s t e n, A.

1960. Relationships between wood density and

other wood and pulp prOperties. Tappi 43: 169-
173.

e b r a n d t, G. 1954. Untersuchungen an
Fichtenbestanden fiber Zuwachs und Ertrag reiner
Holzsubstanz. Berlin.

y, W.E. 1955. Quality in softwoods. Quarterly
Journal of Forestry 49: 159-164.

0 w a y, T r e v o r. 1965. Carbon 14 used to
study wood density. World Wood 6: 30.

s s a 1 o, L a u r i. 1928. Puuluokitus ja har-
vennusasteikko. A tree-classification and

thinning system. Acta Forestalia Fennica 34.38.

s s a 1 0, Y r j 8. 1948. Pystypuiden kuutioimis-
ja kasvunlaskentataulukot. Helsinki.

1956. Suomen metsat vuosista 1921-24 vuosiin
1951-53. Kolmeen valtakunnan metsien inventointiin
perustuva tutkimus. The forests of Finland from
1921-24 to 1951-53. A survey based on three natio-
nal forest inventories. Communicationes Instituti
Forestalia Fenniae 47.1.

1963. IV valtakunnan metsien inventointi. 2. mean
etelapuoliskon metsanhoitolautakuntien alueryhmat.
Fourth national forest inventory. 2. Southern
forestry board districts. Communicationes Institu-
ti Forestalis Fenniae 57.4.

v a, M a t t i. 1933. Suomalaisen mannyn lujuusomi-
naisuuksista. Communicationes Instituti Forestalia
Fenniae 18.7. . '_
1934 a. Havaintoja puun aseman vaikutuksesta puun
ominaisuuksiin. Investigation into the influence
of the position of a tree in the stand upon the_
prOperties of the wood. Acta Forestalia Fennica
40.9. '

1934 b. Suomalaisen mannyn lujuusominaisuuksista.
II. Strength properties of Finnish pine (£19981
€31Y95t21§)- Communicationes Instituti Forestalia
Fenniae 19.6.

104

(I a l a v a, M a t t i. 1946. Suomalaisen mannyn, kuusen,
koivun ja haavan lujuusominaisuuksista. Strenght
prOperties of Finnish pine, Spruce, birch and
aspen. Communicationes Instituti Forestalia
Fenniae 33.3.

- 1952. Puun rakenne ja ominaisuudet. Helsinki.

" - 1959. Puu teollisuuden raaka-aineena. Puun me-
kaanisteknillisia ominaisuuksia. Tapion Tasku-
kirja, pp. 392-397. Helsinki.

:1 a y n e, B e n A. 1958. Effect of site and spacing on

the specific gravity of wood of plantation-~
grown red pine. Tappi 41: 162-166.

.J e n s e n, W., V i r k c l a, N-E., H u i k a r i, 0.
and P a a r l a h t i, K. 1964. Lannoitetussa
metsassa kasvanut puu massan valmistuksen raaka-
aineena. Suomen Puutalous 46: 408-421.

1K a l l a, J u h a n i. 1966. Mantypinotavaran kulutus
valkaisemattoman sulfaattiselluloosan valmistuk-
sessa Pohjois- ja Etela-Suomessa. Manuscript.

K’e l s e y, J.E. and S t e e 1 e, R.L. 1956. Shrinkage
and density of plantation grown Pinusmradiata,
Australia, Commonwealth Sci. Ind. Research
Organization, Div. Forest Project T.P. 22, Progr.
Rept. No. 2 (Cited from Tappi Monograph Series--
No. 24. 1962).

K e r r, J. 1931. Summerwood formation in southern pines
dependent on capacity of soil to retain moisture.
Forest Worker 4: 17.

K i n n m a n, G u n n o. 1923. Kvalitetsfordringar pa
pappersved och skogsvardsAtgardernas anassande
darefter. Svenska Skogsvdrdsf8reningens Tids-
krift 21: 201-225.

K 1 e m, G u s t a v G. 1934. Undersflkelser av gran-
virkets kvalitet. Untersuchungen fiber die .
Qualitat des Fichtenholzes. Meddelelser fra Det
norske Skogforseksvesen. Bind 5: 197-348.

