/. Embryol. exp. Morph. Vol. 45, pp. 271-281, 1978
Printed in Great Britain © Company of Biologists Limited 1978
271
Cell division, cell elongation
and the co-ordination of crystallin gene expression
during lens morphogenesis in the rat
By J. W. McAVOY 1
From the University of Oxford, Nujfield Laboratory of Ophthalmology
SUMMARY
A quantitative analysis of cell division and cell elongation was carried out during lens
morphogenesis in the rat. At 13 days of development elongating cells in the posterior part
of the lens vesicle (presumptive fibre cells) have a lower mitotic activity than cells in the
anterior vesicle. By 14 days these elongating cells do not divide. Thus at 14 days of development the lens can be separated into two compartments; a proliferation compartment in the
anterior lens and an elongation compartment in the posterior lens.
The three main groups of lens-specific proteins, a-, /?- and y-crystallins, were localized by
immunofluorescence. Alpha-crystallin is the first crystallin to be detected and is localized in
some lens pit cells at 12 days of development. By 14 days all lens cells contain a-crystallin.
Beta- and y-crystallins are detected later at 12£ days and are localized in some cells situated
primarily in the posterior part of the lens vesicle. At later stages of development these
crystallins are restricted to cells of the elongation compartment, i.e. presumptive fibre and
fibre cells. Possible mechanisms that govern the temporal and spatial distribution of crystallins are discussed
INTRODUCTION
Once cells are committed to give rise to a tissue their differentiation involves
the co-ordinated expression of genes coding for tissue specific proteins. Alpha-,
/?- and y-crystallins are the three major groups of lens-specific proteins in
mammals (see Clayton, 1974; Harding & Dilley 1976; for reviews). In a previous
study, a quantitative analysis of cell division and cell elongation showed 1hat
the lens of the newborn rat can be separated into two compartments; a proliferation compartment in the epithelium of the anterior lens and an elongation
compartment in the posterior lens. By immunofluorescence it was shown that
epithelial cells in the proliferation compartment contained only a-crystallin, but
all three groups of crystallins were detected in cells of the elongation compartment. Cells in the latter compartment are surrounded by the vitreous humour
and it was suggested that an elongation factor(s), probably transmitted by the
neural retina which bounds the elongation compartment, provides the signal
for the synthesis of /?- and y-crystallins (McAvoy, 1978). To explore this
1
Author's address: Nuffield Laboratory of Ophthalmology, University of Oxford, Walton
Street, Oxford 0X2 5AW, U.K.
I 8-2
272
J. W. MCAVOY
possibility further, the distribution of crystallins was examined as the compartments formed during lens morphogenesis. This is the first report on the
temporal and spatial distribution of all three groups of crystallins during lens
development in mammals and the results provide some information on the
co-ordination of crystallin gene expression during development and growth of
the rat lens.
MATERIALS AND METHODS
Preparation of embryos
Vaginal smears were taken daily from albino Wistar rats and animals in
pro-oestrus were mated. The day of the vaginal plug was taken as day 0 so that
the duration of pregnancy was 21 days. Females were killed with ether and the
embryos removed from the uterus, immediately fixed in Carnoys (4 °C) and
processed for histology by the method of Sainte-Marie (1962).
Quantitative studies on cell division and cell elongation
Cell division and cell elongation were analysed in the lenses of 12-, 13-, 14and 15-day embryos. For this purpose, cells in sagittal sections of the lens were
assigned positional numbers. Two cell numbering systems were used. In the
first, at 12 and 13 days of development the developing lens was divided by an
anterior/posterior axis and cells were numbered consecutively from both sides
of the axis (see Fig. 1C and E). The second cell numbering system was used for
14- and 15-day lenses and has been described previously (McAvoy, 1978). Cells
were numbered consecutively from about the lens equator in both anterior and
posterior directions, the latter being given negative signs. Cell number 1 was
taken as the last cell in the epithelium with a columnar shape (indicated by the
dotted line in Fig. 1F). Patterns of mitotic activity were analysed by scoring
mitotic cells according to their positions and then calculating the proportions
of the total mitotic cells in the different cell positions in each of the
developmental stages examined. The results are presented in histograms along
with measurements of cell length in the different positions. The data are from
measurements on both eyes of various numbers of embryos, three at 12 days,
three at 13 days, three at 14 days, four at 15 days. The distribution of mitotic
cells was examined in alternate sagittal sections and cell length was measured
in three sagittal sections from each lens.
