@Copyright 1987 by The Humana Press Inc. All rights of any nature, whatsoever, reserved. 01634984/87/1300-0159502.00
Elemental Microanalysis of Individual Blood Cells P. M. O'BRIEN* AND G. J. F. LEGGE
School of Physics, University of Melbourne, Parkville, Victoria, 3052, Australia ABSTRACT A scanning proton microprobe was used to study single red blood cells in freeze-dried, whole-blood specimens. Simultaneous collection of PIXE and of forward and backward scattered proton data provided information on the heavier elements (Z ~ 27) and on the organic mass under investigation. The trace elemental spectrum of a red cell was found to be reproducible, and no elemental losses were observed during proton bombardment. The spatial resolution of the probe (1.5 ~m for these studies) enabled the red cell's biconcave disk shape to be visualized in quantitative two- and three-dimensional maps of H, C, P, S, C1, K, and Fe. Index Entries: Scanning proton microprobe; elemental mapping; erythrocyte; red blood cell, elemental microanalysis of.
INTRODUCTION The Melbourne Scanning proton Microprobe (MP) has been successfully employed in many interdisciplinary studies, especially in the fields of medicine (1,2) and biology (3,4). Since the spatial resolution of the microprobe is currently I ~m, studies involving the elemental microanalysis of intracellular structure are possible. We report here on one such study, that of erythrocytes or red blood cells. This work is part of an ongoing study of the hemopoietic system, with the aim of establishing whether the trace elemental spectrum, obtained by proton-induced X-ray emission (PIXE), provides a characteristic and reliable "fingerprint" of a cell. Other studies by Lindh et al. (5) and Heck et al. (6) have employed *Author to whom all correspondence and reprint requests should be addressed. Biological Trace Element Research
] 59
VoL 13, 1987
160
O'Brien and Legge
proton microprobes in the trace elemental analysis of erythrocytes. We wanted to extend such measurements to internal structure at this resolution, apart from its shape. Since red blood cells are the most abundant cell type found in whole blood, they are readily available in the form of simple aspiration, requiring no special isolating procedures. A fully mature red cell has a characteristic biconcave disk shape, contains no cell nucleus, and in the hydrated state has a mean thickness of 2.1 ~m and an average diameter of 7.0-7.5 ~m (7), which decreases with age. The cells are actively involved if supplying necessary oxygen to living tissue throughout the body, with their major transportational functionality provided by the Fe-complexed molecule, hemoglobin, a feature that should be reflected in the cell's elemental spectrum.
PREPARATION The choice of blood donor was a normal, healthy human not undertaking any course of prescribed medication or with any recent history of illness. Whole blood samples were taken from the arm via venous aspiration into a clean, sterilized, 5-mL glass syringe, with a large bore needle to minimize any mechanical stress during aspiration. The whole blood sample was then immediately transfered to a lithium-heparin-lined vial to minimize the rate of coagulation in the short term. Specimens were then prepared by smearing a small quantity of this whole blood on to thin, ultraclean, nylon foils (8), producing many localized areas of red cell monolayers ideally spaced for analysis with a scanning proton beam. These specimens were then rapidly immersed in a cryofixative (isopentane, GPR, BDH Chemicals) cooled almost to liquid nitrogen temperature and subsequently freeze dried at a pressure of - 1 0 5 Torr for 24 h. The time involved in smearing and cryofixation was less than 15 s, to minimize any possible stress to the cells relative to their microenvironment and the migration of trace elements important to this work. During the period between preparation and analysis ( - 2 wk) the specimens were stored in a dessicator over silica gel.
