MODIFICATION OF GROWTH CAUSED BY PRESSURE AND TRACTION, 815 
they dry up and fall off, while those which have formed discs increase in thickness 
and become woody. 
I. The retarding effect on growth of an external pressure on the cells is very evident 
in the formation of the annual rings in wood. In the earlier editions of this work 
I called attention to the fact that the larger radial diameter of the wood- cells in the 
portion of the rings formed in the spring, and their smaller radial diameter in the por- 
tion formed in the autumn, might possibly depend on a difference in the pressure from 
the surrounding bark to which the cambium and the wood are subject, this pressure 
being less, as we have shown, in the spring, and constantly increasing during the 
summer. This hypothesis has been fully confirmed by H. de Vries's investigations ^ 
In branches two or three years old he increased the pressure of the bark in the spring 
by firmly winding string round them at particular places. The experiment showed in 
all cases, firstly, that the absolute thickness of the annual ring was less beneath the liga- 
ture than the mean thickness of the same annual ring at some distance above or below 
that spot. In several instances the difference was so considerable that the spot where 
the experiment was made appeared of considerably less diameter even to the naked eye, 
and this effect was increased by the formation of cushions of wood immediately above 
and below the ligature. Secondly, the absolute thickness of the ' autumnal layer ' of 
wood (up to the middle of August, when the increase in diameter of the tree on which 
the observations were made ceased) was always greater, and generally considerably so, 
at the spot where the experiment was made, than the normal thickness. In the trees 
examined {Acer Pseudo-platanus, Salix cinerea, Populus alba, Pa'via) the autumnal wood 
formed at this spot consisted of fibres flattened radially, between which were a smaller 
number of vessels than in the normal wood ; its composition was therefore the same as 
that of the normal ' autumnal wood.' The normal autumnal wood of Ailanthus glandu- 
losa consists almost entirely of wood-parenchyma-cells flattened radially; while the 
autumnal wood formed beneath a ligature made in May consisted of a thick layer of 
flattened fibres, between which a few vessels could be seen. These results show that 
when the pressure is increased, the formation of the autumnal wood begins at a time 
when, under normal pressure, a large-celled woody tissue is still being formed. 
A diminution of pressure is obtained by making radial longitudinal incisions into 
the bast-tissue. The strips of bast contract somewhat tangentially, since their tension 
ceases. Near the incisions the pressure of the bast upon the wood is entirely removed ; 
but in the middle between two adjacent incisions a considerable pressure always remains. 
The fresh portions of tissue which are formed next to the wounds differ to the greatest 
extent in their composition from the ordinary structure of the wood. A layer of wood 
of the ordinary structure is formed, on the other hand, in the portions of the cambium 
at the greatest distance from the incisions, and afterwards also on the outside of the 
abnormal portions of tissue. But it is only the tissue consisting of wood formed under 
artificially diminished pressure that we have at present to consider. The incisions 
were mostly 2 to 3 cm. long, and were made in the periphery of two- to three-year-old 
branches at distances of from 4 to 6 cm. in the middle of June and the middle of July, 
and therefore after the formation of the normal autumnal wood had already begun. 
The eflfect of the decrease of pressure was first of all shown, after the branches had 
been cut off in the middle of August, by a considerably greater increase in thickness 
at the spots than above or below^ them. On the transverse sections the thickness of 
the annual ring was greatest near the incision and decreased gradually from there to 
the middle points between two incisions. The layer of wood formed after the com- 
mencement of the experiment was often more than tvi^ice as thick at the former as 
at the latter spots. P^or a more exact investigation only those pieces were used in 
which a layer of distinctly flattened fibres of autumnal wood had been formed before 
1 H. de Vries, Flora, 1872, No. 16, 