- " - 1944. Planteavstandens innflytelse pa granvedens
og sulfitcellulosens kvalitet. Effect of planting
space on the quality of spruce wood and sulphite

105

pulp. Meddelelser fra Det norske Skogforsdks-
vesen. Bind 8: 257-293.

K 1 e m, G u s t a v G. 1949. Specific gravity of spruce

wood, its variation in wood structure and pulp
degree of delignification (Sieber chlorine no.),
and the effect of these factors on yield and
sulphite pulp quality. Meddelelser fra Det
norske Skogforseksvesen. Bind 10: 369-396.

1952. Planteavstandens virkning pa granvirkets
kvalitet. The influence of spacing on spruce
quality. Meddelelser fra Det norske Skogforsdks-
vesen. Bind 11: 473-506.

1957. Kvalitetsundersekelser av norsk och tysk
gran. The quality of Norway spruce (Picea1abies)
of Norwegian and German origin. Meddelelser fra
Det norske Skogforseksvesen. Bind 14: 285-314.

K 1 e m, G u s t a v G., L 8 s c h b r a n d t,

F r i t h j o f and B a d e, 0 t t o. 1945.
Undersokelser av granvirke i forbindelse med
slipe- 0g sulfitkokeforsdk. Investigations of
spruce wood in connection with mechanical wood
pulp and sulphite pulp experiments. Meddelelser
fra Det norske Skogfoseksvesen. Bind 19: 1-127.

K 1 e m, G u s t a v S. 1964. Effekten av gjodsling pa

K n i g

K o l 1

3 kvalitetsegenskaper hos vanlig gran (Pipes
abigs). Norsk Skogbruk nr 18: 491-494.

1965. Terrvolumvektsvariasjoner hos fremmende
bartreslag og vanlig gran fra Ser- og Vestlan-
det. Variations in the specific gravity of
foreign softwood species and Norway spruce from
South and West Norway. Meddelelser fra Det norske
Skogforseksvesen. Bind 20: 137-169.

g e, W o 1 f g a n g and S c h u 1 2, H o r s t.
1966. Grundriss der Forstbenutzung. Hamburg and
Berlin. ‘
m a n, F r a n z. 1951. Technologie des Holzes
und der Holzwerkstoffe. Berlin, G8ttingen and
Heidelberg.

K r a h m e r, R o b e r t L. and S n o d g r a s s,

K r a u

K u j a

L a r s

L a s s

L i i r

L u n d

Mia r k

106

J a m e s D. 1965. Development of a sampling
procedure for estimating averages of selected
physical prOperties in a forest stand. Forest
Research Laboratory. Oregon State University.
Corvallis, Oregon.

8, J o h n F. and S p u r r, S t e p h e n H.
1961. Relationship of soil moisture to the
springwood-summerwood transition in southern
Michigan red pine. Journal of Forestry 59: 510-
511. '

1 a, V i l j o. 1946. Koivututkimuksia. Some
recent research data on birches. Communicationes
Instituti Forestalis Fenniae 34.1.

o n, P h i l i p R. 1957. Effect of environment
on the percentage of summerwood and specific
gravity of slash pine. Yale University, School
of Forestry, Bulletin No. 63.

1962. A biological approach to wood quality.
Tappi 45: 443-448.

1963. Stem form deve10pment of forest trees.
Forest Science-Monograph 5.

i l a, I. 1929. Metsatyypin vaikutuksesta puun
painoon. On influence of forest type on weight
of wood. Acta Forestalia Fennica 36: 1.

i, O s m o. 1964. Lastulevyn puuraaka-aine.
Tutkimuksia suomalaisen puun raaka—aineen eraiden
ominaisuuksien merkityksesta valmistettaessa
3-kerroksista 1aakapuristettu lastulevya. Wood
raw material of particle board. Investigations
on the significance of some prOperties of
Finnish wood raw material in the manufacture of
3-layer flatpressed particle board. State Insti-
tute for Technical Research. Publication 80.

b e r g, G u s t a v. 1928. Torrvolymvikten hos,
tall- och granved. Das Trockenvolumgewicht bei
Kiefern- und Fichtenholz. Svenska Skogsvdrds-
f8reningens Tidskrift 27: 564-588.

