Though the various cellular and biochemical
constituents of coagulating blood have been subjected to
vigorous and intensive examination for over a century,
the erythrocyte is dismissed in relatively few words by
the specialist works of reference on the subject (1-3).
This is anomalous, since the erythrocyte in many respects
is the foremost constituent of the coagulation process,
and it comprises the overwhelming bulk of the contracted
clot itself (4). It is important to remember that the
erythrocyte may contain materials of significance in the
biochemistry of the process (5). Further, it is evident
that the open network of a plasma clot cannot provide a
fluid-tight haemostatic seal, and therefore the presence
of the solid mass of erythrocytes is clearly the origin
of the haemostatic properties of the clot. Thus there
would appear to be functional, biochemical and structural
reasons why the erythrocyte should be considered not
merely as a histological constituent of the clot, but as
an essential element of the coagulation mechanism and the
key to its successful function. It is known that only
0.15% of the clot is composed, mass for mass, of fibrin.
Thus for this structural integrity to be maintained,
efficient deployment of the protein itself is paramount;
the length of a fibrin thread of mean cross-sectional
area 0.5 Ám available on average for each erythrocyte
(assuming a calculated volume of approximately 90 Ám3)
may be shown by simple arithmetic to be approximately
0.26 Ám. Much of the fibrin of a clot may be observed to
be very much thicker than this, and so the allocation of
fibrin to each constituent cell is meagre. Even on this
scale, the length, of thread (l) necessary to
ensnare each erythrocyte would clearly approximate to:
l=3 (0.5c) Ám [where c = maximum circumference
of the erythrocyte = 42 Ám] =63 Ám.
The enmeshment of erythrocytes as described
traditionally (the network . . . entrapped and
retracted around the RBCs6), cannot operate with
such small allocations of fibrin available per
erythrocyte Thus Zweifach (7) and others have spoken of a
gelatinous exudate arresting the
erythrocytes, although its nature was obscure; more
recent research has revealed that the erythrocytes are
arrested in the manner of a balloon on a string, each
suspended cell - penderocyte = literally suspended
cell (9) - being attached to an ultramicroscopic
Blood coagulation is often studied in static in vitro
specimens1. Haemostasis is, by definition, a dynamic
phenomenon and can therefore be studied only in
flowing-blood preparations. Other workers (12) have
described slide-chamber and ring-circuit apparatus which
is useful where conditions of high-velocity flow are
obligatory. In th present work, research into the
problems of studying flow in pseudo-capillaries,
utilising a single drop of blood, has enabled us to
derive a satisfactory technique which dispenses
altogether witt elaborate apparatus and greatly
simplifies the speed and repeatability of the
investigations. The technique now used relies on the
capillarity exerted between a glass slide and a fine
coverslip which produces centrifugal radial flow through
fine capillary-like tubules formed in peripheral ring of
partly-dried blood. Drops of fresh blood are placed on
conventional slides and are allowed to stand for several
minutes At the end of this time the addition of a fine
coverslip induces haemostatic pressure which leads to the
breakdown of the peripheral layer of partly dried,
jelly-like blood. Fine undulating tubules form and in
these blood flow may be observed with ease and rapidity.
At presen we find a waiting period of 7 minutes is
optimum for the observation described; the procedure is
very much simpler than those we have employed earlier and
is immediately adaptable to the concept of
penderocyte test, by which the presence or
absence of these cells may be observed rapidly from a
sample of finger-prick blood.
Possible artefacts induced by the drying appear to be
(a) the phenomena observed using the technique are
identical with those seen using sealed-chamber
(b) the amount of crenation observed is negligible in any
preparation and in 60% it is nil.
The drawtube, extended to 180 mm throughout the
experiments, was fitted with baffles to exclude fogging
by extraneous light scattered from the specimen which was
illuminated by a 750W forcibly cooled projector lamp
unit. The unit was screened to give a blue-predominant
spectrum (since visualisation at this extreme
magnification is a function of the illuminating
wavelength) and also to reduce nfrared which would
otherwise unduly heat the specimen.
Thermocouple sensors indicated that the specimens were
maintained at or near 37C. The use of 15x and l0x
eyepiece attachments enabled a final magnification to be
attained in the region of 5000x
with remarkably little emptiness, an extremely gratifying
Figure 1. Light-ground appearance of the
penderocytes; they are seen as 'pear-shaped'
erythrocytes, their apices pointing upstream.
Erythrocytes may be seen within these haemodynamic
tubules. They flow through the passages easily.
Occasional examples will be observed to lie in the
flowing blood with their apices (see Figure 1) pointing
upstream; these are the penderocytes and they
steadily increase in number as coagulation proceeds (in
7-minute specimens coagulation is already evident). Their
superficial appearance is similar to that of erythrocytes
entrapped in a gelatinous exudate; however,
they may be observed to undergo lateral displacement when
jostled by flowing erythrocytes, and thus describe an arc
about the length of the suspending thread; and their
position and the configuration of the pointed apex varies
with the speed of flow and thus as the elastic thread
undergoes changes in length. Irrigation of the
preparation allows the reversal of flow to be induced,
which unequivocally demonstrates the presence of fine
threads anchoring the erythrocytes in position. During
normal procedures the technique enables flowing blood to
be observed at length, and the entire processes of
occlusion and haemostasis due to the blood elements may
For detailed studies (9) oil-immersion dark-ground
microscopy has been utilised, which has demonstrated the
presence of the threads by direct observation (11). They
are of extreme fineness, of the order of 0.2 Ám, and
thus are markedly more tenuous than those of the normal
threads of fibrin observed within coagulating blood. The
dark-ground apparatus was adjusted optically to give a
working numerical aperture of exactly 1.0 in order to
derive maximum theoretical benefits from the technique.
