A simplified method for determination of radioactive iron in whole-blood samples

For studies on iron absorption in man radioisotopes represent an easy and simple tool However, measurement of the orbital electron emitting radioiron, 55Fe, in blood is difficult and insufficiently described in the literature. The present study describes a relatively simple method for simultaneous determination of 55Fe and 59Fe in blood, using a dry-ashing procedure and recrystallization of the remaining iron. The detection limit of the method permits measurements of 0.1 Bq/ml blood thus allowing detection of less than 1% absorption from a 40 kBq dose, which is ethically acceptable in humans. The overall recovery of radioiron from blood is more than 90%, and the coefficient of variation, as judged by the variation in the ratio 55Fe/59Fe is in the order of 4%. Combined with whole-body counting of 59Fe and direct gamma-counting of 59Fe on blood samples, this method represents a sensitive method for studying the intestinal absorption of 55Fe and 59Fe in man and at the same time allows estimation of the amount of radioiron located in the vascular compartment.     

Direct analysis of human blood (mothers and newborns) by energy dispersive X-ray fluorescence

This work is an application of energy dispersive X-ray fluorescence (EDXRF) as analytical technique for trace element determination in human tissues. Potassium (K), calcium (Ca), iron (Fe), copper (Cu), zinc (Zn), bromine (Br), rubidium (Rb) and lead (Pb) were determined directly in blood samples from 66 mothers at delivery after full-term pregnancies. The corresponding 66 cord-blood samples of the newborns were also analysed, in order to find element correlations between maternal and newborn blood at birth. The studied samples were obtained from mothers aged between 15 and 39 years old, the gestational age being between 35 and 41 weeks and the newborns' weight between 2.310 and 4.310 kg. Samples were lyophilised and analysed without any chemical treatment. Very low levels of Pb were found both in maternal and fetal cord blood samples. Cu values ranged from 3 to 13 microg g-1, both for mothers and children. A correlation between Cu and Fe concentrations in maternal and fetal cord blood was found. Zn is considered as one of the key elements in newborn health. Concentrations between 10 and 40 microg g-1 were measured. A positive correlation between Br levels in mothers and children was observed. Positive correlations for mothers were observed between Zn and Rb as well as K and Fe. The corresponding correlations in fetal cord blood samples were not observed, however positive correlations were found between Ca and K; Cu and Fe. The mean concentrations for each element were similar in maternal and in fetal cord blood, except for Cu and Zn, being higher in maternal samples. No correlations between element concentrations and pathologies of the mothers were observed.     

In search of non-random X inactivation: studies of fetal membranes heterozygous for glucose-6-phosphate dehydrogenase.

Extraembryonic membranes and fetal tissues were obtained from 55 specimens of 5--11 weeks conceptual age. The glucose-6-phosphate dehydrogenase (G6PD) electrophoretic phenotype was determined and correlated with that of maternal blood. Fifteen specimens were heterozygous for G6PD A, and for nine of these the maternal allele could be determined. In none of these specimens did the isozyme pattern of the membraneous chorion or chorionic villi differ significantly from that of fetal tissue. We have obtained no evidence of non-random inactivation in extraembryonic membranes of human fetal specimens at this stage of development.     

Consequences of fetomaternal haemorrhage after intrauterine transfusion.

Fetomaternal haemorrhage was studied after 68 consecutive fetal intravascular transfusions performed in 20 patients with Rh isoimmunisation. alpha Fetoprotein concentration was assayed in maternal blood taken before, and immediately after each transfusion and three and 24 hours later. An increase of 50% or more in the concentration in any of the samples after transfusion was considered to indicate fetomaternal haemorrhage. Fetal alpha fetoprotein concentration in blood sampled before transfusion was also assayed and the amount of fetomaternal haemorrhage calculated. Fetomaternal haemorrhage occurred in 21 of 32 patients with an anterior placenta and in six of 36 with a posterior or fundal placenta. The mean estimated volume of haemorrhage was 2.4 ml, which was on average equal to 3.1% of the total fetoplacental blood volume. When the volume of fetomaternal haemorrhage at the first transfusion was greater than 1 ml there was a greater increase in maternal Rh (D) antibody titres and a greater fall in fetal packed cell volume. Sampling of fetal blood should not be routinely done early in patients with Rh isoimmunisation, and intrauterine transfusion should be delayed as long as possible. Sampling sites other than the placental cord insertion reduces the risk of fetomaternal haemorrhage.     

