[ 448 ]

THE ANATOMY OF THE SYMPATHETHIC TRUNKS IN MAN BY MARTIN WRETE Histological Department, The University of Uppsala, Sweden

INTRODUCTION Even a cursory study of the anatomical descriptions of the cervical parts of the sympathetic trunks given in modern text-books or articles discloses that, now as earlier, great confusion exists with respect to terminology. This applies even to monographs and more specialized presentations. The primary cause of this confusion is the very marked variability of the trunks in the region, which gives wide scope for arbitrary interpretations of the arrangement; some uncertainty about the terminology and notation of other parts of the trunks also persists. It is true that the terms to be used for the sympathetic nervous system were fixed by the International Anatomical Nomenclature Committee (Nomina Anatomica, Paris, 1955). This does not, however, prevent some of the individual terms being used to denote different anatomical units, and for practical reasons (such as limiting printing costs) comprehensive explanations could not always be given in the annota- tions to the Parisian Nomina Anatomica. As one of the three members of the Sub- Committee responsible for the nomenclature of the peripheral nervous system, I wish to define more exactly my views on the terminology adopted for the sympathetic trunks. I also take this opportunity of revising a few terms I used in certain papers published some twenty years ago. In Nomina Anatomica the term truncus sympathicus is followed by the names of its ganglia, ganglia trunci sympathici, as well as of its connecting rami interganglio- nares. But, also under the heading ganglia trunci sympathici, the term ganglia intermedia is used to denote ganglia on the rami communicantes and certain ganglia on the trunks in the rami interganglionares between the other ganglia-namely the cervicale superius, ganglion cervicale medium, ganglion cervicothoracicum (s. stellatum), ganglia thoracica, ganglia lumbalia, ganglia sacralia, and . The term ganglion vertebrate is listed under the heading ganglion cervicale medium, without further commentary. The main object of the present paper is to define those anatomical units to which the above terms should be applied.

CERVICAL REGION The cervical parts of the sympathetic trunks differ essentially from the other parts, because their segmentation has become so obliterated, owing to fusion and division of the segmental ganglia, that it has been necessary to give each ganglion a special name. In many text-books it is stated that the individual corre- spond to certain fixed segments, as evidenced by their macroscopically demonstrable connexions with certain spinal through the communicating rami. In the The anatomy of the sympathetic trunks in man 449 cervical region it is not possible to analyse the connexions between the spinal nerves and the trunk solely by macroscopic dissection (Wrete, 1934 a) because the con- nexions consist partly of spinal- branches to the prevertebral muscles, which are joined to the grey communicating rami in a highly complicated way (Fig. 1). However, by microscopic studies of embryos and foetuses (Wrete, 1934b) it was possible to make a segmental analysis of the cervical sympathetic trunks (Fig. 2).

Va

6 t

b d

a V

Va

* -~~~~~~~~Stg

Fig. 1. Human foetus, 39*6 mm. Profile reconstruction from the left half of the body, seen from the median plane. C5, C6, etc., parts of cervical spinal nerves; Va, ; Vg, vertebral ganglion; Str, ; Sig, ; a, grey communicating ramus; b, spinal-nerve branch to prevertebral muscles (the arrows mark the terminal branches of the spinal-nerve branches running towards the muscles); c and d, junction of a and b (the broken line denotes part of a muscle branch which follows a grey communicating ramus); Img, intermediate ganglion on this communicating ramus.

This was done by investigating the embryonic segmental parietal arteries, and par- ticularly the grey communicating rami accompanying them, but unfortunately little attention has been paid to these observations. These communicating rami, which are present only in the lower part of the cervical region and the uppermost part of the thoracic region, are characterized by their division into two branches, one to the segmental artery corresponding to the , and one to the cranial artery in immediate succession; I therefore introduced the term 'rami communi- cantes grisei bipartiti' for these rami. From a linguistic point of view, the term is 29 Anat. 93 450 M. Wrete not particularly appropriate, and it does not seem to have become generally accepted. (I have, in fact, seen it only in two text-books.) Consequently, in view of the fork-like mode of branching of the nerves in question, I now suggest that they be named 'rami communicantes grisei bifurcati'. Throughout the literature the term superior cervical ganglion is used to denote the most cranial of the ganglia on the cervical part of the sympathetic trunk,

