An application of mathematical morphology to analysis of the size and shape of nuclei in tissue sections of non-Hodgkin's lymphoma

Using principles from the theory of mathematical morphology, a semiautomatic analysis of the size and shape of cell nuclei on tissue sections was carried out on a Leitz Texture Analysis System (Leitz-TAS). The four parameters proposed here are more discriminatory than conventional shape evaluation by the form factor (FF), which is based on the ratio of perimeter squared to area. The parameters quantified, respectively, nuclear elongation (ND), narrow (R1) and wide (R2) irregularities, and the distribution of R1 and R2 along the nuclear contour (ID). The properties of these parameters were tested nucleus-by-nucleus on 24 nuclear models. The methodology was then illustrated by a study of lymph node nuclei in non-Hodgkin's lymphoma (NHL). Prior to analysis, 45 lymphomas were classified into five categories of nuclear size and shape according to the International Working Formulation (IWF). Two hundred nuclei were measured on each lymph node section. Statistical interpretation was based upon an analysis of the nuclear surface area on sections and upon the mean values of R1, R2, and ND, the standard deviations of R1 and R2, and the percentage of cleaved nuclei detected by ID. The mean value of R2 discriminated best between the two sets of populations with regular and irregular nuclear contours, respectively. Parameters R1, ND, and ID permitted the distinction of certain NHL cases among populations with irregular nuclei. Nuclear invaginations decreased in depth as the nuclear area became greater. The median surface area was well correlated to the IWF, and the skewness coefficient (third statistical moment of the nuclear surface area distribution) was related to the number of nuclear size or shape subpopulations.     

Nuclear morphometry in canine acanthomatous ameloblastomas and squamous cell carcinomas

The aim of this study was to evaluate whether morphometrical analysis can be of diagnostic value for canine acanthomatous ameloblastoma. We calculated, by means of an automated image analyser, some morphometric nuclear parameters, in particular: mean nuclear area (MNA), mean nuclear perimeter (MNP), maximum and minimum diameters (MDx and MDm) coefficient of variation of the nuclear area (NACV), largest to smallest dimension ratio (LS ratio), and form factor (FF), in 8 canine acanthomatous ameloblastomas, and we compared these morphometric data to those of 13 squamous cell carcinomas of canine gingiva. The results indicated a progressive increase of the MNA, NACV, MNP and MDm proceeding from acanthomatous ameloblastomas (MNA: 42.11+/-8.74; NACV: 28,36+/-7,23; MNP: 24.18+/- 2.68; MDm: 5.69+/-0.49) to squamous cell carcinomas (MNA:49,69+/-9,10; NACV: 30,89+/-7,75; MNP: 25.63+/-2.54; MDm: 6.64+/-0.73). On the contrary, the LS ratio and the FF resulted greater in acanthomatous ameloblastomas (LS ratio: 1,63+/-0,12; FF: 1,13+/-0,002) than in SCCs (LS ratio: 1,40+/-0,12; FF:0.91+/-0.38). Moreover, the MNA, MNP,MDx and MDm resulted similar (MNA: p=0.89; MNP: p=0,65; MDm: p=0,16; MDx: p=0,13) in a subset of four acanthomatous ameloblastomas with cellular atypia (MNA:49,01+/-6,88; MNP: 26,28+/-1,99; MDm: 6.08+/-0.41; MDx: 10.18+/-0.88) and in squamous cell carcinomas (MNA:49.69+/-9,10; MNP: 25.63+/-2.54; MDm: 6.64+/-0.73; MDx: 9.26+/-1.05). While the NACV values resulted higher in typical acanthomatous ameloblastoma (29,99+/-6,06) than in atypical acanthomatous ameloblastoma (26,74+/-8,84) and similar to those of the SCCs (30,89+/-7,75). These results seem to confirm that acanthomatous ameloblastoma is a malignant or potentially malignant lesion and emphasizes that nuclear morphometry analysis can be an useful diagnostic and prognostic method in canine oral pathology.     

Comparison of life span of erythrocytes in some inbred strains of mouse using 14C-labelled glycine.