Diurnal variations in plasma arginine vasotocin (AVT) concentrations in the ovine fetus

Previously we have demonstrated the presence of AVT in the blood of fetal sheep. The source is not clear, but AVT has been identified in fetal pineal and pituitary glands. In view of the circadian secretory pattern of the pineal gland, we questioned whether fetal plasma AVT levels might vary diurnally. Plasma samples from five chronically catheterized ovine maternal ewes and fetal lambs 129-135 days' gestation were obtained at 3-19 hourly intervals for 1-2 days (mean +/- S.E.M. = 35 +/- 6 hours. Plasma AVT levels were determined by radioimmunoassay. Results were analyzed by nonlinear curve fitting procedures to relate hormone levels with time of day. Plasma AVT values for maternal ewes did not vary during the day in response to light/dark periods. The curve for mean fetal plasma AVT plotted against time showed oscillations with a period of about 25 hours (p less than 0.05). Peak fetal AVT levels were observed at 1600 hours and minimal levels at 0400 hours. These results indicate that ovine fetal AVT secretion varies diurnally. The site of AVT secretion may be the pineal gland; however, confirmation of this and identification of the physiological stimuli for secretion of fetal plasma AVT require further information.     

The impact of intermittent umbilical cord occlusions on the inflammatory response in pre-term fetal sheep

Fetal hypoxic episodes may occur antepartum with the potential to induce systemic and cerebral inflammatory responses thereby contributing to brain injury. We hypothesized that intermittent umbilical cord occlusions (UCOs) of sufficient severity but without cumulative acidosis will lead to a fetal inflammatory response. Thirty-one chronically instrumented fetal sheep at ～0.85 of gestation underwent four consecutive days of hourly UCOs from one to three minutes duration for six hours each day. Maternal and fetal blood samples were taken for blood gases/pH and plasma interleukin (IL)-1β and IL-6 levels. Animals were euthanized at the end of experimental study with brain tissue processed for subsequent counting of microglia and mast cells. Intermittent UCOs resulted in transitory fetal hypoxemia with associated acidemia which progressively worsened the longer umbilical blood flow was occluded, but with no cumulative blood gas or pH changes over the four days of study. Fetal arterial IL-1β and IL-6 values showed no significant change regardless of the severity of the UCOs, nor was there any evident impact on the microglia and mast cell counts for any of the brain regions studied. Accordingly, intermittent UCOs of up to three minutes duration with severe, but limited fetal hypoxemia and no cumulative acidemia, do not result in either a systemic or brain inflammatory response in the pre-term ovine fetus. However, fetal IL-1B and IL-6 values were found to be well correlated with corresponding maternal values supporting the placenta as a primary source for these cytokines with related secretion into both circulations. Female fetuses were also found to have higher IL-1β levels than males, indicating that gender may impact on the fetal inflammatory response to various stimuli.     

Pharmacological Studies with Lincomycin in Late Pregnancy

The placental transmission of lincomycin was studied in 60 patients in late pregnancy. A peak maternal blood level of 12·5 μg/ml was recorded 45 minutes after injection, and detectable levels were still present up to 42 hours after a single injection. A peak cord blood level of 2·7 μg/ml was recorded 55 minutes after injection; cord blood levels were about a quarter of the maternal blood levels, and in most cases no levels were detectable 24 hours after a single injection. The passage of lincomycin into and out of the liquor was slower and more variable, but some hours after injection the liquor levels were always higher than the maternal or cord blood levels, and detectable levels were still present in the liquor 52 hours after a single injection. Repeated injections did not lead to any significant accumulation of lincomycin. The only side effect was a possible case of neuromuscular block in a mother delivered by caesarean section. No infant was adversely affected.     

