Bulletin 419, May, 1923
Histological Studies and Practical Considerations

L. H. MacDaniels
Published by the Cornell University Agricultural Experiment Station Ithaca, New York

The main crotch angle is too narrow, the branches are too-nearly equal in size,  and too many branches arise near the same point. Corrective pruning of the tree when young would have prevented this condition. Bracing may yet save the tree.

It is well recognized that one of the chief aims in pruning a young tree is to produce a framework which can support maximum crops when the tree comes to bearing age. From this standpoint, one of the most important considerations is the strength of the main crotches of the young tree. Occasionally a limb will break without the crotch spitting, but ordinarily the crotch will give way first. It is therefore of considerable practical importance to determine the factors affecting the strength of crotch structure.

a, Structure of a node as seen in a longitudinal section through a bud and a twig after the first season's growth
b, The same node after the second season's growth. The new layer of wood and bark formed by the cambium is shown

Experienced horticulturists have long been aware of a number of practical considerations in the pruning of trees to produce a satisfactory framework. Thus, it is known that it is bad practice to leave a crotch in which the two branches are of the same size; also, that the angle between the two arms of a crotch should not be too narrow. It is a matter of common observation, too, that varieties differ widely in the nature and strength of the crotch union. Some varieties make crotches with strong unions, while others are weak, depending on the growth habit.

The purpose of this study was to find the more fundamental reasons for weakness in crotches, especially the histological and anatomical structure underlying; to ascertain how the structure of the wood of a crotch differs from that of normal wood; and to investigate other questions of a similar nature. The observations are based on the examination of microtome sections of the tissue of different types of crotches, macroscopic dissection, and field tests.

Wherever a lateral bud is formed, there is the possibility of a crotch developing. The bud is itself a potential branch, and, with its elongation, a crotch is formed at the point of junction of the secondary shoot and the main branch. The structure of the normal node is illustrated in figure 1, a, which shows a part of a twig cut longitudinally through the pith and a bud, as it would appear in the winter after one summer's growth. It will be noted that there is a gap in the woody cylinder of the twig, so that the pith is continuous out into the growing point of the lateral bud. The cambium layer which covers the whole woody cylinder extends up around the bud, ending in the meristematic tissue of the growing point. When growth starts in the spring, the growing point rapidly elongates, a cambium layer is formed around the new shoot which is continuous with that of the main shoot, and, as the season advances, a sheet of woody tissue is laid down, covering both the main branch and the secondary branches. At the end of the second season, the appearance of the node is as shown in figure 1, b. With each succeeding season's growth, a new layer of wood, or annual ring, is laid down by the cambium over the entire woody structure of the crotch, and, of course, additional layers of bark (phloem) are formed on the outside of the cambium.

In discussing crotch structure, certain terms may need explanation. Thus, the point of insertion is the place where the pith of the secondary branch joins the pith of the main limb. The plane of the crotch is a plane passing through the pith of both branches and the point of insertion. The angle of the crotch is the upper, or inside, angle which the secondary branch makes with the main branch in the plane of the crotch. The axis of the crotch is a line in the plane of the crotch passing through the point of insertion and bisecting the angle of the crotch.


One of the most important factors in relation to the strength of a crotch is the crotch angle. In general, it may be said that the narrower the angle, the weaker is the union, although this does not always follow, as other factors may enter in. The reason for this is found, in large part, in the way in which the annual layers of wood and bark are laid down over the woody cylinder. This is illustrated graphically in figures 2, 3, 4, and 5, which show the successive layers of woody growth on different types of crotches. The diagrams are taken from actual specimens. In figure 2 is shown the condition found in a strong crotch from a French crab seedling. In a wide-angle crotch of this type, the cambium in the angle of the crotch in the region of the axis is able to build up the tissue of the angle rapidly enough to prevent the increase in size of the two arms of the crotch from forcing the bark of the two limbs together. A crotch of a somewhat narrower angle, but one in which the crotch cambium is still able to fill up the angle and give a strong union, is shown in figure 3. In a crotch of this type, it will be noted that the annual layers of wood formed in the crotch angle are much thicker than those formed elsewhere on the woody cylinder. In figure 4, taken from a Northern Spy tree, is shown what happens in a narrow-angle crotch where the growth in diameter of the two arms is so rapid that the bark is forced together before the angle is filled in by the crotch cambium. In such cases, the adjacent bark of the two limbs is subjected to considerable pressure by the increase in diameter of the woody cylinders of the two branches. This results in the death of parts of the bark which are forced together and most tightly squeezed. In many cases, the pressure is such that much of the old bark is forced out and the cambium of the two branches unites again. Usually, in such a case, parts of the old dead bark are left embedded in the wood, as shown in the diagram, the cambium growing up around this dead bark and eventually sealing it in. It is evident that such a condition makes a weak union, as, wherever parts of the bark are included, there is no actual union of the tissue, the woody part of the two arms not being joined at all.