The erythrocyte life span in four inbred strains of mice--C57BL/10ScSnPh, B10.LP, BALB/c and CBA/JPh--was determined by means of erythrocyte labelling with 14C-glycine. Experimental data [decrease of 14C-activity of washed erythrocytes], corrected for blood loss, reincorporation of the label and delay in label incorporation during the initial period, were treated by a novel mathematical procedure based on the death probability function of the form: mu [t] = a + btlambda. The mean erythrocyte life span calculated using this function was, in the given sequence of strains, 42.1 +/- 0.6, 41.3 +/- 1.2, 39.3 +/- 0.9 and 38.6 +/- 0.6 days respectively. The rate of "random" destruction of erythrocytes was 1.20, 1.31, 0.70 and 0.63% of the total number of erythrocytes per day and a "mean potential" erthrocyte life span was found to be 58.9, 59.9, 46.2 and 44.5 days respectively. All the given parameters have similar numerical values in related strains and are apparently genetically conditioned. The erythrocyte life span determined simultaneously using DF32P labelling in the C57BL/10ScSnPh strain was 47.4 +/- 1.0 days. The implications of the results yielded by these two techniques is discussed.     

Endocrine parameters of human follicular fluid and fertilization capacity of oocytes

The induction of multiple follicular growth during ovarian stimulation for in vitro fertilization (IVF) implies follicular asynchrony. As a consequence oocytes of different quality are obtained. The maturity and fertilizability of oocytes cannot sufficiently be predicted by their morphological appearance under the light microscope. Looking for additional parameters of oocyte quality, FSH, hCG, estradiol (E2), progesterone (P), testosterone (T), prolactin (PRL) and cAMP were analysed in human follicular fluid (FF) containing a morphologically mature oocyte. The evaluation of the relationship between FF values and oocyte fertilization showed the following results: no differences of FSH, hCG, E2, P and T concentrations in FF between the group of fertilized and not fertilized ova. However, significant differences were detected for PRL and cAMP. In FF of fertilized oocytes PRL content was higher (38.8 +/- 2.2 vs 29.7 +/- 2.3 ng/ml, P less than 0.01) and cAMP level was lower (32.7 +/- 1.9 vs 59.8 +/- 7.4 pmol/ml, P less than 0.01) as compared with FF of unfertilizable oocytes. In conclusion PRL- and cAMP concentration of FF might be additional parameters of oocyte maturation and fertilizability.     

Differences in the plasma concentrations of FSH and LH in ovariectomized Booroola FF and ++ ewes

During 12 sampling days before ovariectomy the mean plasma FSH but not LH concentrations in FF ewes were higher (P less than 0.01) than those in ++ ewes (16 ewes/genotype). After ovariectomy increases in the concentrations of FSH and LH were noted for ewes of both genotypes within 3-4 h and the rates of increase of FSH and LH were 0.18 ng ml-1 h-1 and 0.09 ng ml-1 h-1 respectively for the first 15 h. From Days 1 to 12 after ovariectomy, the overall mean +/- s.e.m. concentrations for FSH in the FF and ++ ewes were 8.1 +/- 0.6 and 7.1 +/- 0.4 ng/ml respectively and for LH they were 2.7 +/- 0.3 and 2.1 +/- 0.2 ng/ml: these differences were not statistically significant (P = 0.09 for both FSH and LH; Student's t test). However, when the frequencies of high FSH or LH values after ovariectomy were compared with respect to genotype over time, significant F gene-specific differences were noted (P less than 0.01 for both FSH and LH; median test). In Exp. 2 another 21 ewes/genotype were blood sampled every 2nd day from Days 2 to 60 after ovariectomy and the plasma concentrations of FSH and LH were more frequently higher in FF than in ++ ewes (P less than 0.01 for FSH and LH). The F gene-specific differences in LH concentration, observed at 21-36 days after ovariectomy were due to higher mean LH amplitudes (P less than 0.025) but not LH peak frequency in FF than in ++ ewes.(ABSTRACT TRUNCATED AT 250 WORDS)     

Environmental monitoring for genotoxicity: in vivo sister chromatid exchange in the house mouse (Mus musculus)

Sister chromatid exchange (SCE) values were determined in bone marrow cells isolated from mouse (Mus musculus) femurs after injections of 5-bromo-2'-deoxyuridine (BrdU) and 5-fluorodeoxyuridine (FrdU). Male mice of C3H/J, C57BL/6J, and DBA/2 strains maintained in the laboratory gave mean SCE values of 3.42 +/- 0.07, 3.62 +/- 0.08, and 3.97 +/- 0.13, respectively. Males obtained from natural populations of southwestern Ontario had a higher mean SCE value (6.02 +/- 0.16), as did inbred males maintained in outdoor enclosures for at least 3 weeks (5.07 +/- 0.22). Wild mice housed in the laboratory for 9 months or longer had SCE values similar to laboratory bred mice (3.46 +/- 0.05). The SCE values in wild-caught mice were inversely proportional (r = -0.49) to the distance between the sites where these animals were collected and the nearest major industrial center. Based on these results, SCE analysis in mice is proposed as a possible first-line monitoring procedure for the detection of general changes in environmental genotoxicity.     