Figure 2. Dark-ground micrograph showing
the suspensor microfibrils anchoring the erythrocytes
within the fibrinous coagulum in vitro.
The work of Bull (12) describes some aspects of the
relationship between fibrin precipitation and the
occurence of fragmented red blood cells in
microangiopathic haemolytic anaemia, although they do not
refer to an interesting if superficial study of similar
erythrocytes by Bell13. The paper by Bull et al shows
red-cell fragments attached to fibrin strands
although there is no unequivocal evidence to confirm this
In order to ascertain whether the microfibrils observed
in the present study were composed of fibrin, coverslips
were specially prepared by drilling apertures in them,
approximately 0.5 mm in diameter and irrigating through
this saline containing the fibrinolytic enzyme plasmin
(Kabi). The cells were observed to detach and to float
downstream in familiar erythrocytic outline.
Controls irrigated with saline or with serum remained
with the anchoring microfibrils intact. Microchemical
analysis of the fibrils is impracticable in view of their
fine structure; only dark-ground microscopy enables them
to be visualised, though not resolved, by reflected light
(8) and such tenuous structures cannot therefore be
conventionally analysed. However the action of plasmin
tends to confirm that they are composed of fibrin.
In over 100 samples of normal blood examined by this
and earlier techniques, penderocyte formation has been
observed within 6-8 minutes from the commencement of the
procedure, i.e. at the moment of obtaining the blood
sample (whether by finger-prick or intravenously). The
number of cells is at first small but as time proceeds
the proportion increases until the lumen of the tubule
through which the blood is flowing becomes entirely
occluded by masses of adjacent penderocytes. It is clear
that a mechanism of this sort must obtain in normal
haemostasis, since: firstly, a spongy, porous plasma clot
alone is not an effectively fluid-tight plug and
secondly, it is only by the suspension of cells in the
manner of balloons that an open channel filled with
flowing blood can be blocked. Normally blood commences to
coagulate in contact with the trauma site itself, by
which time the blood has already been expressed from the
wound and therefore the precipitation of fibrin within it
is of no avail. Platelet-plug formation clearly cannot
apply to blood loss from vessels much larger than
capillaries. The observations and films that have been
made during the present series of investigations
demonstrates the mechanical significance of the processes
with great clarity.
Suspended erythrocytes (termed in this research
penderocytes) may be observed (10) by a
simple, repeatable technique which demonstrates the
dynamic aspects of the haemostatic mechanism in flowing
blood. Critical dark-ground microscopy at very high
magnifications reveals that the cells are held by
ultra-fine fibrils which specific-enzyme tests suggest to
be composed of fibrin.
I would like to acknowledge the assistance of Dr. W.
Shang Ng, Senior Registrar, United Cardiff Hospitals, for
his help and advice; Dr. Roger Seal, Pathologist, Sully
Hospital, for the use of facilities during part of this
work; Messrs. Kabi Pharmaceuticals for a generous supply
of plasmin (Kabi), and Mrs. Lynn McCarthy for secretarial
1: Biggs, R. and MacFarlane, R. G., (1962). Human
Blood Coagulation and its disorders. Blackwell, Oxon.
2: Hardisty, R. M. and Ingram, G. 1., (1965). Bleeding
Disorders. Blackwell, Oxon.
3: Spaet, T. H., (1966). Hoemostatic homeostasis. Blood,
XXVIII 1 112-123.
4: MacFarlane, R. G., (1939). A simple method of
measuring clot retraction. Lancet, 1 1199.
5: Gaarder, A., Johnson, J. and Owren, P. A., (1961).
Adenosine triphosphate in red cells as a factor in the
adhesiveness of human blood platelets. Nature 192,
6: Buckley, M., et al, (1965). Ultrastructure of
platelets and fibrin in the in vivo formation of the
haemostatic plug. Blood, XXV, 4 604.
7: Zweifach, B. W., (1954). Sib Conference on
Connective Tissues, Josiah Macey Jnr. Foundation, N.
8: Ford, B. J., (1965a). The role of the erythrocyte in
coagulation and haemostasis (British Microcirculation
Society conference address). J. Roy. Micr. Soc., 84
9: Ford, B. J., (1965b). New observations on the role of
the erythrocyte in coagulation and haemostasis. J.
Roy. Micr. Soc., 84, 4 423-426.
10: Ford, B. J., (1968a). The concept of
antipoint applied to submicroscopic fibrillar
structures. Proc. Roy. Micr. Soc., 1, 3,
11: Ford, B. J., (1968b). Micrographic study of
penderocytes. International Yearbook of science and
technology, McGraw Hill, N. Y.: 90.
12: Bull, B. S., Rubenberg, M. L., Dacie, J. V. and
Brain, M. C., (1967). Red blood-cell fragmentation in
microangiopathic haemolytic anaemia: in-vitro
studies. Lancet 2 (7525), 1123-1 125.
13: Bell, R. E., (1963). The origin of burr
erythrocytes. Brit. J. Haemat., IX, 4, 552-555.
*Presented at INTER/MICRO-69, 9-11 September, 1969,
Imperial College, London, England.