The effect of right paralumbar fossa exploratory celiotomy and omentopexy on peritoneal fluid constituents was studied in 22 adult dairy cows. Six cows were eliminated on the basis of physical examination findings (n = 2), surgical findings (n = 2), or inability to obtain a sufficient volume of peritoneal fluid (n = 2). Sixteen cattle had normal results of Csc and serum biochemical analysis, and a minimum of 1 ml of peritoneal fluid was obtained by abdominocentesis. Abdominocentesis was repeated on days 1, 2, and 6 after surgery. Statistical analysis for repeated measures was performed, using a significance level of P < 0.05. Stage of gestation was evaluated for interaction with time. Mean total nucleated cell count was 3,200 cells/1 before surgery, was significantly increased 2 days after surgery (16,336 cells/microliter), and continued to increase through day 6 (20,542 cells/microliter). Mean polymorphonuclear cell count was 1,312 cells/microliter before surgery and was significantly higher at 2 (11,043 cells/microliter) and 6 (10,619 cells/microliter) days after surgery. Mean lymphocyte count was 254 cells/microliter before surgery and was significantly increased 2 days (1,911 cells/microliter) after surgery. By day 6, lymphocyte numbers were similar to preoperative values. Mean mononuclear cell count was 770 cells/microliter before surgery and was significantly increased on days 1 (3,084 cells/microliter), 2 (3,285 cells/microliter and 6 (2,349 cells/microliter) after surgery. Mean eosinophil numbers were 1,388 cells/microliter before surgery and were significantly increased on day 6 (6,347 cells/microliter) only. Interaction between time and stage of gestation was found only for specific gravity and total protein concentration. In general, specific gravity and total protein concentration increased after surgery (mean before surgery, 1.016 and 3.6 g/dl; mean after surgery, 1.021 and 5.6 g/dl).

The effect of right paralumbar fossa exploratory celiotomy and omentopexy on peritoneal fluid constituents was studied in 22 adult dairy cows. Six cows were eliminated on the basis of physical examination findings (n = 2), surgical findings (n = 2), or inability to obtain a sufficient volume of peritoneal fluid (n = 2). Sixteen cattle had normal results of Csc and serum biochemical analysis, and a minimum of 1 ml of peritoneal fluid was obtained by abdominocentesis. Abdominocentesis was repeated on days 1, 2, and 6 after surgery. Statistical analysis for repeated measures was performed, using a significance level of P < 0.05. Stage of gestation was evaluated for interaction with time. Mean total nucleated cell count was 3,200 cells/1 before surgery, was significantly increased 2 days after surgery (16,336 cells/microliter), and continued to increase through day 6 (20,542 cells/microliter). Mean polymorphonuclear cell count was 1,312 cells/microliter before surgery and was significantly higher at 2 (11,043 cells/microliter) and 6 (10,619 cells/microliter) days after surgery. Mean lymphocyte count was 254 cells/microliter before surgery and was significantly increased 2 days (1,911 cells/microliter) after surgery. By day 6, lymphocyte numbers were similar to preoperative values. Mean mononuclear cell count was 770 cells/microliter before surgery and was significantly increased on days 1 (3,084 cells/microliter), 2 (3,285 cells/microliter and 6 (2,349 cells/microliter) after surgery. Mean eosinophil numbers were 1,388 cells/microliter before surgery and were significantly increased on day 6 (6,347 cells/microliter) only. Interaction between time and stage of gestation was found only for specific gravity and total protein concentration. In general, specific gravity and total protein concentration increased after surgery (mean before surgery, 1.016 and 3.6 g/dl; mean after surgery, 1.021 and 5.6 g/dl). Left paralumbar fossa celiotomy performed 7 days after surgery did not reveal complications of repeated abdominocentesis, and pregnancy status was unchanged. Peritoneal fluid constituents are highly variable after exploratory celiotomy and omentopexy in cattle. However, results of this study may provide a reference for interpretation of postoperative peritoneal fluid sample findings in cattle.

We evaluated the effectiveness of 2 dopamine antagonists as treatments for fescue toxicosis in horses. Sixteen gravid mares were assigned by breed and expected foaling date to 1 of 3 treatment groups: endophyte-infested control 1.1 mg of domperidone/kg of body weight/d; and 3.3 mg of sulpiride/kg/d. Mares were pastured on endophyte-infected fescue and received 0.454 kg of a corn and dried molasses carrier containing the drug treatment. Treatment started 30 days prior to expected foaling date and continued until parturition. Blood samples were collected, and mammary gland scores were recorded every 5 days. Body weight and body condition scores were obtained every 28 days. Serum was analyzed for prolactin, progesterone, and estradiol-17beta concentrations. Domperidone-treated mares had shorter (P = 0.09) gestation duration and foaled closer (P = 0.07) to their expected parturition date than did control mares. Mammary gland scores were higher (P < 0.05) for domperidone-treated mares than for control mares. By 4 and 9 days after the start of treatment, serum prolactin concentration was higher P < 0.05) in domperidone-treated mares and sulpiride-treated mares, respectively, than in control mares. Domperidone- and and sulpiride-treated mares had higher (P < 0.05) serum progesterone and lower (P < 0.01) estradiol-17beta concentrations than did control mares. These results indicate that domperidone may offer considerable potential as a treatment for fescue toxicosis in horses.

