Diffuse form of dysembryoplastic neuroepithelial tumour: the histological and immunohistochemical features of a distinct entity showing transition to dysembryoplastic neuroepithelial tumour and ganglioglioma.

AIMS: A diffuse variant of dysembryoplastic neuroepithelial tumour (dDNT) has previously been described, which although composed of oligodendroglia-like cells (OLC), astrocytes and mature neurones, lacks the multinodularity and 'specific component' of typical DNT. The dDNT poses a significant challenge to the neuropathologist. This study was undertaken to further characterize the histological and immunohistochemical features of dDNT. MATERIALS AND METHODS: Review of our archived material from epilepsy surgery identified 16 cases, in which features of dDNT predominated. Their histological and immunohistochemical features, including CD34 and nestin immunohistochemistry, were analysed. RESULTS: Seven cases had the characteristics of pure dDNT. A further two cases of dDNT showed extension into the white matter with occasional dysplastic neurones. Two additional cases had similar features but with the presence of either single, or multiple small nodular clusters of OLC, in keeping with transition to classical DNT. Five cases showed ganglioglioma-like areas, of which three cases had micronodule formation but with predominant dDNT pattern. In all the cases the dDNT areas showed strong CD34 and less intense nestin immunoreactivity and microglial activation highlighting the full extent of the lesions. There was variable overlap between CD34 and nestin positivity within the micronodular and/or ganglioglioma-like areas. CONCLUSIONS: Immunoreactivity for CD34 and nestin characterizes the dDNT and helps to distinguish it from other lesions associated with epilepsy. Histological evidence indicative of transition of dDNT to other forms of DNT and ganglioglioma suggests that dDNT might be an early histogenetic form of these glioneuronal tumours.

Glucose production in the human placenta.

Glucose transfer from mother to fetus by placental facilitated diffusion is the dominant mechanism by which the fetus acquires glucose. In small for gestational age pregnancies, fetal glucose concentrations tend to be lower than normal and this persists following delivery. GLUT1 is the major glucose transporter in human placenta but there is no evidence of GLUT1 deficiency as a cause of the lower fetal glucose concentration in small for gestational age pregnancy. The physiological and pathological roles of the other glucose transporters (and there are 14 currently described) are unknown. In recent years, the possibility has been raised that the placenta is itself capable of supplying glucose for fetal needs. This hypothesis derived from glucose isotope studies in normal pregnancy, where dilution of glucose isotope was demonstrated in blood samples taken from the fetal circulation during intravenous infusion of glucose isotope in the mother. Although other gluconeogenic enzymes were known to be present, the placenta was previously considered incapable of glucose secretion because it lacked functional glucose-6-phosphatase. Recent studies, however, have suggested that specific glucose-6-phosphatase may be present in placenta but it may be the product of a different gene from conventional hepatic glucose-6-phosphatase. The presence of the specific transporters necessary for glucose-6-phosphatase activity is currently being investigated. The role of placental glucose secretion in normal and growth-restricted pregnancies is an area of current study.

Increased leucine turnover in women during the third trimester of uncomplicated pregnancy.

Plasma amino acid concentrations decrease in pregnancy despite increased requirements for both maternal and fetal protein accretion. There are few published data on changes in amino acid turnover in pregnancy. The aim of this study is to test the hypotheses that (1) whole body nonoxidative leucine disposal (NOLD) is higher and (2) whole body oxidative leucine disposal (OLD) is lower in the third trimester of pregnancy than in the nonpregnant state. After an overnight fast 8 pregnant women between 33 and 35 weeks gestation had a primed infusion of labeled [1-(13)C] leucine and a prime dose of NaH(13)CO(3). Carbon dioxide production was measured using indirect calorimetry. Gas chromatography-mass spectrometry was used to determine (13)CO(2) enrichment of expired air and oxidative and nonoxidative leucine turnover by measuring (13)C keto-isocaproate plasma enrichment, which reflects intracellular leucine enrichment. Women acted as their own controls after the puerperium. Whole body leucine turnover expressed per unit body weight was increased in pregnancy (median [interquartile range or IQR]: pregnant = 103.1 [14.9] v nonpregnant = 90.1 [10.9] micromol/kg/h). The mean (+/-SD) of the differences was 11.4 +/- 5.6 micromol/kg/h, P =.0006. NOLD was increased in pregnancy (pregnant = 86.8 [10.1] v nonpregnant = 73.3 [9.5] micromol/kg/h). The mean (+/-SD) of the differences was 10.6 +/- 5.4 micromol/kg/h, P =.0008. OLD was not significantly altered in pregnancy (pregnant = 17.3 [4.5] v nonpregnant = 15.91 [2.4] micromol/kg/h). The mean (+/-SD) of the differences was 0.84 +/- 1.94 micromol/kg/h, P =.26. In conclusion, women have significantly higher NOLD in the third trimester of pregnancy than when not pregnant but there are no significant changes in OLD.