The drawing shows the relative thickness of the successive annual layers of wood laid down by the cambium in the crotch angle, and also the irregular convolutions of the cambium layer (C) and the live bark (L.B.) at this point   In a crotch of this type, it is still possible for the cambium to fill in the crotch angle with wood fast enough to prevent the bark of the side arms from coming together
In this case the bark of the side arms in the crotch angle comes together before the crotch angle is filled with woody tissue. The result is that parts of the bark (D.B.) are squeezed off and become embedded in the wood, thus weakening the crotch union   The wood in the crotch angle is everywhere separated by bark, which near the point of insertion is rotten (R). This grades into dead bark (D.B.), which gives place farther up to living bark (L.B.). A crotch of this type is practically sure to split

A more extreme example of the same sort is shown in figure 5. In a crotch of this type the angle is so narrow that often, after the first few years, there will be no actual union of the wood of the two arms of the crotch along the axis. The bark nearest the point of insertion is decayed, much of the remainder is dead, and it is only a matter of time until more of the live bark in the crotch angle will die. Occasionally, in crotches of this type, the pressure of the lateral growth of the adjacent limbs on the inside of the crotch may be so great as to force the limbs apart and cause the crotch to split. Such a crotch has practically no support other than the actual strength of the tissue on the under side of the branch. This condition is, of course, extreme, yet it is found fairly frequently in some varieties. A somewhat similar condition is sometimes found in crotches of wider angle when the crotch cambium has been winterkilled. In these cases, however, the condition is likely to be even more serious, because decay starts almost immediately in the dead tissue and may spread rapidly into the heart of the tree, causing breakage before the living cambium has had time to unite over the dead tissue.

The type of angle formed varies widely with different varieties and can be counted as a varietal characteristic. Thus, such varieties a McIntosh characteristically form medium wide angles and strong unions, and need little attention in this respect. Varieties like Northern Spy are typically narrow and frequently cause trouble. At the other extreme are such varieties as Rhode Island Greening, which practically always produce wide angles.

In order to find the relation of width of angle to strength of crotch, twenty-six crotches were tested by finding the weight required to break them. The method was to strap the main limb of the crotch to an upright post, and then hang weights on the side arm, at a point one foot perpendicular distance from the main limb so that the leverage of the weight was the same in each case. Measurements were taken on the width of the angle and on the diameter of each arm of the crotch. In order to make the results comparable, the breaking weights were computed by simple arithmetical proportion to correspond to side-arm diameters of 1 inch in each case. Thus, if the side-arm diameter was 1 1/4 inches and the breaking strength 155 pounds, by using the proportion 1 1/4: 155 = 1: x the value of x is 124, which would be an indication of the breaking strength if the side arm was 1 inch in diameter. Arranging the different results in order of width of angle from the narrowest to the widest, and averaging the weights in groups of four, there was an increase in breaking strength from 6 pounds, in the case of narrow angles around 30 degrees, to 107.7 pounds with angles of about 75 degrees. In two of the wider angles, the side branch broke at a point away from the crotch, showing that the crotch union was stronger than the branch.


Crotch no. Crotch
Diameter of
main branch
of side
Proportional reduction
Diameter of main
branch (inches)
Diameter of side
branch (inches)
Breaking weight
Average by
groups of four
8 29 1.31 1.12 68 1.17 1 61  
7 29 0.94 0.87 40 1.08 1 46  
2 34 1.37 1.12 80 1.22 1 71  
3 35 1.12 1.00 48 1.12 1 48  
19 37 1.7 1.12 110 1.56 1 98 56.5
21 39' 1.37 1.00 60 1.37 1 60  
23 39 1.31 0.87 40 1.50 1 46  
1 39 1.25 0.87 45 1.44 1 52 64.0
17 39 1.00 0.94 70 1.06 1 74  
20 41 1.12 1.00 75 1.12 1 75  
22 43 1.06 1.06 55 1.00 1 52  
18 43 1.75 1.06 64 1.65 1 60 65.2
24 44 1.00 1.00 155 1.00 1 55  
4 45 1.12 1.00 60 1.12 1 60  
5 45 1.12 1.00 72 1.12 1 72  
26 49 1.87 1.25 135 1.50 1 108 73.7
12 50.5 1.18. 0.94 65 1.26 1 69  
11 52 1.12 1.12 102 1.00 1 91  
25 55 1.67 1.25 142 1.34 1 114  
16 57 1.37 1.06 107 1.29 1 101 93.7
14 57 5.50 1.12 117 1.34 1 104  
9 67 1.37 1.06 88 1.29 1 83  
13 75 1.30 1.12 135 1.16 1 120  
10 78 1.12 1.25 155 0.90 1 124 107.7

The experimental error in such a test was naturally very large, as the crotches were not of the same variety, and other variable factors were concerned which could not be standardized. Nevertheless, it would seem as if the results had some significance, in that they point consistently in the same direction.