w a r d t, L.J. and P a u 1, B e n s o n H.
1946. Methods of determining the specific gravity

M a r k

M a r t

M i t c

1 t

107

of wood and wood-base materials. Proc. Amer. Soc.
Test. Mat. 46: 1520-1539 (Cited from Spurr and
Hsiung 1954).

w a r d t, L.J. and W i 1 s o n, T.R.C. 1935.
Strength and related properties of woods grown
in the United States. U.S. Department of Agri-
culture. Technical Bulletin No 479.

s, R a l p h 0. 1951. Influence of crown
reduction on springwood and summerwood distri-
bution in longleaf pine. Journal of Forestry 49:
183-189. '

h e 1 1, H a r 0 l d L. 1958. Wood quality
evaluation from increment cores. Tappi 41: 150-
156.

1961. A concept of intrinsic wood quality, and
non-destructive methods for determining quality
in standing timber. Forest Peoducts Laboratory
Report No 2233. ‘

1964. Patterns of variation in specific gravity
of southern pines and other coniferous Species.
Tappi 47: 276—283.

1965. Highlights of results of the southern
wood density survey. Proceedings of the symposium
on density -- a key to wood quality: 38-64.
Madison, Wisconsin.

h e 1 1, H.L. and W h e e 1 e r, P.R. 1959. Wood
quality of Mississippi’s pine resourches. Forest
Products Laboratory Report No. 2143.

e s e n, P. 1957. Tyndingens indflydelse pa
vedkvaliten. Svenska Skogsvardsf8reningens Tids-
krift 55: l-l6.

r k, E 1 i a s, 1928. Die Qualitat des Fichtenholzes

unter besonderer Rficksichtsnahme auf Schleif-
und Papierholz. Papier Fabrikant 26: 741-747.

M u r t 0, J a a k k 0 0. 1949. Piirteita sulfaattisellu—

loosan 1aadun kehittamismahdollisuuksista. Suo-
men Paperi- ja Puutavaralehti 31: 307-308.

" l 1 e r - S t 0 1 l, W.R. 1949. Photometrische

Holzstrukturuntersuchungen, II. Mitteilung:

N y 1 i

108

fiber die Beziehungen der Lichtdurchlassigkeit von
Holzschnitten zu Rohwichte und Wichtekontrast.
Forstwissenschaftliches Centralblatt 68: 21-63.

n d e r, P e r. 1953 a. fiber die Verteilung von
Rohwichte und Holzfeuchtigkeit in Kiefern- und
Fichtenstammen und ihre Bedeutung ffir die
F18sserei in Schweden. Holz als Roh- und Werkstoff
11: 354-361.

1953 b. Volymviktsvariationer hos planterad gran.
Variations in density of planted spruce. Medde-
landen fran Statens Skogsforskningsinstitut.

Band 43: 3. .

1959. Synpunkter pa produktionens kvalitet.
Kungl. Skogsh8gskolan - institutionen f8r virkes-
lara. Nr U 2.

1961 a. Om trad— och vedegenskapers inverkan pa
ravolymvikt och flytbarhet. I. Tall. Influence of
tree features and wood properties on basic
density and buoyancy. I. Scots pine (Pinus_
si1vestris). Kungl. Skogsh8gskolan - institutio-
nen f8r virkeslara. Uppsatser nr R 35.

1961 b. Om trad- och vedegenskapers inverkan pa
ravolymvikt och flytbarhet. II. Gran. Influence
of tree features and wood properties on basic
density and buoyancy. II. Norway spruce (Picea
éb逧)° Kungl. Skogsh8gskolan - institutionen

f8r virkeslara. Uppsatser nr R 36.

N y l i n d e r, P e r and H a g g 1 u n d, E r i k.

N y y s

1954. Stdndorts— och tradegenskapers inverkan

pa utbyte och kvalitet vid framstallning av
sulfitmassa av gran. The influence of stand and
tree prOperties on yield and quality of sulphite
pulp of Swedish spruce (Pigea_exce1sa). Meddelan-
den fran Statens Skogsforskningsinstitut. Band
44: 11. '

s 8 n e n, A a r n e. 1954. Hakkauksilla kasitel-
tyjen mannik8iden rakenteesta ja kehityksesta.