Metabolic effects of Troglitazone in patients with diet-controlled type 2 diabetes.

BACKGROUND: In order to study the mechanisms of action of Troglitazone (TGZ) in vivo in Type 2 diabetes, its effects were studied on glucose metabolism, lipolysis and very low-density lipoprotein (VLDL) apolipoprotein B100 (apoB) kinetics. MATERIALS AND METHODS: A placebo-controlled, double-blind study was performed in 24 diet-treated patients randomized to receive TGZ 600 mg day(-1), TGZ 200 mg day(-1) or placebo for 8 weeks. Glucose and glycerol turnover were assessed after an overnight fast, and during sequential low-dose insulin infusions (0.01 U kg(-1) h(-1) followed by 0.015 U kg(-1) h(-1)) using 6,6-2H Glucose and 1,2,3-2H Glycerol. Very low-density lipoprotein apoB secretion was measured using l-13C-leucine, monitoring isotopic enrichment by gas chromatography-mass spectrometry. Treatment effects were analyzed by analysis of covariance, adjusting for baseline. RESULTS: Therapy resulted in a significant group differences in fasting plasma glucose adjusting for baseline (P=0.039). This was most evident at TGZ 600 mg daily [glucose decrease from (mean +/- SD) 9.2 +/- 2.7 to 6.6 +/- 0.9 mmol L(-1)]. HbA1c and insulin levels did not change significantly. Plasma nonesterified fatty acid (NEFA) levels decreased (P=0.045), most evidently at TGZ 200 mg daily, but glycerol was not significantly affected. Although no significant effects were observed on VLDL apoB or triglyceride concentrations, there were treatment differences in the absolute secretion rate of VLDL apoB of borderline (P=0.056) statistical significance, with a decrease observed at TGZ 600 mg daily [geometric mean, SD range, 0.94 (0.41-2.15) to 0.40 (0.14-1.13 mg kg(-1) h(-1))]. Very low-density lipoprotein apoB fractional secretion rate and pool size were unaffected. The VLDL triglyceride: apoB molar ratio differed between treatment groups (P=0.013), being higher in the TGZ 600 mg group [5714 (4128-7741) to 8092 (5669-11552)]. Neither glucose nor glycerol rates of appearance were significantly altered by TGZ and nor did TGZ affect their suppression by insulin. DISCUSSION: The PPARgamma agonist, troglitazone, decreases fasting glucose and NEFA levels in diet-treated Type 2 diabetes. It may also decrease VLDL particle secretion. These effects would be considered beneficial. The biological importance of the increase in VLDL-triglyceride enrichment warrants further study.

Hypopituitarsim is associated with triglyceride enrichment of very low-density lipoprotein.