With proper care in pruning from the time the trees are set, injury to the trees from narrow crotch angles can be largely controlled. First of all, in pruning the two-years-old tree at the time of setting, preference should be given, in so far as spacing allows, to the branches that already have sufficiently wide angles. During the next years the trees should be carefully watched and important limbs that tend to form bad crotch angles suppressed. While the tree is young, it is possible to suppress one limb and train another to take its place without excessive pruning. In many cases the leading shoots of the main branches can be forced out by allowing the growth on the inside of the crotches in the center of the tree to remain. The direction in which any shoot grows depends in large part on light, and if the center of the tree is already occupied the new growth will be forced to extend more nearly horizontally. Of course, care must be taken to prevent branches in the center of the tree from making so much growth that they shade out the limbs which are to be the permanent scaffold. Another reason for leaving the growth which occurs along a limb above the inside of the crotches, is that this growth contributes largely to the growth of the limb on that side and also to the building-up of the tissue in the crotch angle, as well as hastens the maturity of the bark around the crotch, which is particularly susceptible to winter injury.

The broken surfaces are nearly smooth, and the tissue of one half does not extend around nor interlock with that of the other

In this early pruning, it should be remembered that as soon as the tree reaches bearing age the crotch angles will be materially widened by the weight of the crop. It is therefore neither necessary nor desirable to make the young tree spread to what would be an ideal form before crops are borne. What is essential is to make sure that the main branches have sufficiently wide angles and enough room to allow proper filling-in of the crotch angles by the crotch cambium without pinching or including the bark.


Another very important factor in the strength of crotches is the relative size of the two arms. Other things being equal, it is generally true that crotches with both limbs of the same size are weak and are likely to split, whereas those with the two limbs unequal form a much stronger union and have a greater breaking strength even though the limbs concerned may be actually smaller. This is due to the fact that in the case of unequal arms the woody tissue of the side branch is actually embedded in that of the main axis; whereas with the arms equal, this is not the case, the strength of the union in such crotches depending entirely on the strength of the tissue where the two arms meet. This condition is readily shown photographically in figures 6, 7, and 8. The type of union usually found where the arms are of the same size is shown in figure 6. It is seen that the split surface is practically smooth, showing that there was no interlacing of fibers. The wood of the two branches comes together along the line of cleavage, but at no point do the annual rings overlap. In figure 7 the condition is that in a crotch with the side branch smaller than the main limb. Here the base of the side branch is well embedded in the wood of the main limb, making a strong union. The appearance of these two types as seen in cross section is illustrated in figure 8. Where the arms are equal, the split is straight across, as in the section at the left in the photograph; where the arms are unequal, it is seen that the larger branch has grown around the smaller on either side, giving the condition seen in the other two sections.

The base of the smaller arm is embedded in that of the larger, giving a strong union

On the left, a crotch with equal arms showing a straight line of cleavage. In the center and on the right crotches with unequal arms showing how the wood of the larger arm extends around that of the smaller

Crotches with equal arms can, of course, be readily corrected by pruning, especially when the tree is young. Heading in, cutting off some of the lateral branches, or otherwise pruning one side, will usually dwarf that side so that in a few years the unpruned side will be appreciably larger. From the standpoint of producing unequal crotches, the central-leader or modified-central-leader type of head is of great advantage, in that such a method always aims to have the leader larger than any of the side branches.

In this connection, the undesirability of allowing the main branches to me from the same point of the trunk may be brought out. When this condition is allowed, the base of any one branch is not embedded in the woody tissue of a larger limb, but merely lies adjacent to the wood of other limbs coming off at the same point. When all the limbs are loaded, the strain comes upon the crotch union from both sides, with the result that several limbs may go down at the same time, as frequently happens with young overloaded trees. Such a condition is shown in figure 9, where it will be seen that the tissue making up the trunk belongs entirely to the branches which come off at nearly the same point. There is no wood between any two adjacent branches which does not belong to one or the other of them; hence, When all are loaded, there is little to support them except the strength of the crotch union and the wood on the under side of the branch. In pruning trees at the time of setting, the limbs should be spaced as far apart along the axis as is possible without getting the lower branches too low. Six or eight inches is not too much.

There are too many limbs arising at the same point, so that the wood of each limb merely lies adjacent to that of the next one instead of being embedded in the upright

The number of main branches is also important in this connection. With two or three main branches other than the leader, it is usually possible to have them spaced far enough apart so that they will have their bases embedded in and supported by woody tissue belonging to the main axis, and not merely adjacent to that of other branches originating at nearly the same point. When four or five branches besides the leader are left to form the scaffold of the tree, the undesirable crowded condition shown in figure 10 is likely to result.

With many nursery trees it is impossible to space the limbs in an ideal manner because there are only a few crowded limbs from which to choose. In such cases a very satisfactory tree can be produced by leaving two lateral branches only, one on each side of the leader. These may arise at the same level without damage to the later framework of the tree, because there is ample room for both to become embedded in the tissue of the leader. If the leader is not headed back as severely as the laterals, it is almost certain to make the largest growth and to continue to do so for a number of years without further attention.

It is a matter of common observation that the critical age for serious damage from splitting of the main crotches of apple trees is about that when they bear their first heavy crops. It is rather rare for older trees to go down under a load unless they have been weakened by decay from one cause or another. This would, of course, be partially explained by the fact that in the case of the older trees, it has been a survival of the fittest and the weaker ones have disappeared. It would seem, however, that there are other factors involved. For one thing, in the case of young trees that have been growing rapidly there has been a greater proportional increase in length of branch as compared to girth than in the older trees, where growth in length is slower. There are, however, no experimental data to show that this would not be compensated for by the greater leverage in the case of the old tree, where the bearing surface is farther from the trunk. Another consideration is that crotches are undoubtedly made stronger by the formation of heartwood, for this tissue is intrinsically stronger than sapwood, as has been often shown in the case of other kinds of wood.