On the structure and deve10pment of Finnish pine
stands treated with different cuttings. Acta

O 1 s o

O r,m a

109

Forestalia Fennica 60.4.

n, A.R., P o 1 e t i k a, N.V. and H i c o c k,
H.W. 1947. Strength properties of plantation--
grown coniferous woods. Conn. Univ. Storrs Agr.
Expt. Sta., Bull. No. 511. (Cited from Tappi
Monograph Series-N0. 24. 1962).

tula from Whakarewarewa Kaingaroa State Forests.

New Zealand Forest Products Research Notes 1:
1‘24 e

O r t h, T.M. 1965. Use of wood density data by the sout-

P a u 1,

P a u l,

hern pine plywood industry. Proceedings of the
symposium on density - - - a key to wood quality:
124-128. Madison. Wisconsin.

B e n s 0 n H. 1930. The application of
silviculture in controlling the specific gravity
of wood. U.S. Department of Agriculture. Techni-
cal Bulletin No. 168.

1932. Improving the quality of second—growth
Douglas fir. Journal of Forestry 30: 682-686.
1960. The juvenile core in conifers. Tappi 43:
1-2.

1963. The application of silviculture in
controlling the specific gravity of wood. U.S.
Department of Agriculture. Forest Service,
Technical Bulletin No 1288.

B e n s 0 n H. and M a r t s, R a l p h 0.
1931. Controlling the pr0portion of summer-
wood in longleaf pine. Journal of Forestry 29:
784-796.

P e c h m a n n, H. 1954. Untersuchungen fiber Gebirgs-

fichtenholz. Forstwissenschaftliches Central-

P e c h m a n n, H. and W u t z, A. 1960. Haben Mine- .

raldfingung und Lupinenanbau einen Einfluss

auf die Eigenschaften von Fichte- und Kiefern-
holz? Forstwissenschaftliches Centralblatt

79: 91-105.

P h i l 1 i p s, E.W.J. 1960. The beta ray method of

P i l 1

R e n d

R i s i,

R u n k

S a n d

S c h a

n

110

determining the density of wood and the pr0p0r-
tion of summerwood. Journal of the Institute of
Wood Science. No 5: 16-28.

1963. Timber improvement by tree selection and
breeding. World consultation on forest genetics
and tree improvement. FAO/FORGEN 63-7/5.
Stockholm.

0 w, M.Y. 1954. Specific gravity relative to
characteristics of annual rings in loblolly pine.
Forest Products Laboratory Report No. 1989.

1 e, B.J. 1958. Some recent work on factors
affecting the quality of softwoods. Quarterly
Journal of Forestry 52: 308-311.

1959. Fast-grown coniferous timber - some
anatomical considerations. Quarterly Journal of
Forestry 53: 116-122.

1960. Juvenile and adult wood. Journal of the
Institute of Wood Science. No. 5: 58-61.

J. and Z e l l e r, E. 1960. Specific gravity
of the wood of black spruce (Pigea mariana, Mill.
B.S.P.) grown on a Hy1ocomium—Cornus site type.
Laval University Forestry Research Foundation
Contribution 6, Quebed, Canada.

e l, R. 1942. Zellwandforschung unter dem
Blickpunkt des technischen Einsatzes. Zugleich
IV. Mitteilung ”Zur kenntnis der Zellwande
tropischer Laubh8lzer". Holz als Roh- und
Werkstoff 5: 413-427.

e r m a n n, W., S c h w e e r s, W. and

H o h e i s e l, H. 1963. Uber Dichteuntersuch-
ungen an Holz und Holzwerkstoffen mit,g-strahlen.
Paperi ja Puu 45: 681-688.

f e r, E.R. 1949. The influence of volume of
summerwood and rate of growth on the specific
gravity of southern pine pulpwood. Southern 4
Pulp and Paper Manufacturer 12: 166-168.

1961. Effect of condition and kind of wood on
groudwood pulp quality. Forest Products
Laboratory Report No. 2220.

S c h m

111

i d t, H. 1953. Kronen- und Zuwachsuntersuch-
ungen an Fichten des Bayerischen Alpenvorlandes.
Forstwissenschaftliches Centralblat 72: 276-286.