The dyslipidemia associated with hypopituitarism may contribute to increased vascular mortality. The atherogenic potential of lipoproteins is determined not only by concentration but also by their composition. We therefore studied very low-density lipoprotein composition and apolipoprotein B kinetics in 16 hypopituitary subjects and 16 controls. Hypopituitarism was associated with reduced high-density lipoprotein cholesterol (0.98[0.82-1.18] vs. 1.35[1.15-1.41] mmol/liter, P < 0.001) and increased triglyceride concentrations (1.64[1.09-2.77] vs. 1.12[0.66-1.67] mmol/liter, P = 0.01). Total (P = 0.76) and low-density lipoprotein cholesterol (P = 0.56) concentrations were similar. Very low-density lipoprotein- triglyceride was significantly increased (1.48[1.02-2.55] vs. 0.9[0.31-2.30] mmol/liter, P = 0.004), but very low-density lipoprotein cholesterol levels were similar (P = 0.93). The molar ratios of very low-density lipoprotein-triglyceride:apolipoprotein B (6193[4283-9566] vs. 3599[3188-6854], P = 0.005) and very low-density lipoprotein-triglyceride:cholesterol (2.8[1.98-3.78] vs. 1.6[1.44-2.80], P < 0.003) were significantly increased; very low-density lipoprotein-cholesterol:apolipoprotein B molar ratios (P = 0.93) were similar. Very low-density lipoprotein apolipoprotein B fractional synthetic rate (a measure of apolipoprotein B catabolism, P = 0.42) and pool size (P = 0.63) were similar. The very low-density lipoprotein apolipoprotein B absolute synthetic rate (a measure of apolipoprotein B synthesis) tended to be higher in hypopituitarism (17.7[2.91-19.50] vs. 26.6[19.64-28.05] mg/kg per day, P = 0.24) but failed to reach statistical significance. The absolute synthetic rate, and hence very low-density lipoprotein production, correlated with very low-density lipoprotein triglyceride:apolipoprotein B ratio (P = 0.02, Rs = 0.63), suggesting that triglyceride enrichment of very low-density lipoprotein is important in the mechanism underlying very low-density lipoprotein overproduction in hypopituitarism. Because triglyceride-enriched lipoproteins are proatherogenic, this may contribute to the vascular mortality observed in hypopituitarism. The reasons for these observations are unknown; GH deficiency or routine endocrine replacement may be important.

Effects of growth hormone treatment on very-low density lipoprotein apolipoprotein B100 turnover in adult hypopituitarism.

Adult hypopituitarism is associated with hyperlipidemia, mainly due to an increase of very-low-density lipoprotein (VLDL) and low-density lipoprotein (LDL) levels. Recent studies have shown that such patients exhibit increased hepatic secretion of VLDL apolipoprotein B100 (VLDL apo B100). To examine the effects of growth hormone (GH) replacement on VLDL apo B100 turnover, 13 GH-deficient hypopituitary patients (8 women and 5 men; aged 47 +/- 3 years, mean +/- SEM; body mass index [BMI], 30 +/- 2 kg/m2) entered a double-blind placebo-controlled study for 6 months (GH 0.125 IU/kg/wk for 4 weeks, and then 0.25 IU/kg/wk). GH was subsequently used in all patients for a further 6 months. A 6-hour [1-13C] leucine infusion was administered at baseline and at 6 months. The secretion rate of VLDL apo B100 was derived by kinetic analysis following quantitation of isotopic enrichment by gas chromatography/mass spectrometry. The GH-treated group (6 patients) demonstrated a similar fractional secretion rate (FSR) for VLDL apo B100 at 0 and 6 months. The pool size and absolute secretion rate (ASR) also were unaffected significantly by GH therapy. No significant changes were observed in the placebo group (7 patients). Treatment with GH for 6 months caused an increase in the high-density lipoprotein (HDL) cholesterol concentration (13 patients, 1.27 +/- 0.13 v 1.16 +/- 0.10 mmol/L, respectively, P = .05), whereas total cholesterol and triglyceride concentrations did not change. Nonesterified fatty acids (NEFAs) increased during GH therapy (471 +/- 43 micromol/L at 6 months v 349 +/- 49 micromol/L at baseline, P < .0005). The data suggest that GH does not affect VLDL apo B100 turnover in a significant way.

Suppression of the nocturnal rise in growth hormone reduces subsequent lipolysis in subcutaneous adipose tissue.

BACKGROUND: The aim of this study was to examine the effect of the nocturnal rise in growth hormone (GH) concentration on lipolysis in adipose tissue the following morning. METHODS: Eight healthy subjects were studied on two occasions (control vs. suppression of GH secretion) and six were studied on a third occasion (control vs. replacement of GH). Lipolysis in the whole body was assessed by measurement of systemic glycerol turnover. Lipid metabolism in the subcutaneous adipose tissue of the anterior abdominal wall was studied by measurement of arterio-venous differences. RESULTS: Suppression of the nocturnal rise in GH did not affect systemic glycerol turnover. However, in subcutaneous abdominal adipose tissue it led to a significant reduction in the veno-arterial differences in nonesterified fatty acid (NEFA, P = 0.041) and glycerol (P = 0. 014) concentrations, reflecting a reduction in intracellular lipolysis (P = 0.011). Although arterialized plasma triacylglycerol (TG) concentrations were reduced in the absence of the nocturnal GH pulse, the extraction of TG in subcutaneous abdominal adipose tissue remained unchanged. CONCLUSION: We conclude that the normal nocturnal rise in plasma GH concentration leads to site-specific regulation of lipolysis in adipose tissue on the following day, with preferential fat mobilization from central depots.