i1 Va

Fig. 2. Schema showing the course of the rami communicantes grisei bifurcati in man. Left sympathetic trunk seen from the lateral view. The transverse process and ribs are sawn through, and the lateral parts of the cut surface are removed (broken lines indicate the cut surfaces). Pc, costal process of 7th cervical ; PI, lateral process of same vertebra; Co, neck of first rib; Pt, transverse process of first thoracic vertebra; Va, vertebral artery; C1 C4, etc., cervical spinal nerves; Th,, etc., thoracic spinal nerves; Mcg, ; Scg, superior cervical ganglion; Stg, stellate ganglion; Vg, vertebral ganglion; Tg2, etc., ; 1mg, intermediate ganglion; a, b, deep rami communicantes; c, vertebral nerve; d, e, f, g, caudal rami communicantes grisei bifurcati. The anatomy of the sympathetic trunks in man 451 irrespective of variations in shape, size and position. There is no unanimity about these variations. In most text-books the term stellate ganglion is used to denote that ganglion formed by fusion of the lowest ganglion on the cervical trunk and one or more (usually one to four) of the adjacent thoracic ganglia. These most cranial thoracic ganglia can be identified with the help of the rami communicantes grisei bifurcati. The upper part of the stellate ganglion was formerly known as the inferior cervical ganglion. This term has sometimes given rise to confusion, since it has also been used, incorrectly, for a ganglion cranial to the . The fact that the term has been discarded in Nomina Anatomica is justified by the fact that the in- ferior cervical ganglion appears only exceptionally as an independent formation. The part of the stellate ganglion which corresponds to it is characterized by its position directly caudal to the subclavian (and the root of the vertebral) artery, and by the fact that it gives off two rami communicantes grisei bifurcati (Wrete, 1934b); the cranial ramus, which is developed along the 7th cervical segmental artery and is named the vertebral nerve (nervus vertebralis), joins the spinal nerves C7 and C6; the caudal ramus, developed along the 8th cervical artery, joins C8 and C7. It is not unusual to find a constriction on the stellate ganglion, marking the borderline between the parts representing the inferior cervical ganglion and the 1st thoracic ganglion. The suggestion put forward by Lazorthes & Cassan (1939), and adopted by Guerrier (1944), that the term cervicothoracic or stellate ganglion should include the vertebral ganglion cranial to the origin of the vertebral artery is not warranted, as Mitchell (1953) has justifiably stressed. In my opinion this also applies to a similar suggestion put forward by Axford (1927-8) and Woollard & Norrish (1933), i.e. that it should also include this ganglion, although they failed to differentiate the verte- bral from the middle cervical ganglion. The greatest divergences in the nomenclature are encountered with respect to the part of the sympathetic trunk between the superior cervical ganglion and the stellate ganglion. In most descriptions only one ganglion is mentioned, namely the middle cervical ganglion. Some authors state it is situated relatively high up and others place it relatively low down, close to the root of the vertebral artery. Axford (1927-8) refers to high and low middle cervical ganglia. When two ganglia were present, they were described by van den Broek (1908) as a middle cervical ganglion split into two parts. In an earlier publication (Wrete, 1934 b) I also used this less appropriate nomenclature. Mannu (1914) denoted all ganglia between the superior cervical ganglion and the inferior cervical ganglion as intermediate ganglia; he distinguished as particularly characteristic a superior one, the ganglion, and an inferior one, close to the subclavian artery, the vertebral or subclavian ganglion. In some modern text- books of anatomy an attempt has been made to clarify the terminology by denoting the superior, generally larger ganglion as the middle cervical ganglion, and the inferior one as the intermediate cervical ganglion (e.g. Jonnesco, 1923; Hovelacque, 1927; Kuntz, 1946; Brodal, 1948; White & Srnithwick, 1952). Matsui (1925-6) has used the term intermediate ganglion as a synonym of middle cervical ganglion. On the basis of a study of the cervical sympathetic trunks in 120 foetuses and 29-2 452 M. Wrete newborn infants, Laubmann (1931) set up a schema with five