Stimulated cytokine production correlates in umbilical arterial and venous blood at delivery

BACKGROUND: Umbilical venous blood is easy to obtain after delivery, and thus has been commonly used in many studies for cytokine analysis. Our aim was to evaluate whether or not induced cytokine production differs after stimulation in umbilical artery and vein whole blood samples, using two different stimulation protocols. The effect of such stimulation on fetal and maternal blood was also evaluated. METHODS: Blood samples from umbilical artery (UA) and vein (UV), and from the mother were collected from 23 women after delivery at term. Concentrations of cytokines (IL-4, IFN-gamma, IL-6 and TNF-alpha) were measured in plasma and whole blood after PMA/ConA and PMA/ionomycin stimulation. RESULTS: Both in maternal and in fetal samples, cytokine concentrations in unstimulated plasma samples were lower than in stimulated samples, except for IL-4 after PMA/ConA stimulation. UA and UV showed similar, average cytokine levels after stimulation and the correlations were high (r=0.68-0.95). Cytokine concentrations were clearly higher in umbilical blood than in maternal blood after stimulation, but not in plasma. Correlations between maternal and umbilical samples after stimulation were generally low (r<0.41). IFN-gamma was not detectable in unstimulated plasma samples. The production of IL-4 and IFN-gamma was more intense after PMA/ionomycin stimulation than after PMA/ConA stimulation. INTERPRETATION OF THE RESULTS: Concentrations of the cytokines examined are similar in blood from the UA and UV. For IL-4 and IFN-gamma, the stimulant used has a significant effect on the level of cytokine expression, and interestingly, the effect is more pronounced on the fetal than on the maternal side.     

Oxygen consumption is maintained in fetal sheep during prolonged hypoxaemia.

Experiments were conducted in 12 chronically-catheterized pregnant sheep to examine the effect of prolonged hypoxaemia secondary to the restriction of uterine blood flow on fetal oxygen consumption. Surgery was performed at 115 days gestation to place a teflon vascular occluder around the maternal common internal iliac artery and for insertion of vascular catheters. Following a 5-day recovery period, uterine blood flow was reduced in 6 animals for 24 hours and in 6 animals, the occluder was not adjusted. Fetal arterial PO2 decreased from 19.9 +/- 2.0 mmHg to 12.8 +/- 2.0 mmHg and 11.0 +/- 2.0 mmHg at 1 and 24 hours respectively in the experimental group and did not change the control group. Fetal pH decreased from 7.34 +/- 0.01 to 7.25 +/- 0.03 and 7.29 +/- 0.02 at 1 and 24 hours of hypoxaemia respectively. Fetal arterial lactate concentrations remained elevated throughout the experimental period with maximum concentrations of 6.6 +/- 2.1 mmol/l being present at 4 hours compared to 1.3 +/- 0.2 mmol/l during the control period. Umbilical blood flow increased from 186 +/- 19 ml/min/kg to 251 +/- 39 ml/min/kg at 1 h of hypoxaemia and returned to 191 +/- 21 ml/min/kg at 24 h. In association with the progressive fall in oxygen delivery to the fetus, oxygen extraction increased from 0.33 +/- 0.04 to 0.43 +/- 0.04 and 0.54 +/- 0.05 at 1 and 24 hours, respectively. Overall oxygen consumption by the fetus remained unchanged from control values.(ABSTRACT TRUNCATED AT 250 WORDS)     

Effects on maternal and fetal steroid concentrations of induction of parturition in the sheep by inhibition of 3 beta-hydroxysteroid dehydrogenase

Changes in circulating steroid hormones, the incidence of myometrial contractions, and the onset of labour were all monitored after administration of the 3 beta-hydroxysteroid dehydrogenase inhibitor, epostane, to chronically catheterized ewes and fetuses near term. In all animals the drug induced delivery 33-36 h after injection or infusion into the ewe with the birth of live healthy lambs which showed normal subsequent development. Epostane induced immediate, permanent falls in both maternal and fetal plasma progesterone concentrations, accompanied by increased PGF metabolite concentrations in the uterine vein beginning 15 min after treatment. Of the other hormonal changes observed, the most striking was the pronounced drop in both maternal and fetal plasma cortisol. In the fetus this fall was followed by increasing concentrations of circulating ACTH which eventually restored the cortisol levels. By 12-24 h after epostane a substantial overshoot had occurred and at 27-30 h the fetal plasma cortisol concentrations were as high as those seen during normal parturition at term. No significant changes in maternal plasma oestradiol-17 beta could be detected after epostane treatment or during labour. The incidence of slow myometrial contractions increased significantly during the second 3-h period after epostane, although their duration did not change. Contraction patterns typical of first stage labour were seen from 20 to 24 h. These results show that epostane may be used as a safe, predictable inducing agent in sheep if given 6-10 days before term; the lambs showed no signs of prematurity despite their lowered plasma cortisol concentrations which persisted for some hours before labour was induced.     