Glucose production by the human placenta in vivo.

The human placenta transports glucose by facilitated diffusion down a concentration gradient from mother to fetus. It has previously been considered incapable of glucose synthesis. However, recent work has demonstrated the presence in placental tissue of glucose-6-phosphatase, which is required for the final step in the synthesis of glucose. Following continuous intravenous infusion into the maternal circulation of the stable isotope, 6,6-(2)H(2)glucose, during elective caesarean section, we have observed isotope dilution in the umbilical vein, without further dilution in the umbilical artery. Using a mathematical model containing maternal, placental and fetal compartments, the data were compatible with the release of glucose by the placenta. We conclude that the human placenta at term can produce glucose.

Very-low-density lipoprotein apolipoprotein B100 kinetics in adult hypopituitarism.

Hypopituitarism is associated with hyperlipidemia, the mechanisms of which are not fully known. One possible mechanism is an increased hepatic secretion of very-low-density lipoprotein (VLDL) apolipoprotein B100 (apo B100). To investigate this, 13 hypopituitary patients (seven women and six men; age, 46 +/- 3 years [mean +/- SEM]; body mass index [BMI], 29 +/- 2 kg/m2) and 13 matched controls (seven women and six men; age, 43 +/- 3 years; BMI, 28 +/- 2 kg/m2) were investigated in a stable-isotope study. [1-(13)C]leucine (1 mg/kg body weight) was administered, followed by a continuous 6-hour infusion of [1-(13)C]leucine (at a rate of 1 mg/kg/h). Patients had a similar fractional secretion rate (FSR) of VLDL apo B100 versus controls (0.37 +/- 0.05 v 0.38 +/- 0.06 pools/h, respectively), but they had a significantly larger pool size (3.4 +/- 0.3 v 1.9 +/- 0.3 mg/kg) and higher absolute secretion rate ([ASR] 27.8 +/- 2.9 v 16.0 +/- 2.5 mg/kg/d). The increase in hepatic VLDL production may explain the lipid abnormalities found in hypopituitarism. Fasting circulating nonesterified fatty acids (NEFAs) were decreased in the patients (284 +/- 26 v 664 +/- 92 micromol/L, P < .001) despite the increase in VLDL secretion. An inverse relationship was observed between the NEFA level and VLDL apo B100 FSR in the patients (r(s) = -.85, P < .005).

Reproducibility and comparability of insulin sensitivity indices measured by stable-label intravenous glucose tolerance test.

We have investigated the reproducibility of (1) insulin sensitivity (S*I) and glucose effectiveness (S*G) as measured by the stable-label (one compartment) minimal model, and (2) insulin sensitivity (S*Ib), plasma clearance rate (PCR), basal hepatic output (HGOb), and total hepatic glucose output (HGO0-240) as measured by the novel stable-label two compartment model of glucose disappearance during labelled intravenous glucose tolerance test (IVGTT) using 6,6-(2)H-glucose. Ten normal male subjects were studied on two occasions one week apart. Both models provided estimates of all indices with acceptable precision (CV of parameter estimates < or =50%). The within subject CVs of S*I and S*Ib were comparable (17% vs 19%) as were the within subject CVs of S*G and PCR (13% vs 16%). A highly significant linear relationship was observed between S*Ib and S*I (0.303 +/- 0.046 ml kg(-1) min(-1) per mU l(-1) vs 13.04 +/- 1.89 10(-4) min(-1) per mU l(-1), y = 0.0037 x + 0.0002, r = 0.90, p < 0.001; mean +/- SE), but not between PCR and S*G (1.98 +/- 0.15 ml kg(-1) min(-1) vs 0.0089 +/- 0.0005 min(-1), rs = 0.34, NS). The two compartment model provided a plausible time-profile of hepatic glucose output during IVGTT, reproducible estimates of HGOb (1.96 +/- 0.18 mg kg(-1) min(-1), 15%; mean +/- SE, within subject CV), and a highly reproducible HGO0-240 (7%; within subject CV). We conclude that the stable-label (one compartment) minimal model and the stable-label two compartment model provide reproducible estimates of parameters of glucose kinetics in normal subjects. Insulin sensitivity indices estimated by the two models are strongly linearly related.