basic types (Fig. 3), some of which are stated to be more and others less common. Since there is reason to presume a postnatal reshaping of the cervical sympathetic trunks (Wrete, 1934 b), the incidence figures given by Laubmann may not be fully applicable to adults, nor does this schema cover all the variants which may occur. Despite these drawbacks, Laubmann's schema is of great value. In all the main types and subtypes the lower of the two interjacent ganglia, lying slightly cranial to the origin of the verte- bral artery from the subclavian artery, is indicated. I also found this ganglion to occur with great regularity in extensive macroscopic and microscopic studies of the cervical sympathetic trunks and their segmental branches to the spinal nerves in my series of 41 embryos, foetuses, newborn infants and adults. Laubmann suggested the term vertebral ganglion for this structure, and middle cervical ganglion for the larger ganglion lying more cranially. The use of these two terms, withwhich Iconcur, has the great advantage over the use of only one term in that it recalls the fact that segmentally differing parts of the trunk are involved. AsMitchell (1953) has correctly emphasized in his excellent monograph on the , the term vertebral ganglion is preferable to intermediate ganglion, used by a number ofmodern authors (vide supra), since the ganglion 'is related to and helps to supply the verte- bral artery, and because the term intermediate ganglion is now invariably applied to ganglia on the rami communicantes or ventral nerve roots'. Mitchell's view is based on dissections of about 100 specimens. Substitution of the term vertebral ganglion for intermediate ganglion allows the latter to be reserved for the small ganglion, or ganglia, sometimes present between the ganglia listed, both in the cervical and other parts of the sympathetic trunk (Wrete, 1935, 1941 a, b, 1943, 1951). To understand the anatomical variations of the cervical sympathetic trunks it is necessary to study their ontogenesis. I shall, therefore, give a brief account of my earlier studies (Wrete, 1934 b), since they illustrate particularly well certain problems ofnomenclature. In describing the results I shall use the terminology suggested above, and thus distinguish between middle cervical ganglion and vertebral ganglion. In a 10 mm. human embryo, where the cervical sympathetic trunk appears as a homogeneous ganglionated cord, immediately cranial to the root of the vertebral artery, the primordium of the vertebral ganglion appears as a part separated from the other, more cranial part by a constriction. At this stage the rami communicantes grisei bifurcati have started to develop. In a 15-4 mm. embryo the cranial part of the ganglionated cord, which has a well-defined constriction in the middle, has been displaced some distance cranially, and is now separated from the vertebral ganglion by a distinct interganglionic ramus. In a 29 mm. foetus, this ramus has become further elongated. In both the latter stages, the caudal part of the cranial ganglio- nated mass gives rise to the most cranial of the rami communicantes grisei bifurcati, corresponding to the 6th cervical artery, showing that this caudal end consists of the cranial part of the 6th cervical segment of the trunk, whereas the vertebral ganglion comprises the caudal part of the segment. In specimens from later stages of foetal life, as well as in preparations from new- born infants and adults, marked individual variations are seen. In some cases development of the ganglia has been arrested at the aforementioned stage. In The anatomy of the sympathetic trunks in man 453 other cases development has proceeded yet another step, so that the cranial gangliQ- nated mass has been divided into an upper, superior cervical ganglion, and a middle cervical ganglion caudal to it. This produces a state corresponding to the most common type in Laubmann's schema (type I), which comprised about one-third of all the cases in his material. I have named this the basic type (Fig. 4A). Here, the chief characteristic of the middle cervical ganglion is that it gives off one or two rami communicantes grisei bifurcati. The superior cervical ganglion then consists of the four most cranial segments of the trunk, the middle cervical ganglion comprises the 5th and upper part of the 6th, and the vertebral ganglion is formed from the lower part of the 6th.