Short course in grading based upon standard grading

S i i m

S i i m

S i r e

S k r 1

rules for southern pine lumber. 1963. Southern
Pine Inspection Bureau, New Orleans.

a s, F.E. 1938. Suomalaisen mantypuun rakenteelli-
sista ja fysikalisista ominaisuuksista. Erikoi-
sesti kutistumis- ja 1aajenemisilmi8sta puun vesi-
pitoisuuden vaihdellessa. Helsinki.

1952. Strength grading rules for Finnish
structural sawn timber. Helsinki.

1965. Economic advantages of stress-grading.
Reports presented at the symposium on the
economic aspects of, and productivity in, the
sawmilling industry. Supplement 2 to volume 18
of the Timber Bulletin for Europe. Volume 3,
Geneva.

e s, F.E. and L i i r i, O s m o. 1952. Puun
lujuustutkimuksia I. Pienet virheett8mat manty-
koekappaleet. Investigations of the strength
prOperties of wood. I. Tests on small clear
specimens of Finnish pine (Pinus_§i1vestris).
Valtion teknillinen tutkimuslaitos. Tiedoitus
103.

n, G u s f a f. 1952. Hakkuun vaikutuksesta kuu-
sipuun rakenteeseen korpimailla. On the effect
of releasing cutting upon wood structure of
spruce on peat-moors. Communicationes Instituti
Forestalis Fenniae 40.32.

1959. Eraita havaintoja keskisuomalaisen ja
paikallisen mantyrodun biologisista ja teknilli-
sista ominaisuuksista Para-Pohjolassa. Some
observations on the biological and technical _
properties of the local and Central-Finnish pine
provenances in North Finland. Silva Fennica
96.1.

p e n, J. and R i a s 0 v a, T. 1958. Stavba
dreva a objemova vaha borovice obyéajnej (Pipus_

S m i t

S n e d

S p u r

112

si1vegtg1g) in Slovakia. The wood structure and
density of Pinu§_sylvestr1s in Slovakia.
Drevarsky vyskum 3: 27-48 (Cited from Forestry
Abstracts 20: 2353. 1959).

h, D.M. 1954. Maximum moisture content method
for determining specific gravityof small wood
samples. Forest Products Laboratory Report

No. 2014.

1955. A comparison of two methods for
determining the specific gravity of small
samples of second-growth Douglas-fir. Forest
Products Laboratory Report No. 2033.

1965. Rapid measurement of tracheid cross--
sectional dimensions of conifers: Its applica-
tion to specific gravity determinations. Forest
Products Journal 15: 325-334.

e c o r, G e o r g e W. 1957. Statistical
methods. Ames, Iowa.

r, S t e p h e n H. and H s i u n g, W e n - -
Y e u. 1964. Growth rate and specific gravity
in conifers. Journal of Forestry 52: 191-200.

Standard grading and dressing rules for Douglas-fir, west

coast hemlock, Sitka spruce, and western red
cedar lumber. Revised May 1, 1962. Number 15.
1962. Portland, Oregon.

Standard grading rules for southern pine lumber. 1963.

S t a u

S t o c

S t o n

Effective February 1, 1963. Southern Pine
Inapection Bureau, New Orleans, Louisiana.

f f e r, 0. 1892. Untersuchungen fiber spezi-
fisches Trockengewicht, sowie anatomischen Bau
des Holzes der Birke. Forstlich-naturwissen-
schaftliche Zeitschrift 1: 145-163.

k m a n, L e n n a r t. 1962. The influence of
some morphological factors on the quality of .
spruce sulphite and pine sulphate pulp. Svensk
Papperstidning 65: 978-982.

e c y p h e r, R o y, C e c h, F r a n k 1 i n
C. and Z 0 b e 1, B r u c e J. 1964. Inheri-
tance of specific gravity in two- and three--

'HI,’

.1..\ ‘wfyll ..

 

+4

dfoJe 3 rd.

..Jiet. .ri 141..-.w‘ 2.0 I”, “I

.‘. [fit-[iv

11.

HI! huhflfllq... ..
A

., .. ....P. .