Preparation of antibodies
Alpha-, /?- and y-crystallins were separated from lens homogenates by
agarose gel chromatography and antibodies were raised in 2-3 kg rabbits as
previously described (McAvoy, 1978). The anti a-, anti /?- and anti y-antibodies
were specific for a-, /?- and y-crystallins respectively as shown by immunoelectrophoresis and immunodiffusion.
Development of the rat lens
273
Immunefluorescence
The anti-crystallin antibodies were used to localize a-, /?- and y-crystallins in
sections of 11-, 12-, 12!-, 13-, 14-and 15-day rat embryos by the indirect immunofluorescence technique of Weller & Coons (1954). Antibodies were applied to
4/<m sagittal sections followed by goat anti-rabbit gamma globulin antibody
conjugated with fluorescein isothiocyanate (FITC, Nordic Immunological
Laboratories, The Netherlands). To determine the specificity of the fluorescence,
the following two controls were routinely carried out. Serum from nonimmunized rabbits was substituted for specific anti-crystallin anti-sera. The
other control was to apply FITC directly to sections without prior incubation
in serum. In both controls, there was no fluorescence except in some cells in
the region where the vitreous humour forms and in clumps of cells throughout
the embryos. These cells, which are probably blood cells, had an orange-yellow
fluorescence compared with the apple green fluorescence of FITC.
The temporal and spatial distribution of the a-, /?- and y-crystallins was
established using serial sections. Sections were cut at 4 ^m and three sections
were mounted per slide. Consecutive slides were used for immunofluorescence
whenever possible.
RESULTS
Patterns of cell division and cell elongation
In rats the presumptive lens cells are morphologically distinguishable from
other head ectoderm cells at about 11 days of development. Ectoderm cells that
overlie, and that are probably in contact with the optic plate (see McKeehan,
1951; Hendrix & Zwaan, 1974; Silver & Wakely, 1974; Wakely, 1977), elongate
to form the lens placode (Fig. 1B). Placode cells invaginate to form the lens pit
(Fig. 1 C) at about 12 days of development and there are no major differences
in mitotic activity between cells in different positions of the lens pit (Fig. 2 A).
The pit closes to form a vesicle which separates from the ectoderm (Fig. ID).
By 13 days, cells in the posterior of the lens vesicle in proximity to the optic cup
(presumptive retina and its derivatives) elongate to fill the cavity of the lens
vesicle and form the primary lens fibres (Fig. 1E). This cavity is closed by about
14 days. In the 13-day embryos, the elongating cells in cell positions 1-10 and
10-20 average in length 111 and 83/tm respectively compared with 25 /im in
cell positions 40-50. Elongating cells have a lower proportion of mitotic figures,
particularly in positions 1-10 where only 2 % of the total lens mitoses are found
(Fig. 2 B). At 14 days 28 % of the total lens mitoses are found in the first ten cell
positions above the lens equator. These cells average 27 fim in length. In the
first ten cell positions below the equator the cells average 59 [im in length and
only 9 % of the lens mitoses occur here. There is no mitotic activity in the
elongating fibre cells beyond cell position (-10). In the lenses of 15-day-old
embryos there is a similar distribution of dividing cells. The ten cells above the
lens equator average 30 fim in length and account for 25 % of total lens mitoses.
274
J. W. MCAVOY
B
11 days : lens placode
10 days
D
12 days : lens pit
ays: early lens vesicle
13 days : late lens vesicle
15-day lens
Fig. 1. (A-F) Diagrams of sagittal sections of the main stages in lens development.
Two cell numbering systems were used to determine patterns of cell division and cell
elongation in the developing lenses; (i) in lenses of 12- and 13-day (as well as 14-day)
old embryos, cells were numbered from either side of an anterior/posterior axis,
(ii) in 15-day lenses cells were numbered from either side of the equator. The diagrams are schematic and only a small proportion of the cells are represented.