I A B A B C IV V

Fig. 3. Laubmann's type schema. In other cases the development of the cervical part of the trunk is arrested at a still earlier stage, and the course of the rami communicantes grisei bifurcati shows that the cranial ganglion, generally termed the superior cervical ganglion, is actually the result of fusion of this ganglion and the middle cervical ganglion. Consequently this anomalous ganglion should rightly be denoted as a medio-superior cervical ganglion (ganglion cervicale medio-superius) (see Fig. 4B). This type of sympathetic trunk, corresponding largely to Laubmann's type II, is-as he pointed out-a prinli- tive type. The ganglion is sometimes short and thick, and in such cases the rami communicantes grisei bifurcati are given off from the interganglionic ramus at a vari- able distance caudal to the ganglion (Fig. 4 C): its closest analogue in Laubmann's schema is type IV which comprised about one-fourth of all his cases. In still other cases the course of the rami communicantes grisei bifurcati shows conclusively that the 5th and 6th segments of the trunk have remained in a low position, forming a ganglion which should be denoted as a medio-vertebral cervical ganglion (ganglion cervicale medio-vertebrale) (see Fig. 4D). In this type a superior cervical ganglion is present which corresponds to the ganglion of the basic type with the same name. This anomaly is represented by type III in Laubmann's schema; he found it to be more common than type II, but less common than type IV. It is obviously a misleading simplification to use invariably the term superior cervical ganglion (ganglion cervicale superius) for the most cranial ganglion of the cervical part of the trunk. Moreover, in the basic type and the medio-vertebral type, this most cranial ganglion is not implicated in the sympathetic supply to the vertebral plexus and the arm, whereas the medio-superior ganglion type is involved 454 M. Wrete in this supply. This circumstance should be borne in mind in planning surgical denervation of the blood vessels of the arm. The communicating rami visible in Fig. 4 have long been known as the deep communicating rami (rami communicantes profundi). They are the most cranial of the groups of nerves that I have called rami communicantes grisei bifurcati, situated cranially to the subclavian artery and the root of the vertebral artery. I found them to be absent in only 4 of 55 sides (Wrete, 1934b). Very often only the cranial one was lacking, whereas absence of the caudal one alone was uncommon. These rami

SCg MScg MSCg 5Sg

AC4(C CD a aa

2C / 0_-c a Meg4 b bb

Fig. 4. Type schema set up on the basis of the segmental structure of the cervical sympathetic trunk. C,, C5, C6, cervical spinal nerves; a, b, rami communicates grisei bifurcati; Sua, subclavian artery; Scg, superior cervical ganglion; Meg, middle cervical ganglion; Mscg, medio-superior cervical ganglion; Mvcg, medjo-vertebral cervical ganglion; Stg, stellate gan- glion; Vg, vertebral ganglion. are easily identified by the fact that they pierce the longus colli muscle on their course from the sympathetic trunk to the spinal nerves, which they join close to the vertebral artery. They may be united with branches of the spinal nerves passing to the prevertebral muscles (Wrete, 1934a) but, despite this, they should be identifiable even at operation. Laubmann's type V, which is uncommon and must be regarded as a fairly gross developmental disturbance, probably appears only in association with anomalies in development of the segmental arteries (Wrete, 19.34b). I found that such vascular anomalies could be associated with marked variations in the anatomy of the cervical sympathetic trunk. It is therefore essential, in a study of the normal anatomy of this part of the trunk and of the stellate ganglion, to disregard cases with vascular anomalies. The anatomy of the sympathetic trunks in man 455