T a m m

113

year-01d seedlings of loblolly pine. Tappi 47:
405-407.

i n e n, 2 a c h r i s. 1962. Fuktighet, volym-
vikt m.m. hos ved och bark. I. Tall. Moisture
content, density and other properties of wood
and bark. I. Scots Pine. Kungl. Skogsh8gskolan-
institutionen f8r virkeslara. Uppsatser nr R 41.
1964. Fuktighet, volymvikt m.m. hos ved och
bark. II. Gran. Moisture content, density and
other properties of wood and bark. II. Norway
spruce. Skogsh8gskolan. Institutionen f8r
virkeslara. Rapporter nr R 47.

Tapion Taskukirja. 1959. Tarkeimmat puutavaralajit.

Py8rea puutavara. Pp. 241-245. Helsinki.

Tappi Monograph Series - No. 24. 1962. The influence of

environment and genetics on pulpwood quality.
An annotated bibliography.

s, M i c h a e 1 A. and W a h 1 g r e n,

H a r 0 1 d E. 1963. A comparison of increment
core sampling methods for estimating tree
Specific gravity. U.S. Forest Service Research
Paper SE-7.

b j o r n s e n, E y v i n d. 1961. Variation
in density and fiber length invwood of yellow
p0p1ar. Tappi 44: 192-195.

T 0 d a, R y 0 0 k i t i. 1958. Variation and herita-

bility of some quantitative characters in
Cryptomeria, Silve Genetics 7: 87-93.

d e l e n b u r g, R e i n h a r d. 1939. Das
Holz als Rohstoff. Mfinchen.

d e 1 e n b u r g, R. and M a y e r - W e g e -
1 i n, H. 1955. Das Holz als Rohstoff. Mfinchen.
m i s, G e o r g e. 1964. Microscopic measure-
ment of the amount of cell wall substance in
wood and its relationship to specific gravity.
Tappi 47: 675-677.

b u l l, J.M. 1937. Variations in strength of
pine timbers. South African Journal of Science

33: 653-682.

T u r n

114

b u l 1, J.M. 1942. A study in stem design.
The Journal of the South African Forestry
Association 8: 112-117.

1947. Some factors affecting wood density in
pine stems. Union of South Africa, Depart-
ment of Forestry. Publication No. 82.

Uittoteho ry. and 0y Keskuslaboratorio. 1966. Eraita

tutkimustuloksia manty- ja kuusipaperipuun
tilavuus- ja painosuhteista seka niiden vai-
kutuksesta massan saantoon Etela- ja Pohjois-
Suomessa.

U.S. Forest Service. 1965 a. Southern wood density

survey. 1965 status report. U.S. Forest
Service Research Paper. FPL-26.

U.S. Forest Service. 1965 b. Western wood density survey.

V 0 1 k

V u 0 k

V u o r

W a h 1

Report Number 1. U.S. Forest Service Research
Paper. FPL-27.

e r t, E r i k. 1941. Untersuchungen fiber Gr8s-
se und Verteilung des Raumgewichts in Nadel-
holzstammen. Schriftenreihe der Herman-G8ring-
Akademie der Deutschen Forstwissenschaft. Band
2: 1-133.

1 1 a, Y r j 8. 1960. Siperialaisten 1ehti-
kuusikoiden kehityksesta ja merkityksesta
maamme metsataloudessa. On deve10pment 0f
Siberian larch stands and their importance to
Forestry in Finland. Communicationes Instituti
Forestalis Fenniae 52.5.

i s t o, I l m a r i. 1937. Suomen puuvanuke-
teollisuuden kayttaman kuusipaperipuun laatutut-
kimuksia. Untersuchungen fiber die Qualitat des
von der Zellulose- und Schleifholzindustrie
verbrauchten Fichtenpapierholzes in Finland.
Communicationes Instituti Forestalis Fenniae
23.2.

g r e n, H.E. 1965. Forest Service western wood
density survey: Phases II (increment core
processing) and III (estimating tree specific
gravity). Proceedings of the symposium on

W a h 1

W a 1 1

W a 1 t

W e g e

W h e e

Wijk

W i l d

115

density --- a key to wood quality: 21-35.
Madison, Wisconsin.

g r e n, H a r o 1 d E. and F a s s n a c h t,
D 0 n a l d L. 1959. Estimating tree specific
gravity from a single increment core. Forest
Products Laboratory Report No. 2146.

d é n, P a u 1. 1934. Tutkimuksia_koivupuun
anatoomisen rakenteen ja teknillisten ominai-
suuksien keskinaisesta riippuvaisuudesta solu-
mittauksien perusteella. Untersuchungen fiber
die Abhangigkeit der technischen Eigenschaften
vom anatomischen Bau des Birkenholzes nach
Zellenmessungen. Acta Forestalia Fennica 40: 14.
e r s, C.S. and B r u c k m a n n, G. 1964.