ANALYSIS The specimens were analyzed by MP (9,10), with a 3-MeV beam of protons from a National Electrostatics Corporation (NEC) 5U Pelletron charged-particle accelerator. For these measurements, a 100-150 pA proton beam was focused to a spot of approximately 1.5 ~m in diameter and evenly scanned over the area of interest. A Si(Li) detector mounted at 135~ to the incident beam was used to collect X-rays, a surface barrier detector (SBD) mounted at 135~ to detect Rutherford backscattered protons (RBS), a SBD mounted at 60~ to detect Rutherford forward-scattered Biological Trace Element Research
VoL 13, 1987
Elemental ,Microanalysis of Individual Blood Cells
161
protons (RFS), and a secondary electron detector to study target surface features. The backward and forward mounted SBD's were used to investigate the organic matrix in which the inorganic trace elements were located (3,4,9). The secondary electron detector was used to identify each individual cell, which was then centered in the image, and the scan size was collapsed to - 1 2 x 12 ~m around the cell, for maximum data collection efficiency. Each cell was irradiated for approximately 2 h, with the X-ray and back- and forward-scattered proton data collected simultaneously, together with the spatial coordinates for each independent event. These events were stored in real time on magnetic tape to be analyzed later by the Melbourne data reduction and computer graphics system
(11).
RESULTS AND DISCUSSION In the X-ray spectra of all 18 cells examined, the elements P, S, C1, K, Ca, Fe, and traces of Zn at the 1 ppm level were observed, as shown in Fig. 1 for such a cell. Figure 2 shows a set of elemental maps extracted from the irradiation of a single red cell. The e or secondary electron image was used both to study any surface features of the cell and to position the cell with micron accuracy. The Scan map is a smoothed Y- modulation map for C1 that displays a pseudo-three-dimensional image and more clearly outlines the cell relative to its surroundings. The remaining I
'
I
'
I
'
I
'
I
'
I
Single Red Blood Cell
1200
CI 1000
x10
Z o
(,3
800
Fe
4-,
o
600
L eJ E
n
400
Z
K 200
ll]~l
0 I
0
,
I
2O0
,
I
400
,
I
600
,
iIZIn I
800
,
II I
1000
ChQnnel Number-
Fig. 1. The X-ray spectrum of a single red blood cell extracted from a scan subregion. Biological Trace Element Research
Vol. 13, 1987
O'Brien and Legge
162
[ ;,.3
e-
, / # ~ ' ~ .-.-#-.J~_X%~',~/'K
Scan
H
~s i'
9
%;,
,~'..?~ ,',/:~ ~.'
"',--/," .-..~.~
,%.9
9 :,'~
"vS~'}L~t
C
P
-
"
.i ..~.
. '~ ~,4,-
S
F
,f
CI K Fe Fig. 2. A set of maps obtained by scanning a single red blood cell (the cell's X-ray spectrum is displayed in Fig. 1) with a 150-pA, 3-MeV beam of protons focused to a beam diameter of 1.5 p~m - 3 h. The e (secondary electron image) map, the Scan map (Y-modulation map for C1), and the remaining elemental maps all display a "ring" structure, which is consistent with this cell's characteristic biconcave disk shape. maps show the distribution of various elements in the scanned area. Each map displays the characteristic biconcave disk shape of a red cell. A cell being of the order of 8 p~m in diameter, the beam resolution must be better than 2 ~ m for this "ring" structure to be observed. The map for Fe shows that this element is intrinsic to the cell, for little is found in the cell surrounds, a distinguishing feature of this cell type. Mechanical and electronic stability had to be maintained within 1 p~m over the 3-h period of the scan. Figure 3 shows a set of three-dimensional maps of the same cell, with the third-dimension Z being element intensity. The maps were produced by a 135 ~ rotation about Z followed by a 60 ~ rotation about X and scaled in Z proportionally to X-ray yield. These three-dimensional maps
Biological
Trace Element Research
Vol. 13,
1987
Elemental l~icroanalysis of Individual Blood Cells
163
II
P
$
K
Fe