Prostaglandin F and E levels in the conceptus, uterus and plasma during early pregnancy in the ewe

Concentrations of prostaglandins E and F (PGE and PGF) were measured in the embryo or fetus, extra embryonic or fetal membranes (membranes), intercaruncular and caruncular endometrium and plasma collected from uterine and ovarian arterial and venous vessels from separate groups of ewes laparotomized at 5 day intervals from day 10 to day 55 of pregnancy. Our purpose was to investigate the role of prostaglandins E and F in the maternal recognition of pregnancy, implantation and early placental function. Our data suggest that the initial maintenance of the corpus luteum in the pregnant ewe does not involve a reduction in PGF production, compared to pregnant ewes; but a change in the pattern of PGF secretion. This is accompanied by an elevation in PGE production of similar magnitude to that observed in non pregnant ewes. The extra embryonic/fetal membranes appear to be the major source of elevated PGF levels in the maternal circulation prior to day 30 of pregnancy. Between days 35 and 55 of gestation the rising PGF levels in maternal serum probably come from the fetus. Over the same period PGE levels rise in the fetus and intercaruncular endometrium, but PGE secretion into the maternal circulation is not enhanced. A role for PGF and PGE in fetal, placental and uterine growth is suggested; placental and uterine endocrine function may also be targets.     

Prostaglandin F and E levels in the conceptus, uterus and plasma during early pregnancy in the ewe

Concentrations of prostaglandins E and F (PGE and PGF) were measured in the embryo or fetus, extra embryonic or fetal membranes (membranes), intercaruncular and caruncular endometrium and plasma collected from uterine and ovarian arterial and venous vessels from separate groups of ewes laparotomized at 5 day intervals from day 10 to day 55 of pregnancy. Our purpose was to investigate the role of prostaglandins E and F in the maternal recognition of pregnancy, implantation and early placental function. Our data suggest that the initial maintenance of the corpus luteum in the pregnant ewe does not involve a reduction in PGF production, compared to pregnant ewes; but a change in the pattern of PGF secretion. This is accompanied by an elevation in PGE production of similar magnitude to that observed in non pregnant ewes. The extra embryonic/fetal membranes appear to be the major source of elevated PGF levels in the maternal circulation prior to day 30 of pregnancy. Between days 35 and 55 of gestation the rising PGF levels in maternal serum probably come from the fetus. Over the same period PGE levels rise in the fetus and intercaruncular endometrium, but PGE secretion into the maternal circulation is not enhanced. A role for PGF and PGE in fetal, placental and uterine growth is suggested; placental and uterine endocrine function may also be targets.     

Fetal toxicity of cadmium chloride: the pharmacokinetics in the pregnant Wistar rat

The pharmacokinetics of a single subcutaneous injection of 40 mumol/kg CdCl2 in pregnant rats on day 18 were studied during an 18-hour time period. Previous studies demonstrated that this dose of CdCl2 induced fetal death, placental necrosis, and a reduction of uteroplacental blood flow without maternal lethality; direct fetal injections of CdCl2 indicated that the site of toxic action was not in the fetus. Cadmium was rapidly absorbed from the intrascapular subcutaneous depot with the highest blood levels occurring within 5 minutes of injection. The release from the depot appeared to be multiphasic with both rapid and slow absorption components observed in the blood. Uptake of cadmium was greatest in the liver, followed by the placenta, kidney, and pancreas. The best fit to the data was obtained by assuming the existence of two pools of cadmium in the organs: one freely exchangeable with the blood and the other nonexchangeable. Deviations from model predictions were observed for the placenta and adrenals; these deviations are consistent with the concomitant diminished blood flow to these organs. Cadmium uptake by the fetus was also investigated, and the results support the hypothesis that the placenta is relatively impermeable to the toxicant. It is concluded that the rapid and extensive accumulation of cadmium by the placenta may play a role in the development of early placental cellular damage and the eventual induction of fetal death through uteroplacental dysfunction.     