THORACIC, LUMBAR AND SACRAL REGIONS At first sight the terminology of the parts of the sympathetic trunk caudal to the cervical region seems to present no difficulties, because the thoracic region may have 12 ganglia, the lumbar region 5, and the sacral region 5, i.e. the same number as the segments. When this applies, the ganglia can simply be denoted as the lst-12th thoracic ganglia, lst-5th lumbar ganglia and lst--5th . In actual fact, however, their number is highly variable. Thus there may be 9 or 13 ganglia in the thoracic region, 1 or 7 in the lumbar region, and 2 or 6 in the sacral region, and then difficulties arise in naming the individual ganglia. If each ganglion were joined by one or several communicating rami to a single spinal nerve, it could be given the name of the relevant nerve, but this is the case only in the upper part of the thoracic region. Attempts have been made to number the ganglia according to the spinal nerve to which each is joined by the communicating rami (van den Broek, 1908), by its white communication ramus (Zuckerman, 1937-8), or by its grey communi- cating ramus (Langley, 1896; Ranson & Billingsley, 1918; Sheehan & Pick, 1942-3). The last-mentioned authors have dealt fairly extensively with the question of terminology in connexion with a study of the communicating rami in Rhesus monkeys. It is more convenient to use the grey instead of the white rami, since, as Sheehan & Pick have pointed out, 'the fibres entering the sympathetic trunk through one white ramus may be distributed to 9-10 successive ganglia and, furthermore, certain ganglia receive two and even three white rami from several spinal nerves'. However, the use of the grey communicating rami is also associated with draw- backs. It is not always possible to distinguish the white and grey rami from each other macroscopically, since they are often mixed, and Sheehan & Pick admitted that histological examination was often necessary. They had to adopt arbitrary rules in order to bring about uniformity in the descriptions of the various dissections and to permit accurate comparison, even if microscopic studies were available. The same principles were applied by Pick & Sheehan (1946) in a later study of the com- municating rami in man. The procedure of Sheehan & Pick is applicable only in the extremely rare situa- tion when all communicating rami are examined histologically to determine whether they are of the white or grey type. Even then difficulties occur. On macroscopic dissection the procedure can scarcely be used, in view of the great difficulty of distinguishing between the two kinds of rami. Souli6 (1901) has gone so far as to state that, in man, this is possible only in exceptional cases; van den Broek (1908) came to the same conclusion in a study of a large number of mammalian species. In fresh material the difference could occasionally be observed, but in preserved material no difference was detectable. In my opinion, it is more appropriate to choose another procedure, which facilitates exact terminology, and which is applicable even when deviations are present from the simple basic pattern of 12 thoracic, 5 lumbar and 5 sacral ganglia. In other words, the terminology should be entirely independent of the communi- cating rami, in view of their highly variable course, and the difficulties in dis- tinguishing between them. On a wax reconstruction, according to Born's method, of the sympathetic trunks 456- M. Wrete in a 10-3 mm. embryo (Wrete, 1930), the parts of the trunk caudal to the cervical region are distinctly segmented, and the ganglia lie in the spaces between the seg- mental, parietal branches of the aorta. These subsequently give rise to the 1st to 11th posterior , the subcostal artery and the , as well as lateral branches of the median sacral artery. Not only in human embryos and foetuses, but also in the newborn and adult, all these arteries cross the sympathetic

Str

Th22 l-T 12 LI 31.

L2 >Lg2

L3Lg3 L4trrniL

L5°m

-g5g Sob.

Fig. 5. Schema showing the thoracic, lumbar and sacral regions of the sympathetic trunk. Th12-SI, segmental parietal arteries; Str, sympathetic trunk; Th12, L1-L., S,, segmental parietal arteries; Tg1,2, Lg2-Lg,, Sg,, thoracal lumbar and sacral ganglia of the sympathetic trunk. trunk and lie close to it. This permits convenient, exact division of the trunk within its segments. Thus the term 4th thoracic ganglion can be used to denote the ganglion lying in that part of the trunk between the 4th and 5th intercostal arteries, the 12th thoracic ganglion is the ganglion lying between the subcostal and the 1st lumbar arteries, and so on. Finally, it is reiterated that rami communicantes grisei bifurcated are present in the The anatomy of the sympathetic trunks in man 457 three or four uppermost thoracic segments (Wrete, 1934 a) as they have developed along the segmental arteries. They can therefore be used for an analysis of the stellate ganglion, to determine the number of thoracic ganglia in it. Application of the terminology I have suggested is illustrated schematically in Fig. 5 which shows some segmental arteries (Th12-SA), as well as a corresponding length of the sympathetic trunk. The uppermost of the ganglia is the 12th thoracic ganglion. The 1st lumbar ganglion is lacking, and the 2nd is split into two parts. The 3rd and 4th are fused into a common ganglion. The caudal one next in order should be termed the 5th lumbar ganglion, since only a minor part of it projects above the 5th segmental artery; it is not necessary in such a case to describe the upper part of the ganglion as a caudally situated part of the 4th lumbar ganglion. Obviously a segmental artery may occasionally have an anomalous course and, in this event, it should be possible to determine approximately how it would have run normally and to number the relevant ganglion accordingly.