Differences in ovarian activity between booroola X merino ewes which were homozygous, heterozygous and non-carriers of a major gene influencing their ovulation rate

Differences in the function and composition of individual ovarian follicles were noted in Booroola Merino ewes which had previously been segregated on at least one ovulation rate record of greater than 5 (FF ewes, N = 15), 3-4 (F+ ewes, N = 18) or less than 3 (++ ewes, N = 18). Follicles in FF and F+ ewes produced oestradiol and reached maturity at a smaller diameter than in ++ ewes. In FF (N = 3), F+ (N = 3) and ++ (N = 3) ewes, the respective mean +/- s.e.m. diameters for the presumptive preovulatory follicles were 3.4 +/- 0.3, 4.1 +/- 0.2 and 6.8 +/- 0.3 mm and in each of these follicles the respective mean +/- s.e.m. numbers of granulosa cells (X 10(6)) were 1.8 +/- 0.3, 2.2 +/- 0.3 and 6.6 +/- 0.3. During a cloprostenol-induced follicular phase, the oestradiol secretion rates from FF ewes with 4.8 +/- 0.4 'oestrogenic' follicles, F+ ewes with 3.2 +/- 0.2 'oestrogenic' follicles and ++ ewes with 1.5 +/- 0.02 'oestrogenic' follicles were not significantly different from one another. Moreover, the mean total numbers of granulosa cells from the 'oestrogenic' follicles from each genotype were identical, namely 5.4 X 10(6) cells. Irrespective of genotype the mean weight of each corpus luteum was inversely correlated to the ovulation rate (R = 0.91, P less than 0.001). Collectively, these findings support the notion that the maturation of greater than or equal to 5 follicles in FF ewes and 3-4 follicles in F+ ewes may each be necessary to provide a follicular-cell mass capable of producing the same quantity of oestradiol as that from 1-2 preovulatory follicles in ++ ewes.     

Effect of single alien chromosome from shallot (Allium cepa L. Aggregatum group) on carbohydrate production in leaf blade of bunching onion (A. fistulosum L.)

We used a complete set of Allium fistulosum - shallot (A. cepa Aggregatum group) monosomic addition lines (FF+1A - FF+8A) to identify shallot chromosomes affecting the production of sugars. In the alien addition lines grown over two years in an experimental field at Yamaguchi University (34 degrees N, 131 degrees E), shallot chromosomes 2A and 8A altered sugar contents in leaf-bunching onion (A. fistulosum). Except for FF+2A, every monosomic addition accumulated non-reducing sugars in winter leaf blades. FF+8A caused an increase in the amounts of non-reducing sugars in the winter. FF+2A hardly produced non-reducing sugar throughout the two-year study. These results indicated that genes related to non-reducing sugar metabolism are located on the 2A and 8A chromosomes. The results of regression analyses using 2002 data on A. fistulosum and the monosomic addition set revealed a correlation (r = 0.63 +/- 0.07; mean +/- SE., n = 9) between reducing sugar and monosaccharide (Glc+Fru) contents but no correlation between non-reducing sugar and sucrose contents. This result indicates the existence of other polysaccharides (e.g., scorodose) as non-reducing sugars in the leaf blade.     

DNA flow cytometry of endoscopically examined colorectal adenomas and adenocarcinomas

DNA ploidy of 64 colorectal adenomas and 49 adenocarcinomas, examined endoscopically, was studied by flow cytometry. We found DNA aneuploidy in none of the 105 normal mucosa samples (0%), in 20 adenomas (31%), and in 36 adenocarcinomas (74%). DNA ploidy of adenomas correlated with size (P = 0.02) and degree of dysplasia (P less than 0.01) but not with histologic type. Adenomas had a 45% incidence of DNA aneuploid stem lines in the DNA index range of 0.80-1.20, compared with 8% in the case of adenocarcinomas. The distribution of the DNA index values of adenocarcinomas was approximately normal, with a mean value 1.63 +/- 0.28. The mean DNA index for the three cases of "carcinoma in adenoma" with invasion of the stalk of the adenoma was 1.52 +/- 0.18. These results, using DNA flow cytometry, provide evidence for the progression of colorectal adenoma to adenocarcinoma. The classification of adenomas according to DNA ploidy may be information of considerable practical value to the clinician in predicting risk of further adenomas and/or risk of cancer.     