"0 Fig. 3. A set of three-dimensional representations of the same red blood cell displayed in Fig. 2. The Z dimension is proportional to elemental intensity and provides a more realistic picture of the cell. give a more realistic picture of the original specimen and provide an enhanced visual representation of the concentration and distribution of an element in that specimen. High-spatial-resolution proton beams are obtainable at the expense of beam current, but nevertheless have high current densities. In a study of single cells it is necessary that the cells do not overlap and are suitably separated to scan singly or in small clusters, since large area scanning is inefficient. Continuous scanning of the beam is used to minimize any elemental losses, and such losses are monitored. The simultaneous collection of RBS and RFS allows such monitoring of the major specimen constituents H, C, N, and O, with C providing a normalization to mass density variations. Figure 4 shows, for elements C1 and K, the n u m b e r of counts plotted as a function of the C count. Other studies (3,12) have Biological TraceElement Research
VoL 13, 1987
164
O'Brien and Legge
s h o w n C to be a stable element u n d e r proton beam b o m b a r d m e n t and, hence, it essentially provides a time axis i n d e p e n d e n t of beam current fluctuations. As Fig. 4 shows, the elemental concentrations of these ele60
'
I
'
I
,
I
'
I
'
I
,
I
,
I
'
I
,
5O
E Z
10 0
I
0
200
400 600 Channel Number
t
B00
1000 :3
hrs
12.0
"t
10.0
5 eo 3
~o 6.o
_~
4.0
7= 2.0 O0
I
200
I
I 400 600 Chonnel Number
I
i
BOO
10 0 0 =3
hrs
Fig. 4. The X-ray yield for the elements C1 and K as a function of carbon count (effectively time), showing that these elements are stable under proton beam bombardment, and no losses are observed. Biological TraceElementResearch
VoL 13, 1987
Elemental Microanalysis of Individual Blood Cells
165
ments remained constant over the period of irradiation and no losses were observed. One of the major considerations associated with the elemental sensitivity of PIXE analysis is the introduction of contaminants during specimen preparation; conventionally used techniques may not be desirable since they are not developed with such considerations in mind. The X-ray energy spectrum extracted from the displayed cell (Fig. 1) shows a distinct Si peak, which was attributed to contamination from the plastic syringe used for aspiration, and subsequent use of an ultraclean glass syringe removed this peak from the spectrum. A distinctive feature of the two-dimensional maps in Fig. 2 is a "dark band" or "valley" around the cell; this is seen also in Fig. 3. On observation under a light microscope, the cell appeared to have pulled away and separated from the surrounding plasma. The small X-ray contribution from the gap between the cell and the plasma was produced by the thin nylon support. This separation was attributed to the differential rate of cooling between the cell and its surroundings during cryofixation. During data analysis, this band provides a clear outline of the cell for accurate extraction of the X-ray yield from the cell area. Although this band provides a clear cell-plasma boundary, it is difficult to estimate the remaining plasma still adhering to the cell and it is a possible source of error in a quantitative study of elemental concentrations. The alternative to the use of whole blood is washing and resuspending the red cells in suitable buffers to remove the plasma. The problems encountered with this technique are the introduction of contaminant elements from buffers and their adjustment to the correct pH and tonicity to maintain cell integrity, an important requirement for cells of the sensitivity of red blood cells.
CONCLUSION In all the red blood cells examined, the X-ray yield of each element when normalized to C or the cell's effective mass, was observed to be consistent and reproducible for each cell. Thus, the X-ray spectrum of a single red cell provides a reliable "fingerprint" for this blood cell type, especially with regard to its relatively large Fe concentration. The spatial resolution of the probe enabled the observation of the "ring" structure associated with this cell's shape and the probe's sensitivity enabled the detection of Zn in each cell at the 1-ppm level. The data presented in this paper show only the results for a single cell and illustrate the techniques that are currently available for quantitative single cell analysis.
ACKNOWLEDGMENTS The assistance of V. Ivanov (Royal Melbourne Hospital, Victoria) and J. Cerini in the preparation of specimens was much appreciated. D. Biological Trace Element Research
Vol. 13, 1987
166
O'Brien and Legge
N. Jamieson's contribution to the preparation of the three-dimensional maps is gratefully acknowledged. This work was supported in part by the Australian Research Grants Scheme (B82115688).