As apple trees grow old, there is a tendency for the limbs to become equal in size. This happens regardless of the system of training when the tree was young. As the leader or the more upright limbs reach a certain height, there seems to be some inhibiting influence which automatically checks their growth. The other branches, however, which have not approached this maximum height, continue vigorous growth, with the result that in time they become practically equal to the leader in length and also in girth.

This behavior is frequently observed with other species, each having its particular maximum height at which growth is checked. Thus, a sumac or a choke cherry may make an annual growth of several feet for three or four seasons, but as a height of from fifteen to twenty feet is reached, growth becomes slower and lateral branches begin to overtake the leader, until, as in old plantings, practically all the limbs have attained the same height. In the case of old elms the trees frequently present a flat-topped appearance, with many limbs of approximately the same size.

At the end of the second year, the cambium occupied the position A-B-C. It now occupies the much shorter distance D-E-F. The line B-E shows the change in direction of the line of cleavage due to different relative rates of growth of the two branches. Such overlapping of annual rings strengthens the union

This equalizing process in the case of apple-tree crotches, in view of the principle brought out earlier that equal crotches are fundamentally weak, might at first sight seem to indicate that the crotches of old trees would become weaker as the branches become equal in size. The contrary, however, is the result, for such equalization may result in an even stronger crotch union. This can be understood when it is recalled that in the young tree the base of the smaller side branch is embedded in the larger upright, the annual rings of wood belonging to the upright extending over those of the lateral. This extension of the annual rings of the main branch over those of the lateral will continue as long as the thickness of the annual ring of the main branch in each succeeding year exceeds that of the lateral. As soon, however, as the lateral begins to make the thicker annual ring, the rings of the lateral will begin to overgrow those of the main branch. A continuation of this relative rate of growth will result in an interlocking of the tissue of the successive annual rings, and consequently in a stronger union. In fact, any marked difference in the relative thickness of the annual rings over the two branches helps to make a stronger union by this process of overgrowing.

Bracing would have held the tree together until the bark had united across the wound in the crotch and new layers of wood had been formed. The arrow indicates the swollen bark at the edge of the wound, where healing is in process.

An illustration of the way in which annual rings overlap is seen in figure 10.  The line extending irregularly from the crotch insertion through the crotch angle from B to E shows the approximate points where the tissue of the two branches meets and the line along which cleavage would usually take place. This figure gives but an inadequate idea of the efficiency of this overlapping in increasing the strength of the union, because it shows the intersection of the annual rings in the plane of the crotch only. By referring back to figure 7, it can be seen that the overgrowth extends around the sides of the crotch also, so that the smaller limb extends into the larger as a cone of tissue. When the annual rings on the smaller branch become thicker than those formed on the larger, a reverse cone of tissue will be formed which will help to knit the crotch more firmly.


The question of bracing apple trees to prevent breakage of crotches is worthy of serious attention. The value of a twelve-years-old apple tree is considerable, whether figured on the original cost plus the interest on the value of the land used and the cost of labor for cultivation, or on the value of the crop actually produced. Granted that the net value of the crop for one year is only four or five dollars, still that would be interest on from seventy-five to one hundred dollars. It is therefore well worth while to take every precaution to save trees that are in danger of breaking. The trees that need bracing will be those of varieties which naturally tend to form weak crotches, those which have formed weak crotches through neglect of the considerations previously explained, and those in which the crotches are weak due to some injury. In the last-mentioned case, winter injury is the most frequent cause and is of the greatest importance, as crotches so injured are almost certain to decay and break if they are not braced. The results of neglect after winter injury are graphically shown in figure 11.  The tree shown here could have been saved if it had been properly braced the year following the injury.

There are several methods of bracing which are used by so-called tree surgeons and others, and which accomplish the desired purpose. Possibly the strongest method where two opposite limbs of a crotch are to be fastened together, and where the limbs are large, is to bore holes through these limbs about five to eight feet from the point of insertion, in which either ring bolts or bolts with a hook on the end are inserted. The bolts are anchored by a washer and nut on the outside of the branch. The best practice is to cut away the bark under the washer so that the washer will rest against the solid wood. The two limbs can then be supported by several strands of heavy galvanized wire put through the rings on the bolts and twisted together, or, in the case of very large limbs, by a chain.

Another satisfactory method is to use large wood screws or lag screws with rings or hooks on the ends, the size of the screw depending on the size of the limb to be braced. The hole bored for the screw should be considerably smaller than the screw itself, so that the screw will hold satisfactorily. Wire is fastened to the end of the screw as described in the preceding paragraph.