SUMMARY The individual ganglia on the sympathetic trunks in man have the following characteristics. The cervical part of the trunk generally has four typical ganglia, the superior cervical ganglion, the middle cervical ganglion, the vertebral ganglion, and the stellate ganglion of which the lower part belongs to the thoracic part of the trunk. The middle cervical ganglion (ganglion cervicale medium) gives off a ramus com- municans griseus bifurcatus to the 6th and 5th cervical nerves, and sometimes one to the 5th and 4th as well. The ganglion cranial to it is the superior cervical ganglion (ganglion cervicale superius). These two ganglia are often fused into what should be denoted as a medio-superior cervical ganglion (ganglion cervicale medio-superius). The vertebral ganglion (ganglion vertebrate) is relatively small, and lies directly cranial to the subclavian and the root of the vertebral artery. It is not infrequently fused with the middle cervical ganglion, thus forming a medio-vertebral cervical ganglion (ganglion cervicale medio-vertebrale). The medio-superior and medio-vertebral ganglia are common variants. The upper part of the stellate ganglion (ganglion stellatum s. cervicothoracicum) only exceptionally appears as an independent formation; it was formerly known as the inferior cervical ganglion (ganglion cervicale inferius). Characteristically it is joined respectively to the 8th and 7th cervical nerves and to the 7th and 6th cervical nerves through the caudal and cranial rami of a griseus bifur- catus. The lower part of the stellate ganglion usually consists of the 1st thoracic ganglion, which can be identified by its connexion with Th, and C8 through a ramus communicans griseus bifurcatus. The 2nd and 3rd thoracic ganglia, and even the 4th, may form part of the stellate ganglion. These ganglia can be identified by the communicating rami of a similar kind given off by them, whereas such rami are lacking in the segments caudal to them. The above statements about the cervical sympathetic trunk ganglia are valid only when the cervical segmental arteries are normally developed. Below the stellate ganglion, each ganglion should be named after the space between the segmental, parietal arteries (e.g. intercostal arteries, lumbar arteries) 458 M. Wrete in which it is situated. For example, the ganglion lying between the 3rd and 4th lumbar arteries is denoted as the 3rd lumbar ganglion. The term intermediate ganglion (ganglion intermedium) should be reserved for the small, inconstant ganglia situated in the internodal rami (rami internodiales) between the ganglia listed above, at any site in the trunk, as well as in the com- municating rami or in the main trunks or roots of the spinal nerves.