REFERENCES 1. P. M. O'Brien, G. J. F. Legge, T. R. Bradley, and G. S. Hodgson, Aust. Phys. and Eng. Scs. Med. vol. 5, no. 1, 1982, pp. 30-34. 2. G. J. F. Legge, A. P. Mazzolini, A. F. Rocznick, and P. M. O'Brien, Nucl. Instr. Meth. 197, 191. (1982). 3. A. P. Mazzolini, G. J. F. Legge, and C. K. Pallaghy, Nucl. Instr. Meth. 191, 583 (1981). 4. A. P. Mazzolini, C. K. Pallaghy, and G. J. F. Legge, The New Phytologist, (1985) 100, pp. 483-509. 5. U. Lindh, E. Johansson, and L. Gille, Nucl. Instr. Meth. B3 631 (1984). 6. D. Heck and E. Rokita; IEEE Trans. Nucl. Sci. vol. NS-30, no. 2 1983, 1220-1223. 7. H. Begemann and H.-G. Harwerth, Pr~ktische Hamatologie, 3rd ed., Georg Thieme, Verlag, Stuttgart 1967, p. 47. 8. P. Echlin and R. Moreton, Microprobe Analysis as Applied to Cells and Tissues, (T. Hall, P. Echlin, and R. Kaufman, eds.), Academic, London, New York (1974), pp.159-174. 9. G. J. F. Legge, C. D. McKenzie, and A. P. Mazzolini, J. Micros. 117, 185 (1979). 10. G. J. F. Legge, D. N. Jamieson, P. M. O'Brien and A. P. Mazzolini, Nucl. Instr. Meth. 197, 85 (1982). 11. G. J. F. Legge and I. Hammond, J. Micros. 117, 201 (1979). 12. A. P. Mazzolini, PhD Thesis, University of Melbourne (1983), Unpublished.
Biological Trace Element Research
Vol. 13, 1987
Elemental Microanalysis of Individual Blood Cells P. M. O'BRIEN* AND G. J. F. LEGGE
School of Physics, University of Melbourne, Parkville, Victoria, 3052, Australia ABSTRACT A scanning proton microprobe was used to study single red blood cells in freeze-dried, whole-blood specimens. Simultaneous collection of PIXE and of forward and backward scattered proton data provided information on the heavier elements (Z ~ 27) and on the organic mass under investigation. The trace elemental spectrum of a red cell was found to be reproducible, and no elemental losses were observed during proton bombardment. The spatial resolution of the probe (1.5 ~m for these studies) enabled the red cell's biconcave disk shape to be visualized in quantitative two- and three-dimensional maps of H, C, P, S, C1, K, and Fe. Index Entries: Scanning proton microprobe; elemental mapping; erythrocyte; red blood cell, elemental microanalysis of.
INTRODUCTION The Melbourne Scanning proton Microprobe (MP) has been successfully employed in many interdisciplinary studies, especially in the fields of medicine (1,2) and biology (3,4). Since the spatial resolution of the microprobe is currently I ~m, studies involving the elemental microanalysis of intracellular structure are possible. We report here on one such study, that of erythrocytes or red blood cells. This work is part of an ongoing study of the hemopoietic system, with the aim of establishing whether the trace elemental spectrum, obtained by proton-induced X-ray emission (PIXE), provides a characteristic and reliable "fingerprint" of a cell. Other studies by Lindh et al. (5) and Heck et al. (6) have employed *Author to whom all correspondence and reprint requests should be addressed. Biological Trace Element Research
] 59
VoL 13, 1987
160
O'Brien and Legge
proton microprobes in the trace elemental analysis of erythrocytes. We wanted to extend such measurements to internal structure at this resolution, apart from its shape. Since red blood cells are the most abundant cell type found in whole blood, they are readily available in the form of simple aspiration, requiring no special isolating procedures. A fully mature red cell has a characteristic biconcave disk shape, contains no cell nucleus, and in the hydrated state has a mean thickness of 2.1 ~m and an average diameter of 7.0-7.5 ~m (7), which decreases with age. The cells are actively involved if supplying necessary oxygen to living tissue throughout the body, with their major transportational functionality provided by the Fe-complexed molecule, hemoglobin, a feature that should be reflected in the cell's elemental spectrum.