In the case of young trees, a satisfactory, quick, and cheap method is to fasten the wire with large screw eyes from 2  1/2 to 3 1/2 inches long. Staples may be fairly satisfactory for small limbs, although it is frequently difficult to drive them into the wood without their bending. After a year or so, these screw eyes or staples become so embedded in the wood that it is practically impossible to get them out. A crotch braced in this manner with screw eyes is shown in figure 12 (upper left corner).

Another condition that may be found is one in which there are three or more limbs of about equal size, any one of which is in danger of breaking down. Here, instead of fastening any two limbs together, all the limbs can be fastened to a central ring which will take the strain from all directions. This method of bracing is valuable as a substitute for propping limbs, even when the crotches are not weak, and has proved particularly valuable in bracing prunes, apricots, and peaches on the Pacific coast. It is adapted primarily to trees trained to a vase form or some other type of open head. Care should be used to have the braces placed high in the branches, for if they are S placed low the limbs may break either at or above them. An open-headed apple tree with five limbs wired to a central ring is shown in figure 13.

In this and all the foregoing cases, care must be taken to have a good quality of galvanized wire that will not rust out too quickly. With apple trees, the chances are that if the wire lasts eight or ten years the crotches will be able to support the load without danger even if the wire has rusted out. The trees should, however, be examined and braced again if necessary.

The narrow crotch, which is causing a split in the trunk below, as effectively braced with large screw eyes and wire (upper left of photograph). In the right center are two living braces, made by grafting in the upper ends of water sprouts. In making such a graft, the bark on the water sprout is cut away where it is to touch the main limb. A corresponding part of the bark on the main limb is cut away, and the end of the water sprout is fastened in place with a small ware nail. The cambium of the main limb and of the water sprout must fit closely together. The wounds are then covered with grafting wax

If the need for bracing is not immediate, it is possible to accomplish the desired result by means of a living brace. This may be formed by taking a sucker shoot from the lateral to be supported, and fastening it into the upright in the same way as in making a bridge graft, as shown in figure 13. In most cases this shoot will unite readily with the main trunk and will grow larger from year to year. Care should be taken to keep all shoots from growing on the brace at first, for if a shoot is allowed to grow, it is probable that the upper end of the brace will die. The same result can be obtained by taking a slender branch from each of the limbs to be tied together, and twisting them tightly around each other. As these increase in size, they will grow into a single limb which will act as an effective brace. Union of the two branches can be hastened by scraping the bark away at the points of contact, so that the cambium layers can unite. It is frequently necessary to thin the growth above such a brace by pruning, to keep the brace from being shaded out.

In the case of trees that have had the bark of the crotch killed by winter injury, special treatment is often needed. The problem here is threefold first, to keep the tree from breaking down; secondly, to keep the exposed, dead wood from rotting; and thirdly, to get the dead area to heal over from the edges. To accomplish the first, the tree should be braced as outlined above. The best treatment for the wound is probably to let it alone until the end of the growing season following the injury. This gives the cambium around the dead area an opportunity to start growth and so mark the limits of the wound. The old dead bark should then be removed, and the exposed wood covered with a good white-lead paint or with gas tar. In case the latter is used, it is better not to smear it on the edges of the callus growing over the wound, as some tar is injurious. In cases in which the crotch has partly split down so that a crack or crevice is formed, it may be advantageous to pack or calk this with oakum or rags soaked in tar, to keep out the water and give a surface over which new callus tissue can grow.

This method is adapted to trees of this type only


The ability of a crotch to support a load depends in large part upon the actual strength of the crotch tissue itself, as well as upon the factors previously explained. In this regard, also, there is considerable varietal difference, although the actual arrangement of the cells seems to depend on the crotch angle and on the relative sizes of the two branches more than on variety. The cellular structure of the crotch is markedly different from that of normal wood as seen in regions away from the crotch. This normal structure is here shown in three photomicrographs of apple wood (Plate I, 1, 2, and 3), an understanding of which is necessary before comparison with microscopic crotch structure can be made.

In Plate I, 1, the normal structure of apple wood is shown in cross section. The main features to be observed are the large water-conducting elements the vessels, appearing as open spaces surrounded by the smaller and thicker-walled wood fibers and the wood parenchyma. The wood fibers make up the larger part of the wood, and can be distinguished by their small size and thick walls. The wood parenchyma appears as scattered cells distinguished from the wood fibers by their thinner walls and dark-staining contents. The above-named types of cells extend vertically up and down the tree, and are therefore shown in the figure in cross section. The vessels are the channels through which water, and possibly mineral nutrients, are conducted from the roots to the leaves. The wood fibers serve chiefly for mechanical support, and to them the strength of the wood is largely due. Both vessels and fibers are cells without living contents. The wood parenchyma cells, on the other hand, are living, and each is filled with a protoplast. They function in starch storage and possibly in conduction. The lines of cells extending across the figure from top to bottom are the medullary rays. Each ray forms a band of tissue extending horizontally from some point in the phloem through the cambium and into the wood. A small proportion of them may extend clear through the wood to the pith. The cells are living, and, in the figure, are shown filled with protoplasm, starch, and other substances. The rays function in food storage and in lateral conduction. Water and possibly mineral nutrients, brought up from the roots through the vessels, may pass along the rays out to the phloem, and carbohydrates and other material may pass from the phloem through the rays into the wood. In such normal structure, it is seen that the cells are arranged with more or less regularity; the vessels, the wood fibers, and the wood parenchyma extending vertically up and down the trunk more or less parallel with each other, and the medullary rays extending at right angles to them.