Effects of fish oil supplementation on apolipoprotein B100 production and lipoprotein metabolism in normolipidaemic males.

OBJECTIVE: To evaluate the effects of four weeks of fish oil supplementation on apolipoprotein B100 production and lipoprotein metabolism in normolipidaemic males. DESIGN AND SUBJECTS: Very low density lipoprotein (VLDL) apolipoprotein B100 (apoB100) kinetics in ten healthy, white males, aged 22-43 y (mean 32 y) were investigated using 13C-leucine technique and gas chromatography-mass spectrometry before and after fish oil supplementation. INTERVENTION: All subjects received 10 g (1.8 g EPA, 1.2 g DHA)/d of fish oil concentrate for four weeks. RESULTS: Fish oil supplementation resulted in a decrease of total plasma VLDL (mean +/- s.d. 1.11 +/- 0.41 vs 0.87 +/- 0.28 mmol/l, P < 0.05) and triacylglycerol concentrations (0.74 +/- 0.27) vs 0.48 +/- 0.21 mmol/l, P < 0.01). VLDL apoB100 pool size was decreased without alteration of the fractional synthetic rate but a significant decrease of apoB100 production (2.23 +/- 0.90 vs 1.54 +/- 0.52 mg/dl/h, P < 0.02). Following fish oil supplementation plasma concentrations of glucose and insulin as well as lipoprotein and hepatic lipase activities were unchanged. Fasting plasma concentrations of non-esterified fatty acid (NEFA) were decreased (0.45 +/- 0.12 vs 0.33 +/- 0.10 mmol/l, P < 0.05). CONCLUSIONS: Dietary supplementation with fish oil in healthy males results in decreased VLDL-triacylglycerol concentrations through a decrease in VLDL particle synthesis. The decrease in NEFA substrate supply also contributes.

Effects of physiological hypercortisolemia on the regulation of lipolysis in subcutaneous adipose tissue.

Cortisol is known to increase whole body lipolysis, yet chronic hypercortisolemia results in increased fat mass. The main aim of the study was to explain these two apparently opposed observations by examining the acute effects of hypercortisolemia on lipolysis in subcutaneous adipose tissue and in the whole body. Six healthy subjects were studied on two occasions. On one occasion hydrocortisone sodium succinate was infused i.v. to induce hypercortisolemia (mean plasma cortisol concentrations, 1500 +/- 100 vs. 335 +/- 25 nmol/L; P < 0.001); on the other occasion (control study) no intervention was made. Lipolysis in the s.c. adipose tissue of the anterior abdominal wall was studied by measurement of arterio-venous differences, and lipolysis in the whole body was studied by constant infusion of [1,2,3-2H5]glycerol for measurement of the systemic glycerol appearance rate. Hypercortisolemia led to significantly increased arterialized plasma nonesterified fatty acid (NEFA; P < 0.01) and blood glycerol concentrations (P < 0.05), with an increase in systemic glycerol appearance (P < 0.05). However, in s.c. abdominal adipose tissue, hypercortisolemia decreased veno-arterialized differences for NEFA (P < 0.05) and reduced NEFA efflux (P < 0.05). This reduction was attributable to decreased intracellular lipolysis (P < 0.05), reflecting decreased hormone-sensitive lipase action in this adipose depot. Hypercortisolemia caused a reduction in arterialized plasma TAG concentrations (P < 0.05), but without a significant change in the local extraction of TAG (presumed to reflect the action of adipose tissue lipoprotein lipase). There was no significant difference in plasma insulin concentrations between the control and hypercortisolemia study. Site-specific regulation of the enzymes of intracellular lipolysis (hormone-sensitive lipase) and intravascular lipolysis (lipoprotein lipase) may explain the ability of acute cortisol treatment to increase systemic glycerol and NEFA appearance rates while chronically promoting net central fat deposition.

Constant infusion and bolus injection of stable-label tracer give reproducible and comparable fasting HGO.