PREPARATION The choice of blood donor was a normal, healthy human not undertaking any course of prescribed medication or with any recent history of illness. Whole blood samples were taken from the arm via venous aspiration into a clean, sterilized, 5-mL glass syringe, with a large bore needle to minimize any mechanical stress during aspiration. The whole blood sample was then immediately transfered to a lithium-heparin-lined vial to minimize the rate of coagulation in the short term. Specimens were then prepared by smearing a small quantity of this whole blood on to thin, ultraclean, nylon foils (8), producing many localized areas of red cell monolayers ideally spaced for analysis with a scanning proton beam. These specimens were then rapidly immersed in a cryofixative (isopentane, GPR, BDH Chemicals) cooled almost to liquid nitrogen temperature and subsequently freeze dried at a pressure of - 1 0 5 Torr for 24 h. The time involved in smearing and cryofixation was less than 15 s, to minimize any possible stress to the cells relative to their microenvironment and the migration of trace elements important to this work. During the period between preparation and analysis ( - 2 wk) the specimens were stored in a dessicator over silica gel.
ANALYSIS The specimens were analyzed by MP (9,10), with a 3-MeV beam of protons from a National Electrostatics Corporation (NEC) 5U Pelletron charged-particle accelerator. For these measurements, a 100-150 pA proton beam was focused to a spot of approximately 1.5 ~m in diameter and evenly scanned over the area of interest. A Si(Li) detector mounted at 135~ to the incident beam was used to collect X-rays, a surface barrier detector (SBD) mounted at 135~ to detect Rutherford backscattered protons (RBS), a SBD mounted at 60~ to detect Rutherford forward-scattered Biological Trace Element Research
VoL 13, 1987
Elemental ,Microanalysis of Individual Blood Cells
161
protons (RFS), and a secondary electron detector to study target surface features. The backward and forward mounted SBD's were used to investigate the organic matrix in which the inorganic trace elements were located (3,4,9). The secondary electron detector was used to identify each individual cell, which was then centered in the image, and the scan size was collapsed to - 1 2 x 12 ~m around the cell, for maximum data collection efficiency. Each cell was irradiated for approximately 2 h, with the X-ray and back- and forward-scattered proton data collected simultaneously, together with the spatial coordinates for each independent event. These events were stored in real time on magnetic tape to be analyzed later by the Melbourne data reduction and computer graphics system
(11).
RESULTS AND DISCUSSION In the X-ray spectra of all 18 cells examined, the elements P, S, C1, K, Ca, Fe, and traces of Zn at the 1 ppm level were observed, as shown in Fig. 1 for such a cell. Figure 2 shows a set of elemental maps extracted from the irradiation of a single red cell. The e or secondary electron image was used both to study any surface features of the cell and to position the cell with micron accuracy. The Scan map is a smoothed Y- modulation map for C1 that displays a pseudo-three-dimensional image and more clearly outlines the cell relative to its surroundings. The remaining I
'
I
'
I
'
I
'
I
'
I
Single Red Blood Cell
1200
CI 1000
x10
Z o
(,3
800
Fe
4-,
o
600
L eJ E
n
400
Z
K 200
ll]~l
0 I
0
,
I
2O0
,
I
400
,
I
600
,
iIZIn I
800
,
II I
1000
ChQnnel Number-
Fig. 1. The X-ray spectrum of a single red blood cell extracted from a scan subregion. Biological Trace Element Research
Vol. 13, 1987
O'Brien and Legge
162
[ ;,.3
e-
, / # ~ ' ~ .-.-#-.J~_X%~',~/'K
Scan
H
~s i'
9
%;,
,~'..?~ ,',/:~ ~.'