275
Development of the rat lens
40
A
12 days
B
13 days
30
30
120
20
, s 20
80
10
40
=
U
10 20 30
Cell position
C
10 20 30 40 50
Cell position
14 days
80
60
30
T
20
1
\
40 ob
\
20
10
- 2 0 - 1 0 A 10 20 30
' Lensequator
Cell position
D 15 days
30
20
1\
\
10
U
40 50
60
T
1
40
20
I
-20-10 • . 10 20 30 40 50 60 70. 80
Lens equator
Cell position
Fig. 2. (A-D) The histograms show % mitoses in relation to cell positions in lenses
of 12-day (A), 13-day (B), 14-day (C) and 15-day (D) embryos. Cell length in the
different cell positions is also plotted in the 13-, 14-, and 15-day embryos. See
Fig. 1 for methods of numbering cells. Standard deviations are included.
276
J. W. MCAVOY
The ten cells below the lens equator average 53 ^m in length and only account
for 5-5% of total lens mitoses. There is no mitotic activity in the elongating
fibre cells beyond cell position (— 10).
Thus by 14 days mitotic activity in the lens is found in the epithelial cells
above the equator and, to a lesser extent, in the first ten cells below the equator.
Since the cells below the equator show the greatest degree of elongation, the
lens of 14-day-old embryos can be divided into two functionally distinct compartments; a proliferation compartment and an elongation compartment. This
functional separation coincides approximately with the anatomical separation of
the anterior and posterior parts of the lens by the border of the optic cup. A
decrease in mitotic activity in elongating cells of the lens has also been described
in the developing chick (Modak, Morris & Yamada, 1968).
At 15 days there is a gradual decrease in mitotic activity within the proliferation
compartment from the lens equator to the centre of the lens epithelium, i.e.
from cell positions 1-80.
Distribution of crystallins
Alpha-crystallin
Anti a-crystallin antibodies were used to localize a-crystallin by indirect
immunofluorescence. Apple green fluorescence, indicating the presence of
a-crystallin, was present in some cells of the lens pit of 12-day embryos (Fig. 3 a).
Beta- and y-crystallins were not detected at this stage, therefore a-crystallin is
the first lens-specific protein to be synthesized in detectable amounts in the
developing lens. By \2\ days the lens vesicle has formed and all cells except
the cells in the middle of the anterior part of the vesicle (presumptive lens
epithelial cells) contain detectable amounts of a-crystallin (Fig. 3b). In 13-day
lenses there is a faint fluorescence in many of the presumptive lens epithelial
cells (Fig. 3c) and by 14 days all lens cells contain a-crystallin (Fig. 3d).
Beta- and y-crystallins
These two groups of lens-specific proteins have a similar temporal and spatial
distribution during lens morphogenesis. Using anti /?- and anti y-antibodies,
neither group of crystallins is detected at the lens pit stage (Fig. 3 e and i), but
both groups are present in some cells of the lens vesicle at 12| days of development (Fig. 3 / and /). In both cases faint fluorescence is present in some cells
situated mainly in the posterior part of the vesicle. Only cells in the middle of
the anterior part of the vesicle, those cells in which a-crystallin is not detected,
are completely negative for /?- and y-crystallins. By 13 days most cells in the
posterior part of the lens vesicle fluoresce for both /?- and y-crystallins (Fig. 3g
and k). Fluorescence for y-crystallin is weaker than for /?-crystallin and is only
clearly present in the longest cells. Besides these differences the general distribution of/?- and y-crystallins is similar in that they are both present in detectable amounts only in the elongating cells of the lens vesicle. These cells also
Development of the rat lens
277
Fig. 3. Immunofluorescence localization of a-crystallin {a-d), /?-crystallin (e-h) and
y-crystallin (/-/) in lenses of 12-day (a, e and /), 12^-day (6,/and./) 13-day (c, g and
k) and 15-day {d, h and /) embryos. The localization of a-, /?- and y-crystallins in
12-day and 12-^-day lenses as well as the localization of /?- and y-crystallins in
13-day and 15-day lenses was carried out using serial sections. Some cells outside the
lens autofluoresce (arrows).
278
J. W. MCAVOY
have the lowest mitotic activity in the lens. By 14 days of development, the
proliferation and elongation compartments have formed and /?- and y-crystallins
are detected only in cells beyond about position (-10) in the elongation
compartment (Fig. 3/z and /).