离体培养小鼠神经细胞在有丝分裂过程中的膜电位变化（简报）

Primarily cultured dorsal root ganglion cells and olfactory bulb cells were dissected from 12 to 14-d-old fetal C57 BL/6 J mice. After the cells were cultured for about two weeks, the growing status of cells were observed and the membrane potentials(MP) were recorded. The results show that mean value of the MP in anaphase was -68 +/- 3.1 mV (SE, n = 3), same as in interphase: Beginning from telophase, the cell membrane in the equator contracted gradually to become a concave ditch, and the MP decreased obviously, the mean value was -23.3 +/- 3.3 mV (SE, n = 6). After this, MP recovered gradually, till it divided into two sister cells. MP which were recorded separately in two sister cells were similar. But usually MP did not recovered to their normal values immediately.     

Exposure to lead and cadmium and associated factors in children 0–17 years of age living in an area contaminated by metals

This article analyzes some characteristics and conditions associated with Pb and Cd exposure using an exploratory approach, relating them to the levels of Pb and Cd in blood (blood lead levels [BLL] and blood cadmium levels [BCL]) in children 0–17 years of age living in an area contaminated by metals. BLL and BCL values were determined for each child and questionnaires were applied to their parents. Significant differences were found in mean BLL values according to race (p = .03), family history of intoxication by Pb (p = .004), if a family member was a metallurgy worker (p = .047), if a family member performed activities in the area of metallurgy (p = .03), and mother's employment status (p = .014). The following characteristics were associated with increased risk of BLL above 5.0 mg/dL in children: race, having a family record of intoxication by lead, and having a parent who worked at the plant. BLC values are not significantly associated with any of the studied characteristics. It may be concluded that BLL values are influenced by the social indicators of the population.     

Evaluation of adenosine deaminase seric activity in the diagnosis of bovine tuberculosis

Determination of seric levels of adenosine deaminase (ADA), an enzyme produced by monocytes/macrophages and lymphocytes, has been used in the diagnosis of human tuberculosis (TB). In the present study, ADA seric activity was evaluated comparatively to the comparative tuberculin test in the diagnosis of bovine tuberculosis. Two hundred fifty-six cattle were classified by origin and by the comparative tuberculin test as TB-positive animals (n = 52, from herds where the Mycobacterium bovis had previously been isolated), and TB-negative animals (n = 204, TB-free herds). The mean ADA seric value from the TB-positive group (4.45 +/- 2.33 U/L) was significantly lower (p = 0.008) than that observed in sera from the TB-negative group (6.12 +/- 4.47 U/L). When animals from a herd with clinical cases of enzootic bovine leukosis of TB-negative group were withdrawn from analysis, the mean ADA seric values of TB-negative group (5.12 +/- 3.75 U/L) was not significantly different anymore from that of the TB-positive group (p = 0.28). There was no agreement in the diagnosis of bovine TB between comparative tuberculin test and determination of ADA seric values, using two different cutoff points, being 6.12 U/L and 15.0 U/L, (kappa = -0.086 and kappa = -0.082, respectively). In conclusion, the determination of ADA seric activity was not a good auxiliary test for bovine TB, because it was not able to distinguish between TB-positive and TB-negative animals.     

Scale-independent roughness value of cell membranes studied by means of AFM technique

The roughness of cell membrane is a very interesting indicator of cell's health state. Atomic Force Microscopy allows us to investigate the roughness of cell membrane in great detail, but the obtained roughness value is scale-dependent, i.e. it strongly depends on measurement parameters, as scanning area and step size. The scale-dependence of the roughness value can be reduced by means of data filtration techniques, that are not standardized at nanometric scale, especially as far as biological data are concerned. In this work, a new method, based on the changes of values of some roughness parameter (root mean square roughness and skewness) as a function of filtration frequencies, has been implemented to optimize data filtering procedure in the calculation of cell membrane roughness. In this way, a root mean square roughness value independent of cell shape, membrane micro-irregularities and measurement parameters can be obtained. Moreover, different filtration frequencies selected with this method allow us to discriminate different surface regimes (nominal form, waviness and roughness) belonging to the raw cell profile, each one related to different features of the cell surface.     