In Plate I, 2, is shown a radial section as taken parallel to the medullary rays and the vessels. Here the main water-conducting elements extend in a vertical direction, while the medullary rays, showing their broadest dimension, extend horizontally. Again the regularity of arrangement is apparent. The larger proportion of the wood is made up of thick-walled wood fibers and vessels, with a smaller amount of ray tissue, and scattered vertical lines of wood parenchyma.

The tangential section (Plate I, 3) is taken at right angles to the medullary rays, and parallel to the vessels, the wood fibers, and the lines of wood parenchyma, all of which can be seen extending vertically across the figure, curving around the medullary rays. Such normal wood structure resists transverse breaking, but splits readily in a longitudinal direction.

The ease of longitudinal splitting is one of the chief underlying causes for weakness in crotch structure. With any given crotch, the tissue of both branches runs along parallel up to the point of insertion, and then divides to form a crotch. This means that were there no special tissue formed in the crotch angle, the strain of the load would be applied in a way that would split the wood most easily. The crotch tissue that is formed is, in fact, more resistant to splitting than is normal wood structure. It is not, however, as strong as normal wood in transverse breakage.

In the case of the hardwoods used for purposes where strength is the important consideration, it has been found that, as a rule, strength increases directly with the specific gravity. The woody material making up the cell walls is of about the same density and strength in all cases, differences in strength being due to differences in the thickness of cell walls and in the amount of air space. In order to determine the specific gravity of the apple-crotch tissue as compared with ordinary tissue, tests were made on samples taken from the heartwood of a mature apple tree, with the following results:

  Oven-dry weight
Water displaced
(cubic centimeters)
Crotch tissue 52.25 77 0.67
Normal heartwood 43.35 74 0.58

The greater density of the crotch tissue indicates that its weakness is due, not to lack of woody material, but rather to the arrangement of the cells in crotch structure.

The structure within the crotch is markedly different from the normal structure described above, particularly in the region where the tissue of the two branches lies adjacent. In this location, instead of the usual wood elements such as vessels, fibers, and wood parenchyma in normal proportion, the tissue is made up, in large part, of parenchyma and large ray cells which are packed with starch in the dormant season. What few vessels are present are small and very much distorted, as are the wood fibers. The medullary rays are relatively larger and more irregular in shape than in normal structure, and are composed of larger cells.

A photomicrograph of a section taken at right angles to both the plane and the axis of the crotch is shown in Plate I, . It shows the abundant parenchyma and the medullary ray cells filled with starch. Vessels and wood fibers are few. One vessel only is shown, above the center of the figure. Strands of fibers, cut in oblique section, can be seen extending diagonally across the figure from side to side. The cells for the most part are thinner-walled in proportion to diameter than would be the case with normal fibers. It is through such tissue that breaking usually takes place.

In Plate II, 1, is shown a section taken through the axis perpendicularly to the plane of the crotch along the face of the break when the crotch was split apart. The tissue in the lower half of the figure corresponds to a medullary ray, which in this situation is composed of larger cells than in normal wood. Above the ray, near the center of the figure, is a vessel segment showing several elongated (scalariform) openings in its end wall, and surrounding it are modified vessels, or tracheids. The presence of this type of opening in this situation is interesting from an histological standpoint, as in normal woody tissue of the apple there is but a single round opening in the end wall of each vessel segment, the vessels being of the porous type. In this and in Plate I, 4, it is readily seen that the cells are not arranged in the regular order which is prevalent in normal wood away from the crotch, but are much distorted in form and in direction of extent.

In considering woody tissue of this type, some insight into its structure can be gained by considering its source, namely, the cambium. This tissue, composed of living cells, extends as a sheet over the entire woody cylinder. By division toward the inside, new layers of wood are formed, and toward the outside, new layers of phloem are laid down. In general, it can be said that the cells of the wood and of the bark extend in the same direction and have approximately the same shape as the cambium cells which produced them. There are certain exceptions to this statement, such as the formation of large vessels from small cambium cells, but it is correct in the main. Thus, in Plate II, 2, is shown the normal cambium as seen in longitudinal section parallel to the surface of the woody cylinder. The elongate cells extending vertically are the ones which give rise to the vessels, the wood fibers, and the wood parenchyma in the woody cylinder, and the groups of smaller cells are those which give rise to the medullary rays. Comparing this with the tangential section of wood (Plate I, ), it is seen that the general size and arrangement of cells is the same, with the exception of the wide vessels and the transversely divided wood parenchyma, which have been modified subsequently to their division from the cambium.