DISCUSSION
Ectoderm cells that are in contact with the optic plate elongate to form the
lens plate. The lens'plate invaginates to form the lens pit. This probably depends
on the contraction of cytoplasmic microfilaments at the apex of the cells
(Wrenn & Wessells, 1969) as well as increased adhesion between tissues and
changes in patterns of cell proliferation (Hendrix, 1975). It is about this stage
in development of the rat lens that a-crystallin is detected by immunofluorescence
and thus provides the first evidence of lens-specific protein synthesis.
Distribution of crystallins
Alpha-crystallin is first detected in some lens pit cells at 12 days of development and by 14 days all lens cells contain a-crystallin. Beta- and y-crystallins are
first detected at \2\ days in some cells situated primarily in the posterior part
of the early lens vesicle, i.e. presumptive fibre cells. At later stages of lens
morphogenesis /?- and y-crystallins are exclusive to presumptive fibre and fibre
cells.
This temporal sequence of crystallin synthesis in rats is similar to that in mice
(Barabanov, 1965). In Rana, /?- and y-crystallins are detected before a-crystallin
(McDevitt, 1972) but in normal lens development (Ogawa, 1965) and lens
regeneration in Triturus (Yamada, 1966) a- and /?-crystallins appear before
y-crystallin. In normal chicks a- is the last crystallin to be detected (Zwaan &
Ikeda, 1968; Clayton, 1978). However, in the mutant Hyl chick a- appears
about the same time as the other crystallins (McDevitt & Clayton, 1975;
Clayton, 1978).
The spatial distribution of a-crystallin in rats is similar to that in all other
species examined. Alpha-crystallin is present in presumptive epithelial and
epithelial cells as well as in presumptive fibre and fibre cells in mice (Zwaan,
1975), in chicks (Ikeda & Zwaan, 1967; Zwaan & Ikeda, 1968; Brahma & van
Doorenmaalen, 1971) and during lens regeneration in Triturus (Yamada, 1966).
Although /?-crystallin is not present in presumptive epithelial or epithelial cells
in the rat it has been localized in these cells in Triturus (Yamada, 1966) and in
both normal (Zwaan & Ikeda, 1968) and mutant chicks (McDevitt & Clayton,
1975). The only other localization of y-crystallin during lens morphogenesis in
mammals was also in the rat. Schubert, Trevithick & Hollenberg (1970) did not
detect y-crystallin until the late lens vesicle stage. Moreover, at later stages of
development these workers found y-crystallin in the presumptive epithelium
as well as outside the lens. In the study reported here y-crystallin is also detected
Development of the rat lens
279
in the presumptive epithelium and outside the lens but in these cases, localization
is inconsistent between lenses of the same developmental stage as well as between
sections of the same lens. Because the only reproducible localization of ycrystallin is in the presumptive fibre and fibre cells this is taken to be the true
localization of this crystallin in the rat lens. In all amphibians y-crystallin is
exclusive to presumptive fibre and fibre cells both during normal lens development (Yamada, 1966; McDevitt, Meza & Yamada, 1969; McDevitt & Brahma,
1973; Brahma & McDevitt, 1974a; McDevitt & Brahma, 1977) as well as during
lens regeneration from the iris epithelium (Yamada, 1966; Takata, Albright &
Yamada, 1966; Brahma & McDevitt, 19746).
The appearance of /?- and y-crystallins in cells of the posterior part of the
lens vesicle is associated with a drop in mitotic activity in this region and by
14 days these cells, the presumptive fibre and fibre cells, do not divide. Thus at
these later stages of lens development, and in the newborn rat lens (McAvoy,
1978), the synthesis of detectable amounts of these two groups of crystallins
and cell proliferation might be mutually exclusive processes. This may also be
the case at the lens vesicle stage since cells that initiate synthesis of /?- and ycrystallins, perhaps during or shortly after the S phase of the cell cycle (see
Zwaan, 1974), might complete the division cycle and then remain in a nondividing state. Therefore, as the cells divide asynchronously there is a gradual
decrease in mitotic activity as more cells initiate /?- and y-crystallin synthesis.
A mutually exclusive relationship between the synthesis of y-crystallin and cell
division has also been reported by McDevitt et al. (1969) during lens development in Ranapipiens and by Yamada (1966) during lens regeneration in Tritums
viridescens.