Aminoguanidine prevents fructose-induced arterial stiffening in Wistar rats: aortic impedance analysis

Fructose has been reported as a potent agent in forming advanced glycation end products (AGEs) and, thus, may play a significant role in the pathogenesis of diabetic complications. Herein, we determined the effects of aminoguanidine (AG), an inhibitor of AGEs, on the mechanical properties of the arterial system in fructose-fed (FF) rats, using aortic impedance analysis. Rats at 2 months were given 10% fructose in drinking water for 2 weeks and compared with untreated age-matched controls. Meanwhile, FF rats were treated for 2 weeks with AG (daily peritoneal injections of 50 mg kg(-1)) and compared with the untreated FF group. Neither fructose nor AG affects body weight, blood glucose level, and basal heart rate. In comparison with controls, FF rats showed a decrease in cardiac output in the absence of any significant changes in mean aortic pressure, having increased total peripheral resistance (R(p)), at 51.1 +/- 2.9 versus 66.2 +/- 1.9 mm Hg sec ml(-1) (P < 0.05). Fructose also contributed to an increase in aortic characteristic impedance (Z(c)), from 1.528 +/- 0.094 to 1.933 +/- 0.084 mm Hg sec ml (-1) (P < 0.05) and a decrease in wave transit time (tau), from 22.6 +/- 0.6 to 19.2 +/- 0.7 msec (P < 0.05). The elevated Z(c) and the reduced tau suggest that fructose may cause a detriment to the aortic distensibility in animals. After exposure to AG, FF rats exhibited a significant improvement in physical properties of the resistance vessels, as evidenced by the reduction of 21.3% in R(p). Meanwhile, AG retarded the fructose-induced decline in aortic distensibility, as reflected in the decrease of 16.0% in Z(c) (P < 0.05) and the increase of 18.1% in tau (P < 0.05). By contrast, AG exerted no effects on the mechanical properties of Windkessel vessels, as well as resistance vessels, in normal diet controls. We conclude that AG may prevent the fructose-derived changes in arterial stiffening, possibly through inhibition of the fructose-derived advanced glycation end product formation in Wistar rats.     

Nucleolar organizer regions in aspirates of malignant lymphomas and benign disorders of the lymph nodes.

Silver staining of nucleolar organizer regions (AgNOR) was used to differentiate malignant lymphoma and chronic lymphadenitis. Aspiration smear samples from lymph nodes of 120 cases, including 43 non-Hodgkin's lymphoma, 3 Hodgkin's disease, 56 chronic lymphadenitis, 7 tuberculosis, 6 reactive hyperplasia and 5 samples from other diseases (epidermoid cyst, branchial cyst, mixed tumor, lymphoepithelioma and nodulous disease), were investigated. The number of AgNORs in 200 cells in each sample was counted, and the mean +/- SD in each disease was calculated: non-Hodgkin's lymphoma, 6.58 +/- 2.37; Hodgkin's disease, 4.22 +/- 0.5; chronic lymphadenitis, 1.16 +/- 0.1; tuberculosis, 1.13 +/- 0.14; reactive hyperplasia, 1.48 +/- 0.25; other diseases, 1.47 +/- 0.31. The results indicate that the AgNOR count in malignant lymphoma differed highly significantly from that in benign disease (P less than .001). The size of AgNORs in malignant lymphoma and chronic lymphadenitis was measured, and the maximum diameter and area of lymphocyte and lymphoma cell were: lymphocyte, 0.93 +/- 0.12 microns, 0.61 +/- 0.13 microns 2; lymphoma cell, 0.83 +/- 0.22 microns, 0.50 +/- 0.25 microns 2. The AgNOR sizes in malignant lymphoma were significantly smaller than in chronic lymphadenitis (P less than .001).     

243,244Cm studies in C57BL/Do mice

Three groups of C57BL/Do mice were injected with different activities of 243,244Cm so that the long-term biological effects could be evaluated. The biological retention, R, of injected curium in the skeleton at t days after injection could be represented by the equations R = 0.245e-0.000379t and R = 0.208e-0.000494t for male and female mice, respectively. Effective skeletal retention equations were used to calculate the cumulative mean skeletal dose in rad at 140 days before death in each group of mice. The primary objective of this study was to determine the relative biological effectiveness (RBE) of 243,244Cm compared to 226Ra, using bone sarcoma induction as the end point. Combined data (bone sarcomas per 10(6) mouse-rad) for male and female mice permitted the RBE value +/- SD for 243,244Cm to be calculated as 4.4 +/- 1.8 compared to 1.0 for 226Ra. A comparison of RBE values form a previous study in this mouse strain and the value for 243,244Cm from this study suggests that the trivalent actinides 241Am, 243,244Cm, and 249Cf are about three times less effective for bone sarcoma induction than 239Pu.     