We have investigated the reproducibility of fasting hepatic glucose output (HGO) estimates by use of isotope dilution methodology of stable-label tracers. Six normal subjects were studied on two occasions 1 wk apart. After an overnight fast, the subjects received a bolus injection of 7 mg/kg of [U-13C]glucose and, simultaneously, a primed constant infusion of 0.05 mg.kg-1.min-1 of [6,6(-2)H]glucose. The bolus injection provided one estimate of HGO (HGOBOL), and the constant infusion provided two estimates of HGO, namely, HGO at 2 h (HGOINF2) and HGO at 4 h (HGOINF4), both with the assumption of steady-state conditions. All estimates were similar in value; HGOBOL was highest, followed by HGOINF2 and HGOINF4 [2.30 +/- 0.11 (SE), 2.17 +/- 0.12, and 2.01 +/- 0.13 mg.kg-1.min-1]. The constant infusion gave highly reproducible results. In the case of HGOINF2, the within-subject coefficient of variation (CV) was only 3% compared with 5% of HGOINF4. The reproducibility of HGOBOL was comparable with the within-subject CV of 7%. We conclude that a constant infusion and a bolus injection of stable-label tracer give reproducible and comparable estimates of HGO.

Motor and cognitive function in Lewy body dementia: comparison with Alzheimer's and Parkinson's diseases.

OBJECTIVE: Motor and cognitive function were compared in patients with Lewy body dementia, Parkinson's disease, or Alzheimer's disease, to identify features that may be clinically useful in differentiating Lewy body dementia from Alzheimer's disease and Parkinson's disease. METHODS: A range of neuropsychological function and extrapyrimidal signs (EPS) was assessed in 16 patients with Lewy body dementia, 15 with Parkinson's disease, 25 with Alzheimer's disease, and 22 control subjects. RESULTS: The severity of total motor disability scores increased in the following order: controls approximately = Alzheimer's disease << Parkinson's disease < Lewy body dementia. Compared with patients with Parkinson's disease, patients with Lewy body dementia had greater scores for rigidity and deficits in the finger tapping test, but rest tremor and left/right asymmetry in EPS were more evident in Parkinson's disease. Patients with Lewy body dementia were also less likely to present with left/right asymmetry in EPS at the onset of their parkinsonism. "Sensitivity" to neuroleptic drugs was noted in 33% of patients with Lewy body dementia. Alzheimer's disease and Lewy body dementia groups had greater severity of dementia compared with the Parkinson's disease group and controls. Neuropsychological evaluation disclosed severe but similar degrees of impaired performances in tests of attention (digit span), frontal lobe function (verbal fluency, category, and Nelson card sort test) and motor sequencing in both Lewy body dementia and Alzheimer's disease groups, than Parkinson's disease and controls. In the clock face test, improved performance was noted in the "copy" compared to "draw" part of the test in controls, patients with Alzheimer's disease, and those with Parkinson's disease, but not in the patients with Lewy body dementia, who achieved equally poor scores in both parts of the test. CONCLUSIONS: EPS in Lewy body dementia resemble those seen in idiopathic Parkinson's disease, although less rest tremor and left/right asymmetry but more severe rigidity favours a diagnosis of Lewy body dementia. The unique profile of patients with Lewy body dementia seen in the clock face test suggests that this simple and easy to administer test may be useful in the clinical setting to differentiate Lewy body dementia and Alzheimer's disease.

Frost resistance and susceptibility to ice formation during natural hardening in relation to leaf anatomy in three evergreen tree species from New Zealand.

Foliar frost resistance of three endemic New Zealand land trees, Nothofagus menziesii (Hook. f.) Oerst. (Fagaceae), Pittosporum eugenioides A. Cunn. (Pittosporaceae) and Griselinia littoralis Forst. f. (Cornaceae), was examined as the trees hardened from late summer to midwinter in a lowland forest site. The lowest temperatures causing 50% damage (LT(50)) occurred in late winter and were similar to those recorded for other forest trees native to New Zealand (-11.7 degrees C in N. menziesii, -10.7 degrees C in P. eugenioides, and -10.6 degrees C in G. littoralis). All three species hardened by 4-7 degrees C, with G. littoralis showing the least frost resistance in summer and hence the greatest degree of hardening. Thermal analysis during freezing indicated that all three species became more tolerant of extracellular ice formation in winter. Measurements of chlorophyll a fluorescence correlated well with visible injury. The differing patterns of frost damage development in the three species were related to leaf anatomy: visible injury was localized within the small compartments formed by the highly septate leaves of the most resistant species, N. menziesii, and was somewhat localized in the partially septate leaves of P. eugenioides, whereas damage could be initiated anywhere in the aseptate leaves of G. littoralis,which was the least frost resistant species, particularly in summer.