REFERENCES AXFORD, M. (1927-8). Some observations on the cervical sympathetic in man. J. Anat., Lond., 62, 301-318. BRODAL, A. (1948). Neurological Anatomy in Relation to Clinical Medicine, 496 pp. Oxford: University Press. VAN DEN BROEK, A. J. P. (1908). Untersuchungen uber den Bau des sympathischen Nerven- systems der Saugetiere. Morph. Jb. Part I in 37, 202-288; Part II in 38, 532-589. GUERRIER, Y. (1944). Le sympathique cervical, 245 pp. Montpellier: Imprem. de la Charitk. HovELAcQUE, A. (1927). Anatomie des nerfs craniens et rachidiens et le systeme grand sympathique, 873 pp. Paris: G. Doin. JONNESCO, T. (1923). Le sympathique cervico-thoracique, 88 pp. Paris: Masson. KUNTZ, A. (1946). The Autonomic Nervous System, 3rd ed. 687 pp. London: Baillioere, Tindall and Cox. LANGLEY, J. N. (1896). Observations on the medullated fibres of the sympathetic system and chiefly on those of the gray rami communicantes. J. Physiol. 20, 55-76. LAUBMANN, W. (1931). Anatomische Studie uber den Halssympathicus des Menschen. Z. ges. Anat. 1. Z. Anat. EntwGesch. 96, 787-805. LAZORTHES, G. & CASsAN, L. (1939). Essai de schematisation des 6toilt6 et intermediaire (ganglion cervico-thoracique) d'apres 80 dissections chez l'adulte et chez le foetus. Bull. Ass. anat. 49, 193-210. MANNU, A. (1914). Richerche anatomo-comparative sul Simpatico cervicale nei Mammiferi. Int. Mschr. Anat. Physiol. 30, 49-168. MATSUI, Y. (1925-6). Beitrage zur Kenntnis der Anatomie des sympathischen Nervensystems. IV. Mitt.: Truncus sympathicus cervicalis bei Neugeborenen und Foten. Acta Sch. med. Univ. Kioto, 8, 397-424. MITCHELL, G. A. G. (1953). Anatomy of the Autonomic Nervous System, 356 pp. Edinburgh and London: Livingstone. PICK, J. & SHEEHAN, D. (1946). Sympathetic rami in man. J. Anat., Lond., 80, 12-20. RANSON, S. W. & BILLINGSLEY, P. R. (1918). The superior cervical ganglion and the cervical portion of the sympathetic trunk. The thoracic truncus sympathicus, rami communicantes and in the cat. An experimental analysis of the sympathetic trunk and greater splanchnic nerve in the cat. J. comp. Neurol. 29, 313-358, 405-439, 441-456. SHEEHAN, D. & PICK, J. (1942-3). The rami communicantes in the rhesus monkey. J. Anat., Lond., 77, 125-139. SOULIE, A. (1901). Systeme nerveux grand sympatbique, in Traits d'anatomie humaine, public par P. Poirier et A. Charpy, T. 3, F. 3, pp. 625-1124. Paris: Masson. WHITE, J. C. & SMITHWICK, R. H. (1952), The Autonomic Nervous System, 3rd ed. 569 pp. New York: MacMillan. WOOLLARD, H. H. & NORRISH, R. E. (1933). The anatomy of the peripheral sympathetic nervous system. Brit. J. Surg. 21, 83-103. WRETE, M. (1930). Morphogenetische und anatomische Untersuchungen uber die Rami communi- cantes der Spinalnerven beim Menschen. Inaug.-Diss., Uppsala. Lakarefor. forhandl. N.F. 35, 221-380. WRETE, M. (1934a). Uber die Verbindungen der Cervikalnerven mit den sympathischen Grenz- strangen beim Menschen. Z. mikr.-anat. Forsch. 35, 425-456. WRETE, M. (1934b). Untersuchungen uber die Sympathicusversorgung des Plexus brachialis und die Halsgrenzstrange beim Menschen. Z. mikr.-anat. Forsch. 40, 1-94. WRETE, M. (1935). Die Entwicklung der intermediaren Ganglien beim Menschen. Morph. Jb. 75, 229-268. The anatomy of the sympathetic trunks in man 459 WRETE, M. (1941 a). Die Entwicklung und Topographie der intermediaren vegetativen Ganglien bei gewissen Versuchstieren. Z. mikr.-anat. Forsch. 49, 503-515. WRETE, M. (1941 b). Beitrage zur Kenntnis der Anatomie des Halssympathicus beim Kaninchen. Z. mikr.-anat. Forsch. 49, 317-332. WRETE, M. (1943). Die intermediaren vegetativen Ganglien der Lumbalregion beim Menschen. Z. mikr.-anat. Forsch. 53, 122-133. WRETE, M. (1951). Ganglia of rami communicantes in man and mammals particularly monkey. Acta anat. 13, 329-336. ZUCKERMAN, S. (1937-8). Observations on the autonomic system and on vertebral and neural segmentation in monkeys. Trans. Zool. Soc. Lond. 23, 315-378.