"',--/," .-..~.~
,%.9
9 :,'~
"vS~'}L~t
C
P
-
"
.i ..~.
. '~ ~,4,-
S
F
,f
CI K Fe Fig. 2. A set of maps obtained by scanning a single red blood cell (the cell's X-ray spectrum is displayed in Fig. 1) with a 150-pA, 3-MeV beam of protons focused to a beam diameter of 1.5 p~m - 3 h. The e (secondary electron image) map, the Scan map (Y-modulation map for C1), and the remaining elemental maps all display a "ring" structure, which is consistent with this cell's characteristic biconcave disk shape. maps show the distribution of various elements in the scanned area. Each map displays the characteristic biconcave disk shape of a red cell. A cell being of the order of 8 p~m in diameter, the beam resolution must be better than 2 ~ m for this "ring" structure to be observed. The map for Fe shows that this element is intrinsic to the cell, for little is found in the cell surrounds, a distinguishing feature of this cell type. Mechanical and electronic stability had to be maintained within 1 p~m over the 3-h period of the scan. Figure 3 shows a set of three-dimensional maps of the same cell, with the third-dimension Z being element intensity. The maps were produced by a 135 ~ rotation about Z followed by a 60 ~ rotation about X and scaled in Z proportionally to X-ray yield. These three-dimensional maps
Biological
Trace Element Research
Vol. 13,
1987
Elemental l~icroanalysis of Individual Blood Cells
163
II
P
$
K
Fe
"0 Fig. 3. A set of three-dimensional representations of the same red blood cell displayed in Fig. 2. The Z dimension is proportional to elemental intensity and provides a more realistic picture of the cell. give a more realistic picture of the original specimen and provide an enhanced visual representation of the concentration and distribution of an element in that specimen. High-spatial-resolution proton beams are obtainable at the expense of beam current, but nevertheless have high current densities. In a study of single cells it is necessary that the cells do not overlap and are suitably separated to scan singly or in small clusters, since large area scanning is inefficient. Continuous scanning of the beam is used to minimize any elemental losses, and such losses are monitored. The simultaneous collection of RBS and RFS allows such monitoring of the major specimen constituents H, C, N, and O, with C providing a normalization to mass density variations. Figure 4 shows, for elements C1 and K, the n u m b e r of counts plotted as a function of the C count. Other studies (3,12) have Biological TraceElement Research
VoL 13, 1987
164
O'Brien and Legge
s h o w n C to be a stable element u n d e r proton beam b o m b a r d m e n t and, hence, it essentially provides a time axis i n d e p e n d e n t of beam current fluctuations. As Fig. 4 shows, the elemental concentrations of these ele60
'
I
'
I
,
I
'
I
'
I
,
I
,
I
'
I
,
5O
E Z
10 0
I
0
200
400 600 Channel Number
t
B00
1000 :3
hrs
12.0
"t
10.0
5 eo 3
~o 6.o
_~
4.0
7= 2.0 O0
I
200
I
I 400 600 Chonnel Number
I
i
BOO
10 0 0 =3
hrs
Fig. 4. The X-ray yield for the elements C1 and K as a function of carbon count (effectively time), showing that these elements are stable under proton beam bombardment, and no losses are observed. Biological TraceElementResearch
VoL 13, 1987
Elemental Microanalysis of Individual Blood Cells
165
ments remained constant over the period of irradiation and no losses were observed. One of the major considerations associated with the elemental sensitivity of PIXE analysis is the introduction of contaminants during specimen preparation; conventionally used techniques may not be desirable since they are not developed with such considerations in mind. The X-ray energy spectrum extracted from the displayed cell (Fig. 1) shows a distinct Si peak, which was attributed to contamination from the plastic syringe used for aspiration, and subsequent use of an ultraclean glass syringe removed this peak from the spectrum. A distinctive feature of the two-dimensional maps in Fig. 2 is a "dark band" or "valley" around the cell; this is seen also in Fig. 3. On observation under a light microscope, the cell appeared to have pulled away and separated from the surrounding plasma. The small X-ray contribution from the gap between the cell and the plasma was produced by the thin nylon support. This separation was attributed to the differential rate of cooling between the cell and its surroundings during cryofixation. During data analysis, this band provides a clear outline of the cell for accurate extraction of the X-ray yield from the cell area. Although this band provides a clear cell-plasma boundary, it is difficult to estimate the remaining plasma still adhering to the cell and it is a possible source of error in a quantitative study of elemental concentrations. The alternative to the use of whole blood is washing and resuspending the red cells in suitable buffers to remove the plasma. The problems encountered with this technique are the introduction of contaminant elements from buffers and their adjustment to the correct pH and tonicity to maintain cell integrity, an important requirement for cells of the sensitivity of red blood cells.