Further characterization of the distribution and metabolism of nitrofen in the pregnant rat

Nitrofen (2,4-dichloro-4'-nitrodiphenyl ether) is an herbicide with potent teratogenic activity in rodent species. The present study was an extension of previous efforts to characterize the distribution and metabolism of nitrofen in pregnant rats. Following a single p.o. exposure to radiolabeled compound on day 10 of pregnancy, maternal and embryonic tissues were collected at intervals from 1.5 to 72 hours. Radioactivity was accumulated and retained in maternal fat for over 72 hours. Peak levels were reached in other maternal organs at 3-12 hours. The half-life in maternal plasma was estimated to be 42 hours. Radioactivity was first detected in the embryonic compartment at 3 hours and continued to increase through the 72-hour time point. HPLC analysis indicated that the parent compound is initially deposited in maternal fat and after 48 hours redistributes to other maternal organs and to the embryo. The 5-hydroxy derivative was the major nitrofen metabolite found in maternal tissues, while the 4'-amino and 4'-acetylamine derivatives were found at lower levels and all exhibited single-phase kinetics. The parent compound alone was found in the embryo, and levels increased gradually as nitrofen redistributed from the fat at 48 hours after exposure. The results of this and other studies of nitrofen metabolism in pregnant rats suggest that its teratogenicity is not mediated via generation of mutagenic intermediates through nitroreduction of the parent compound. Rather, the embryo is exposed to the parent compound alone and appears to be a deep compartment for accumulation of nitrofen.     

The role of fetal urinary excretion in the transfer of ethanol into amniotic fluid after maternal administration of ethanol to the near-term pregnant ewe

The objective of this study was to determine whether fetal urinary excretion is a major route of ethanol transfer into the amniotic fluid surrounding the fetus following maternal administration of ethanol. Conscious instrumented pregnant ewes between 130 and 137 days' gestation (term, 147 days) with (n = 3) or without (n = 3) a catheter in the fetal bladder were administered 1 g ethanol/kg maternal body weight as a 1-h maternal intravenous infusion. Maternal blood, fetal blood, and amniotic fluid samples were collected at selected times, and fetal urine was collected continuously from the bladder-cannulated fetus during the 14-h study for the determination of ethanol concentrations. Fetal urinary excretion of ethanol occurred, and the total amount of ethanol excreted represented 0.30 +/- 0.07 (SD)% of the maternal ethanol dose. The renal clearance of ethanol by the fetus was 0.43 +/- 0.06 mL/min. The pharmacokinetics of ethanol in the maternal-fetal unit and the amniotic fluid for the bladder-cannulated fetal preparation were similar to the data for the nonbladder-cannulated preparation. The data indicate that fetal urinary excretion of ethanol is a secondary route of ethanol transfer into the amniotic fluid. It would appear that diffusion of ethanol across membranes from the maternal and fetal circulations is a major route of ethanol transfer into this intrauterine compartment.     

Carboxy- and oxyhemoglobin in pregnant ewe and fetus after inhalation of marijuana, marijuana placebo and tobacco cigarette smoke

We measured carboxyhemoglobin (HbCO) and oxyhemoglobin (HbO2) percent saturations and blood gases in four near-term pregnant ewes and their fetuses, during and for 6 hours after 9-12 minutes of smoke inhalation from one high-potency marijuana cigarette (M), a marijuana placebo cigarette (P), and a reference tobacco cigarette (T). Maternal HbCO reached maximum levels at or soon after the exposure (M, 2.8%; P, 3.5%; T, 6.3% above baseline) and fell to baseline values by 6 hours. Fetal HbCO rose slowly reaching a plateau at 3 hours (M, 0.7%; P, 1.1%; T, 2.0% above baseline) which was maintained for at least three additional hours. Reductions in maternal and fetal HbO2 after exposure to marijuana placebo and reference tobacco cigarettes reflected these rises in HbCO. After exposure to marijuana cigarettes, however, fetal HbO2 dropped precipitously by 17% of baseline and showed a prolonged rate of return to presmoking HbO2 levels. Although P exposure caused a greater change in HbCO in the fetus than did M, it had a less-profound effect on fetal oxygenation.     