Co-ordination of crystallin gene expression
There are two major ways in which the temporal sequence of crystallin
synthesis could be governed. One possibility is that the lens-inducing tissue,
the optic plate/cup, transmits a sequence of factors. The first to be transmitted
provides a specific signal for a-crystallin synthesis; then at a later stage in
development another specific factor(s) is released that signals synthesis of /?and y-crystallins. An alternative explanation is that the sequence in which the
genes are expressed is governed by a hierarchy of gene function. In this case a
factor(s) that induces crystallin synthesis would not be specific for any particular
crystallin gene group and crystallin synthesis may even be part of a general
response to a signal initiating lens differentiation. Genes involved in the proliferative cell cycle may also be included in this hierarchy since there appears to be
a mutually exclusive relationship between cell proliferation and synthesis of
/?- and y-crystallins, at least by 14 days of development. The presence of a
functional hierarchy within gene groups has also been suggested in other differentiating systems (see for example Tsanev, 1975). The different sequences of
crystallin synthesis shown by the various species can be explained in terms of the
280
J. W. MCAVOY
former suggestion by a species-specific sequence of factors from the optic cup
or for the latter suggestion by a species-specific functional hierarchy within
the gene groups.
Alpha-crystallin is present in all lens cells but /?- and y-crystallins are detected
only in non-dividing cells of the elongation compartment, i.e. presumptive fibre
and fibre cells. The elongation compartment is contained within the optic cup
(see Fig. 1F). Therefore although the signal(s) for crystallin synthesis probably
comes directly or indirectly from the optic cup, synthesis of a-crystallin spreads
to cells of the proliferation compartment whereas synthesis of/?- and y-crystallins, at least in detectable quantities, only occurs in cells continuously under the
influence of the optic cup. This indicates that the conditions for the expression
of /?- and y-crystallin genes are continuous induction and/or cessation of the
proliferative cell cycle. In any case the conditions are provided by the optic cup
environment. The presence of /?-crystalltn in the cells of the proliferation
compartment of Triturus and the chick shows that these conditions do not
apply to all species.
Thus the separation of the lens into compartments and either a specific
sequence of signals for crystallin synthesis or a hierarchy of gene function might
co-ordinate the expression of crystallin genes during rat lens morphogenesis.
The latter two possibilities are not necessarily mutually exclusive.
I am grateful to Bridget Galdes and Linda Palfrey for their skilled technical assistance.
Thanks also go to Dr John Harding and Dr Ruth van Heyningen for their helpful comments
on the manuscript. The support for this research was provided by grant no. 1. RO1 .EYO
1548-01 awarded by the N.E.I, to Ruth van Heyningen, to whom I also owe special thanks
for her continued interest in this work.
REFERENCES
V. M. (1965). Immunoelectrophoretic analysis of antigens of the lens in the
embryonic development of mice. Zh. Obshch. Biol. 27, 252-255.
BRAHMA, S. K. & VAN DOORENMAALEN, W. J. (1971). Immunoflorescence studies of chick
lens FJSC and a-crystallin antigens during lens morphogenesis and development. Ophthal.
Res. 2, 344-357.
BRAHMA, S. K. & MCDEVITT, D. S. (1974O). Ontogeny and localization of gamma-crystallins
in Rana temporaria, Ambystoma mexicanum and Pleurodeles waltlii normal lens development. Expl Eye Res. 19, 379-387.
BRAHMA, S. K. & MCDEVITT, D. S. (19746). Ontogeny and localisation of the lens crystallins
in Xenopus laevis lens regeneration. /. Embryol. exp. Morph. 32, 783-794.
CLAYTON, R. M. (1974). Comparative aspects of lens proteins. In The Eye, Volume 5 (ed.
H. Davson & L. T. Graham), pp. 399-494. London: Academic Press.
CLAYTON, R. M. (1978). Divergence and convergence in lens cell differentiation: regulation
of the formation and specific contents of lens fibre cells. In Stem Cells and Tissue Homeostasis (ed. B. Lord, C. Potten & R. J. Cole). Cambridge University Press. (In the Press.)
HARDING, J. J. & DILLEY, K. J. (1976). Structural proteins of the mammalian lens: a review
with emphasis on changes in development, aging and cataract. Expl Eye Res. 22, 1-73.
HENDRIX, R. W. (1975). Chick neurulation and lens invagination: co-operation between cell
division and inter-tissue adhesion. Anat. Rec. 181, 373.
BARABANOV,
Development of the rat lens
281
R. W. & ZWAAN, J. (1974). Changes in the glycoprotein concentration of the extracellular matrix between lens and optic vesicle associated with early lens differentiation.