CONCLUSION In all the red blood cells examined, the X-ray yield of each element when normalized to C or the cell's effective mass, was observed to be consistent and reproducible for each cell. Thus, the X-ray spectrum of a single red cell provides a reliable "fingerprint" for this blood cell type, especially with regard to its relatively large Fe concentration. The spatial resolution of the probe enabled the observation of the "ring" structure associated with this cell's shape and the probe's sensitivity enabled the detection of Zn in each cell at the 1-ppm level. The data presented in this paper show only the results for a single cell and illustrate the techniques that are currently available for quantitative single cell analysis.
ACKNOWLEDGMENTS The assistance of V. Ivanov (Royal Melbourne Hospital, Victoria) and J. Cerini in the preparation of specimens was much appreciated. D. Biological Trace Element Research
Vol. 13, 1987
166
O'Brien and Legge
N. Jamieson's contribution to the preparation of the three-dimensional maps is gratefully acknowledged. This work was supported in part by the Australian Research Grants Scheme (B82115688).
REFERENCES 1. P. M. O'Brien, G. J. F. Legge, T. R. Bradley, and G. S. Hodgson, Aust. Phys. and Eng. Scs. Med. vol. 5, no. 1, 1982, pp. 30-34. 2. G. J. F. Legge, A. P. Mazzolini, A. F. Rocznick, and P. M. O'Brien, Nucl. Instr. Meth. 197, 191. (1982). 3. A. P. Mazzolini, G. J. F. Legge, and C. K. Pallaghy, Nucl. Instr. Meth. 191, 583 (1981). 4. A. P. Mazzolini, C. K. Pallaghy, and G. J. F. Legge, The New Phytologist, (1985) 100, pp. 483-509. 5. U. Lindh, E. Johansson, and L. Gille, Nucl. Instr. Meth. B3 631 (1984). 6. D. Heck and E. Rokita; IEEE Trans. Nucl. Sci. vol. NS-30, no. 2 1983, 1220-1223. 7. H. Begemann and H.-G. Harwerth, Pr~ktische Hamatologie, 3rd ed., Georg Thieme, Verlag, Stuttgart 1967, p. 47. 8. P. Echlin and R. Moreton, Microprobe Analysis as Applied to Cells and Tissues, (T. Hall, P. Echlin, and R. Kaufman, eds.), Academic, London, New York (1974), pp.159-174. 9. G. J. F. Legge, C. D. McKenzie, and A. P. Mazzolini, J. Micros. 117, 185 (1979). 10. G. J. F. Legge, D. N. Jamieson, P. M. O'Brien and A. P. Mazzolini, Nucl. Instr. Meth. 197, 85 (1982). 11. G. J. F. Legge and I. Hammond, J. Micros. 117, 201 (1979). 12. A. P. Mazzolini, PhD Thesis, University of Melbourne (1983), Unpublished.
Biological Trace Element Research
Vol. 13, 1987