Thyroxine binding globulin (TBG) assay by radial immunodiffusion.

Thyroxine-binding globulin was measure by radial immuno-diffusion using ten microliters of serum. The mean value in a normal group was 1.95 +/- 0.38 mg/dl. The values obtained by this method were in good agreement with the thyroxine-binding capacity determined by paper electrophoresis (r = 0.956).     

Method for the quantitative evaluation of the distribution pattern of nuclei in normal and malignant endometrial epithelia

A quantitative method of evaluating the pattern of distribution of nuclei within a cell cluster was developed and applied to endometrial cytology. The distribution pattern (DP) is mathematically expressed using a "DP index," which can be determined as a product of the square root of the nuclear density, square root of n0, and the mean distance between the two nearest nuclei, r. The DP index has a limited range of decreasing values from 1.074 to 0.5 to 0, indicative of regular, random and aggregated patterns of nuclear distribution, respectively. The estimated DP index was 0.831 +/- 0.031 (mean +/- standard deviation) in the clusters of normal endometrial epithelial cells from 16 nonneoplastic cases, but 0.638 +/- 0.041 in malignant epithelia from 19 cases of well-differentiated adenocarcinoma. The index of normal epithelia was close to 0.877 of the regular hexagonal distribution pattern. Contrary to this, the DP index was significantly smaller in well-differentiated adenocarcinoma (P less than .001), approaching 0.5, the theoretical value of a random distribution pattern. These findings suggest that quantitative analysis of the distribution pattern of nuclei can be a useful aid in the cytodiagnosis of endometrial lesions.     

The development and validation of a charge-coupled device laser reflectance system to measure the complex cross-sectional shape and area of soft tissues

Determination of the stresses in soft tissues such as ligaments and tendons under uniaxial tension require accurate measurement of their cross-sectional area. Of the many methods available, there are concerns regarding contact methods which exert external loads and deform the cross-sectional shape of soft tissues. Hence, the area measurements are affected. On the other hand, non-contact methods have difficulties in dealing with complex shapes, especially with concavities. To address these problems, a new measurement system using a charge-coupled device (CCD) laser displacement sensor has been developed and tested. This system measures the complete surface profile of the object by rotating the laser 360° around the soft tissue. Then, the cross-sectional shape is reconstructed and the cross-sectional area determined via Simpson's rule. The system's accuracy was first verified with objects of various cross-sectional shapes and areas (cylinder: 23.1, 76.5, 510.3 mm²; cuboid: 34.3, 163.8, 316.7 mm², and cylinder with concavities: 121.4 mm²). The CCD laser reflectance system's accuracy was within 2.0% for these objects. To test biological application, the goat Achilles tendon and the anteromedial bundle of the porcine anterior cruciate ligament specimens were measured and compared to values obtained using another accepted technique, the laser micrometer system. The areas obtained using the CCD laser reflectance system were 4.4% and 9.7% lower than those obtained with the laser micrometer system respectively. These differences could be mainly attributed to concavities. Thus, the CCD laser reflectance system is an improved method for measuring the cross-sectional shape and area of soft tissues since it can detect and account for concavities without physically contacting the specimen.     

Renal damage in the mouse: the effect of d(4)-Be neutrons

A further study on the response of the mouse kidney to d(4)-Be neutrons (EN = 2.3 MeV) is described. The results confirm and augment the work published previously by Stewart et al. [Br. J. Radiol. 57, 1009-1021 (1984)]; the present paper includes the data from a "top-up" design of experiment which extends the measurements of neutron RBE (relative to 240 kVp X rays) down to X-ray doses of 0.75 Gy per fraction. The mean RBE for these neutrons increases from 5.8 to 7.3 as X-ray dose per fraction decreases from 3.0 to 1.5 Gy in the kidney. This agrees with the predictions from the linear quadratic (LQ) model, based on the renal response to X-ray doses above 4 Gy per fraction. The mean RBE estimate from a single dose group at 0.75 Gy per fraction of X rays is, however, 3.9. This is below the LQ prediction and may indicate increasing X-ray sensitivity at low doses. Data from this study and from those published previously have been used to determine more accurately the shape of the underlying response to d(4)-Be neutrons; an alpha/beta ratio of 20.5 +/- 3.7 Gy was found. The best value of alpha/beta for X rays determined from these experiments was 3.04 +/- 0.35 Gy, in agreement with previous values.