Differentiation 2, 357-362.
IKEDA, A. & ZWAAN, J. (1967). The changing cellular localization of a-crystallin in the lens
of the chicken embryo, studied by immunofluorescence. Devi Biol. 15, 348-367.
MCAVOY, J. W. (1978). Cell division, cell elongation and distribution of a-, -ft-, and ycrystallins in the rat lens. /. Embryol. exp. Morph. 44, 149-165.
MCDEVITT, D. S. (1972). Sequential appearance of a, fi and y-crystallins in embryonic and
metamorphic Rana pipiens. Experientia 28, 1215-1217.
MCDEVITT, D. S. & BRAHMA, S. K. (1973). Ontogeny and localization of the crystallins during
embryonic lens development in Xenopus laevis. J. exp. Zool. 186, 127-140.
MCDEVITT, D. S. & BRAHMA, S. K. (1977). Eye lens development and y crystallins in Discoglossus pictas (Anura). Experientia 33, 1087-1089.
MCDEVITT, D. S. & CLAYTON, R. M. (1975). Ontogeny and localization of a, /? and 8 crystallins in the Hy-1 chick lens. /. Cell Biol. 67, 273a.
MCDEVITT, D. S., MEZA, I. & YAMADA, T. (1969). Immunofluorescence localization of the
crystallins in amphibian lens development, with special reference to the y-crystallins.
Devi Biol. 19, 581-607.
MCKEEHAN, M. S. (1951). Cytological aspects of embryonic lens induction in the chick.
/. exp. Zool. 117, 31-64.
MODAK, S. P., MORRIS, G. & YAMADA, T. (1968). DNA synthesis and mitotic activity during
early development of chick lens. Devi Biol. 17, 544-561.
OGAWA, T. (1965). Appearance of lens antigens during the embryonic development of the
newt. Embryologia 8, 345-361.
SAINTE-MARIE, G. (1962). A paraffin embedding technique for studies employing immunofluorescence. /. Histochem. Cytochem. 10, 250-257.
SCHUBERT, E. E., TREVITHICK, J. R. & HOLLENBERG, M. J. (1970). Localisation of gamma
crystallins in the developing lens of the rat. Can. J. Ophthalmol. 5, 353-365.
SILVER, P. H. S. &WAKELY, J. (1974). Fine structure, origin and fate of extracellular materials
in the interspace between the presumptive lens and presumptive retina of the chick embryo.
J.Anat. 118, 19-31.
TAKATA, C , ALBRIGHT, J. F. & YAMADA, T. (1966). Gamma crystallins in Wolffian lens
regeneration demonstrated by immunofluorescence. Devi Biol. 14, 382-400.
TSANEV, R. (1975). Cell cycle and liver function. In Cell Cycle and Cell Differentiation (ed.
J. Reinert & H. Holtzer), pp. 197-248. Berlin, Heidelberg: Springer-Verlag.
WAKELY, JENNIFER (1977). Scanning electron microscope study of the extracellular matrix
between presumptive lens and presumptive retina of the chick embryo. Anat. Embryol.
150, 163-170.
WELLER, T. H. & COONS, A. H. (1954). Fluorescent antibody studies with agents of varicella
and herpes zoster propagated in vitro. Proc. Soc. exp. Biol. Med. 86, 789-794.
WRENN, J. T. & WESSELLS, N. K. (1969). An ultra structural study of lens invagination in the
mouse. /. exp. Zool. 171, 359-368.
YAMADA, T. (1966). Control of tissue specificity. The pattern of cellular synthetic activities
in tissue transformation. Am. Zool. 6, 21-31.
ZWAAN, J. (1974). Mitotic activity in the lens rudiment of the chicken embryo before and
after the onset of crystallin synthesis. II. Immunofluorescence studies. Wilhelm Roux Arch.
EntwMech. Org. 175, 13-25.
ZWAAN, J. (1975). Cell cycle and onset of cytodifferentiation in the embryonic mouse lens.
Anat. Rec. 181, 517-518.
ZWAAN, J. & IKEDA, A. (1968). Macromolecular events during differentiation of the chicken
lens. Expl Eye Res. 7, 301-311.
HENDRIX,
(Received 1 December 1977, revised 20 January 1978)