A comparison of methods for determining volume
of increment cores. Journal of Forestry 62: '
172-177.

1 i u s, Th. 1939. The presence and prOperties
of knots in Finnish spruce. Investigations
concerning the origin and characteristics of
branch and knot wood in Finnish spruce with
particular consideration given to the raw
material needs of the paper industry. Acta
Forestalia Fennica 48: 1.

1946. Det finska granvirkets egenskaper och
kvalitetsvariationer. Svensk Papperstidning
49: 51-61.

1949. Puun kulutuksesta ja puun 1aadun merki-
tyksesta hiomoissa ja selluloosatehtaissa.
Tapion Taskukirja, pp. 361-363.

1 e r, P.R. and M i t c h e l 1, H.L. 1962.
Specific gravity variation in Mississippi
pines. Forest Products Laboratory Report No.
2250.

a n d e r, A u g. 1897. Untersuchungen der
Festigkeits-Eigenschaften schwedischer
Holzarten. G8teborg.

e, S.A. and P a u l, B.H. 1951. Rate of

growth and composition of wood of quaking

W i l d

W i 1 1

Y a n d

Yleiset

Y 1 i n

Z 0 b e

Z 0 b e

Z 0 b e

116

and largetooth aspen in relation to soil
fertility. Transactions of the Wisconsin
Academy of Sciences, Arts and Letters 11:
245-250. (Cited from Tappi Monograph Series-
No. 24. 1962).

e, S.A., P a u 1, B e n s 0 n H. and M i k 0-
1 a, P e i t s a. 1951. Yield and quality of
jack pine pulpwood produced on different
types of sandy soils in Wisconsin. Journal

of Forestry 49: 878-881.

i a m s, R o b e r t F. and H a m i 1 -

t o n, J o h n R. 1961. The effect of
fertilization on four wood prOperties of slash
pine. Journal of Forestry 59: 662-665.

1 e, D.O. 1956. Statistical evaluation of the
effect of age on specific gravity in loblolly
pine. Forest Products Laboratory Report No.
2049.

vanerikoivujen mittaus— ja laatumaaritelmat.
1961. Jyvaskyla.

e n, A r v o. 1951. Uber den Einfluss des
Spatholzanteils und der Jahrringsbreite auf
die Rohwichte beim finnischen Kiefernholz.
Holz als Roh- und Werkstoff 9: 449-456.

1, B r u c e: 1961. Inheritance of wood
prOperties in conifers. Silvae Genetica 7:
67-70. '

1963. Breeding for wood properties in forest
trees. World consultation on forest genetics
and tree improvement. FAO/FORGEN 63-7/0.
Stockholm.

1, B.J., G 0 g g a n s, J.F., M a k i, T.E.
and H e n s 0 n, F. 1961. Some effects of
fertilizers on wood prOperties of loblolly
pine. Tappi 44: 186-192.

1, B r u c e J. and M c E l w e e, R 0 b e r t
L. 1958. Variation of cellulose in loblolly
pine. Tappi 41: 167-170.

 

117

Z 0 b e 1, B r u c e, W e b b, C h a r 1 e s and
H e n s 0 n, F a y. 1959. Core 0r juvenile
wood of loblolly and slash pine trees.
Tappi 42: 345-356.

 

 

MUCH

119,111,111111 11

1

 

1(le-

.-
‘
-e
v

.‘w-
I“
I\-’
e “A e, ‘

.5":-
'1‘

I
l

’d' , a_
.0 -‘
l
‘ _ w.
‘1 I‘

I“
”‘7 w .u§

.fi““.""5

.‘2

n ._J

4 ~ ."
' ‘
a

‘1. -‘ .-
3%" u“ ‘2 .‘J