1, Transverse section. 2, Radial Section. 3, Tangential section. 4, Section through an apple-tree crotch, taken perpendicularly to both the plane and the axis of the crotch



1, Section of an apple crotch, taken through its axis perpendicularly to the plane of the crotch
2, Tangential section of the cambium of the pear. The cambium of the apple is practically identical in shape and size of cells
3, Section of an apple crotch taken diagonally through the cambium where it joins the wood in the crotch angle at the union of the tissues of the two arms. On the left, the lighter-stained wood; on the right, the cambium.
4, Same as no. 3. but at higher magnification. The lighter-staining wood cells on the left have the same conformation as the cambium cells on the right

In studying the cambium of the crotch, a similar relationship between the arrangement of cambium cells and that of the cells of the woody tissue of the crotch is seen to exist. In Plate II, 3, is shown a section in the cambium region of the crotch angle along the line where the tissues of the two branches meet. The section is taken tangentially to the curved surface of the crotch, and shows on the left the lighter-stained woody tissue, and on the right the darker-stained cambium tissue. A section taken in a similar location is shown at higher magnification in Plate II, 4. From these it can be readily observed that the woody cells are somewhat similar in shape and size to the cambium cells that formed them. Farther away from the cambium in the wood, a somewhat greater differentiation of wood elements is found, as indicated in Plate I, , which is somewhat comparable. The last-named figure, however, shows a higher magnification and so may be confusing.

As the annual layers of woody tissue are laid down over the inside of the crotch, it is evident that there must be a distinct shortening of the cambium and the tissues outside it in this area. For example, in a crotch such as that shown in figure 10, the cambium once extended over the annual ring indicated by A-B-C. In its present position (D-E-F), the length of the cambium intersecting the plane of the crotch is little more than half the length of its original intersection. During the increase in size, therefore, there must be a continual crowding of the cambium and bark in the crotch angle. The results of this action are readily seen in crotches that have not had the dead bark rubbed off for some time. The bark in the crotch angle is crowded up into a ridge where the tissue of the two arms of the crotch comes together, and is shed from the tree, and the living and more plastic bark is forced up into convolutions along with the cambium. Just what happens to the cambium itself it is difficult to determine. There is no evidence that the cambium cells are crushed and sloughed off, as is the case with the dead bark. Sections taken in different planes through the crotch cambium fail to show cambium cells that bear evidence of having been killed by crushing, though usually the cells are much distorted, as shown in Plate II, 3 and . A possible explanation of what takes place is that during the growing season, when the pressure of lateral growth would be most strongly exerted, the cambium is forced out sidewise from the crotch angle, and serves in part to compensate for the increased surface along the sides of the crotch caused by the increased growth. It is evident that when the crotch angle is filled in with woody tissue, the effect is not only to shorten the line of intersection of the crotch cambium with the plane of the crotch, as before explained, but also proportionally to increase the extent of the cambium on the sides of the crotch. The nature of this increase can possibly be made more clear by again referring to figure 10. Suppose G to be a point on the surface of the bark on the side of the crotch horizontally opposite the point of insertion. Then when the crotch was but a year old and the point E was in the position of B, the distance from G to E would have been small, possibly only half an inch. As growth takes place, however, the crotch is filled in and the diameter of the branches is increased, with the result that E in the crotch angle moves upward and G horizontally outward, thus greatly increasing the distance between them. This increase may be compensated for in part by the cambium forced out of the crotch angle, and in part by normal cambial division. At the time of rapid cambial growth there is a region of considerable thickness around the cambium which is more or less plastic, since during the period of rapid division there are always many layers of cells in both xylem and phloem that are unlignified and could be easily displaced and distorted without injury.

During this growing season the cambium region over the crotch might be likened to the region between the skin and the muscles of an animal. The skin is attached at all points, yet there is allowance for a certain amount of slipping underneath. During any one growing season, a comparatively slight slipping of this sort would allow adjustment of the cambium to the shortening of the intersection of the plane of the crotch and the lengthening of the area on the sides of the crotch.

It should be understood that, in an adjustment of this sort, it is not necessary that cell walls actually slide past walls of adjacent cells, for this would involve the splitting of the middle lamellae which the adjacent cells have in common, and disorganization of the plasmodesma and pit connections. When the cells in the cambium region are in a plastic, unlignified condition, it is entirely practicable for adjustments of position to take place by the change in form of the cells, just as a gelatinous mass of frogs' eggs changes its shape without breaking apart.

From a study of the direction of extent of the cambium cells, the wood fibers and vessels, and the elements of the phloem, it is evident that the main lines of conduction up and down the stem are fairly distinct for the two branches of a crotch. The tissue of the two branches is adjacent but is almost entirely separate. This means that along the line where the two limbs of the crotch come together, there is a region which belongs to neither branch, but is a sort of parting of the ways, or dead area, as far as conduction is concerned. The conduction which does take place through this irregular tissue would be somewhat comparable to lateral conduction, and this, while it does take place, is not considered to be as rapid as vertical conduction. The inefficiency of lateral conduction is indicated by the results of various experiments in which only one side of a tree has been fertilized and only that side has shown any effects of the treatment.