Cadmium exposure on day 12 of gestation in the Wistar rat: distribution, uteroplacental blood flow, and fetal viability

The effects of cadmium exposure (40 mumole CdCl2/kg, s.c.) on day 12 of gestation were evaluated in the Wistar rat. At 16-18 hours following such cadmium exposure, blood flow (as determined by radiolabeled microspheres) to the chorioallantoic placenta (CAP) was significantly reduced by 35%; at 24-26 hours, blood flow to the CAP had returned to control levels and was still unaffected at 38-43 hours. Uterine blood flow was not significantly altered at any of these timepoints. Between 16-18 and 24-26 hours after cadmium exposure, the concentration of cadmium in the placenta decreased significantly, while total cadmium content did not change. By 38-43 hours after cadmium exposure, total cadmium content of the placenta had increased significantly, although cadmium concentration was unchanged. There were no adverse effects on fetal viability or growth, as determined on day 20 of gestation. In sharp contrast, near-term (day 18) exposure to 40 or 50 mumole CdCl2/kg (s.c.) resulted in 53% and 82% mean incidences of fetolethality, respectively, within 24 hours. Administration of 50 mumole CdCl2/kg (sc) on day 12 also had no effect on fetal growth but resulted in increased fetolethality (12%). Thus midgestational cadmium exposure and its accompanying alterations in placental blood flow do not compromise fetal viability or growth. The differential response to cadmium at mid- and late gestation, in terms of fetolethality, is not due to maternal cadmium dose.     

The effect of maternal nutrition level during the periconception period on fetal muscle development and plasma hormone concentrations in sheep

The effect of maternal nutrition level during the periconception period on the muscle development of fetus and maternal–fetal plasma hormone concentrations in sheep were examined. Estrus was synchronized in 55 Karayaka ewes and were either fed ad libitum (well-fed, WF, n=23) or 0.5×maintenance (under-fed, UF, n=32) 6 days before and 7 days after mating. Non-pregnant ewes (WF, n=13; UF, n=24) and ewes carrying twins (WF, n=1) and female (WF, n=1; UF, n=3) fetuses were removed from the experiment. The singleton male fetuses from well-fed (n=8) and under-fed (n=5) ewes were collected on day 90 of gestation and placental characteristics, fetal BWs and dimensions, fetal organs and muscles weights were recorded. Maternal (on day 7 after mating) and fetal (on day 90 of pregnancy) blood samples were collected to analyze plasma hormone concentrations. Placental characteristics, BW and dimensions, organs and muscles weights of fetuses were not affected by maternal feed intake during the periconception period. Maternal nutrition level did not affect fiber numbers and the muscle cross-sectional area of the fetal longissimus dorsi (LD), semitendinosus (ST) muscles, but the cross-sectional area of the secondary fibers in the fetal LD and ST muscles from the UF ewes were higher than those from the WF ewes (P<0.05). Also, the ratio of secondary to primary fibers in the ST muscle were tended to be lower in the fetuses from the UF ewes (P=0.07). Maternal nutrition level during the periconception period did not cause any significant changes in fetal plasma insulin and maternal and fetal plasma IGF-I, cortisol, progesterone, free T3 and T4 concentrations. However, maternal cortisol concentrations were lower while insulin concentrations were higher in the WF ewes than those in the UF ewes (P<0.05). These results indicate that the reduced maternal feed intake during the periconception period may alter muscle fiber diameter without affecting fiber types, fetal weights and organ developments and plasma hormone concentrations in the fetus.     

Search for the optimal fetal cell antibody: results of immunophenotyping studies using flow cytometry

Fetal nucleated cells circulating in maternal peripheral blood are a noninvasive source of fetal DNA for prenatal genetic diagnoses. The successful isolation of fetal cells from maternal blood depends upon identification of differences between fetal and maternal cell surface antigen expression. To our best knowledge, a monoclonal antibody that binds only fetal blood cells has not yet been identified. We studied antigens recognized by six different monoclonal antibodies for their biologic expression on fetal blood cells as a function of gestational age, and compared their ability to bind fetal but not maternal cells. The results suggest a relationship between gestational age and nucleated cell surface antigen expression. The monoclonal antibodies FB3-2, H3-3, CD71 and 2-6B/6 are suitable reagents for first or early second trimester fetal cell isolation, although FB3-2 and H3-3 are more specific for fetal cells due to significantly lower expression of these antigens on maternal mononuclear cells. The observation that samples from fetuses with chromosome abnormalities or multiple structural anomalies express higher levels of these antigens indicates that these reagents will potentiate the detection of abnormal fetal cells in maternal blood samples. Received: 23 November 1996 / Accepted: 13 February 1997