It might well be that the relation of this more or less isolated crotch area to conduction may in part account for the susceptibility of the crotch region to winterkilling. Poor conduction may result in delayed maturity, and thus in greater likelihood of winter injury. Analyses of crotch tissue show that, in so far as nitrogen is concerned, there is more of this element in both wood and bark on the inside of the crotch than in the tissue of the limb on the side away from the crotch. Microscopic examination of tissue taken in the early winter also indicates that starch storage tissue is particularly abundant in the crotch region. The presence of an excess of nitrogen might in part account for the susceptibility of the tissue to winterkilling, as it is reasonably well established that an excess of nitrogen delays maturity. The results of the analyses are given in table 2.


  Green weight
of sample
Per cent of
water in
green weight
Milligrams of
per gram
of dry weight
Milligrams of
per gram
of water
  Inside Outside Inside Outside Inside Outside Inside Outside Inside Outside Inside Outside Inside Outside
Wood (crotch no.)                          
4 4.744 6.787 2.769 3.563 1.975 3.224 42.6 47.5 5.46 8.12 2.97 2.28 2.76 2.52
5 1,451 3.365 0.860 1.803 0.591 1.562 40.7 46.4 2.10 4.69 2.44 2.60 3.55 3.00
9 6.573 6.682 3.836 4.004 2.737 2.678 41.6 40.2 9.66 3.78 2.52 0.94 3.53 1.41
10a 3.439 5.720 2.061 3.360 1.378 2.560 40.1 44.7 5.46 2.76 2.65 0.82 3.96 1.08
10b 3,252 3.739 1.900 2.300 1.352 1.439 41.6 38.5 6.72 4.20 3.54 1.83 4.97 2.92
10c 3.750 3.952 2.628 2.314 1.132 1.637 30.2 41.4 5.95 .60 2,27 2.42 5,36 3.42
11 4.215 4.103 2.616 2.517 1.599 1.586 37.9 38.6 4.27 4.55 2.63 1.81 2.67 3.87
Average             39.2 42.5     2.43 1.81 3.81 2.46
              Odds 39-1 †     Odds 27-1 Odds 111-1
Bark (crotch no.)                          
4 2.832 3.062 1.341 1.392 2.491 2.672 52.6 54.6 22,96 11.76 9,66 8.45 8.69 7.04
5 2.364 1.460 0.650 0.651 0.714 0.899 52.3 55.4 6.02 4.90 9.26 7.53 8.43 6.06
9 3.627 4.198 2.746 2.937 1.881 2.261 51.9 53.8 15.47 16.10 8.86 8,32 8.22 7.12
10a 3.497 2.46! 1.659 1.246 1.838 2.315 52.5 53.4 15.96 9.24 9.62 8.06 8.68 7.03
10b   1.6011 2.092 0.754   0.847   52.9 20.70 6.65 9.89 8,83   7.85
II 2.908 3.160 2.529 1.628 2.379 2.532 47.4 48.5 11.90 22.50 7.78 7.68 8.63 8.16
Average             51.3 53.2     9.18 8.24 8.53 7.22
              Odds 199-1     Odds 49-1 Odds 199-1
* These analyses were made in the Plant Physiology Laboratories at Cornell University, in cooperation with Dr. O. P. Curtis.
The probable error, calculated by Student's method, indicates that the chances are 39 to I that the percentage of water will be greater in the wood outside the crotch than in the inside. Odds of 30 to 1 are usually considered to indicate a significant difference.

The presence of moisture and nitrogen in excess in the crotch region might be taken as an argument against the idea that there is poor conduction in that region. On the other hand, it may be considered to support this view, in that the region, being a more or less dead area with relation to conduction, receives these materials faster than they are carried away. In ordinary tissue, conduction goes on in comparatively straight lines, whereas in the crotch there is a condition that might be considered analogous to a whirlpool or a backwater, in which the lines of conduction are confused and indefinite and where material of various sorts might accumulate.


The strength of apple-tree crotches varies directly with the width of the angle between the crotch arms, other factors being equal.

Crotches with equal arms split much more easily than those with arms of unequal size, other factors being equal.

More than two scaffold limbs coming from the trunk at nearly the same point may produce a weak union because the crowded condition prevents the wood of the main trunk from supporting the side branches. The wood of one branch merely lies adjacent to that of the branches on either side, instead of being embedded in the main trunk.

The critical age at which crotches break is just as the tree is coming into heavy bearing. The crotches of older trees tend to become strong due to an equalization of growth and the consequent firm knitting of the crotch tissue.

Most defects of crotch structure can be remedied or prevented by pruning while the tree is young. With older trees, bracing is advisable, and in case of winter injury to the crotches, this treatment is necessary if the trees are to be saved.

Histologically, crotch tissue differs from the normal in having more parenchyma, larger medullary rays, and fewer wood fibers and vessels. All of these tissues are very much distorted in their arrangement as compared with the normal tissue, a distortion found also in the cambium itself.

There is a more or less definite line of cleavage where the tissues of the two branches of a crotch lie adjacent. The lines of conduction for the two limbs of a crotch are separate and distinct.

Tissues of both the bark and the wood of a crotch have, by analysis, a higher nitrogen content than corresponding tissue outside the crotch. This high nitrogen content may have something to do with delayed maturity, and hence with susceptibility to winter injury, of tissues in the crotch region.