We evaluated the effectiveness of 2 dopamine antagonists as treatments for fescue toxicosis in horses. Sixteen gravid mares were assigned by breed and expected foaling date to 1 of 3 treatment groups: endophyte-infested control 1.1 mg of domperidone/kg of body weight/d; and 3.3 mg of sulpiride/kg/d. Mares were pastured on endophyte-infected fescue and received 0.454 kg of a corn and dried molasses carrier containing the drug treatment. Treatment started 30 days prior to expected foaling date and continued until parturition. Blood samples were collected, and mammary gland scores were recorded every 5 days. Body weight and body condition scores were obtained every 28 days. Serum was analyzed for prolactin, progesterone, and estradiol-17beta concentrations. Domperidone-treated mares had shorter (P = 0.09) gestation duration and foaled closer (P = 0.07) to their expected parturition date than did control mares. Mammary gland scores were higher (P < 0.05) for domperidone-treated mares than for control mares. By 4 and 9 days after the start of treatment, serum prolactin concentration was higher P < 0.05) in domperidone-treated mares and sulpiride-treated mares, respectively, than in control mares. Domperidone- and and sulpiride-treated mares had higher (P < 0.05) serum progesterone and lower (P < 0.01) estradiol-17beta concentrations than did control mares. These results indicate that domperidone may offer considerable potential as a treatment for fescue toxicosis in horses.

A study was conducted to investigate whether aspiration of amniotic fluid is associated with a deleterious effect on absorption of colostral immunoglobulins or on blood gas and acid-base values of healthy newborn calves. Fourteen calves purchased from commercial sources were transported to a research facility immediately after birth and fed colostrum with known concentrations of immunoglobulins. Blood samples for gas analyses were collected within 5 hours of birth, 24 hours later, and prior to euthanasia. Between 3 and 5 days of age, calves were euthanatized by an overdose of barbiturates. Eleven calves had evidence of bronchoaspiration of amniotic fluid, as determined by presence of meconium, squamous epithelium, or keratin in histologic sections of fixed lung or by cytologic analysis of bronchoalveolar lavage fluid. Blood gas tensions and pH were within reference ranges in 11 of 14 calves. Aspiration of amniotic fluid could not be linked to any specific changes in blood gas tensions, acid-base status, or absorption of colostral immunoglobulins. Presence of keratin and meconium in the lungs often was accompanied by mild exudative alveolitis and focal atelectasis. It was concluded that aspiration of small amounts of amniotic fluid with or without meconium is common in calves and is not associated with hypoxemia, respiratory acidosis, or failure of passive transfer.

A study was conducted to investigate whether aspiration of amniotic fluid is associated with a deleterious effect on absorption of colostral immunoglobulins or on blood gas and acid-base values of healthy newborn calves. Fourteen calves purchased from commercial sources were transported to a research facility immediately after birth and fed colostrum with known concentrations of immunoglobulins. Blood samples for gas analyses were collected within 5 hours of birth, 24 hours later, and prior to euthanasia. Between 3 and 5 days of age, calves were euthanatized by an overdose of barbiturates. Eleven calves had evidence of bronchoaspiration of amniotic fluid, as determined by presence of meconium, squamous epithelium, or keratin in histologic sections of fixed lung or by cytologic analysis of bronchoalveolar lavage fluid. Blood gas tensions and pH were within reference ranges in 11 of 14 calves. Aspiration of amniotic fluid could not be linked to any specific changes in blood gas tensions, acid-base status, or absorption of colostral immunoglobulins. Presence of keratin and meconium in the lungs often was accompanied by mild exudative alveolitis and focal atelectasis. It was concluded that aspiration of small amounts of amniotic fluid with or without meconium is common in calves and is not associated with hypoxemia, respiratory acidosis, or failure of passive transfer.

Effects of the following treatments on abomasal and duodenal myoelectric activity in yearling cattle were studied: 2 ml of 0.9% sodium chloride solution (NACL); 0.07 mg of bethanechol (BET)/kg of body weight; 0.1 mg of metoclopramide (MET)/kg; and 0.07 mg of bethanechol and 0.1 mg of metoclopramide (BETMET)/kg. All treatments were administered SC during the early part of phase I of the migrating myoelectric complex Myoelectric signals were recorded for 4 hours after administration of the treatments from 1 electrode in the antrum and 3 electrodes in the duodenum. For the antral spike rate (ASR), there was no significant difference among treatments during the first hour, but the ASR was significantly (P < 0.05) greater during hours 2 to 4 after treatment with BETMET, compared with ASR for MET alone. The duodenal spike rate (DSR) was significantly (P < 0.05) greater during the first hour after administration of BETMET than after the other treatments. After administration of BET, DSR was significantly (P < 0.05) greater than after MET or NACL. There was no difference in DSR after MET, compared with DSR after NACL. There was no significant difference in DSR among treatments during the second and third hours. The total antegrade propagating spike (TAPS) count was greater after administration of BETMET in all hours, compared with the other treatments. The ratio of TAPS to total spikes on the orad-most duodenal electrode was significantly (P < 0.05) greater after BETMET during hours 1 and 2.

Effects of the following treatments on abomasal and duodenal myoelectric activity in yearling cattle were studied: 2 ml of 0.9% sodium chloride solution (NACL); 0.07 mg of bethanechol (BET)/kg of body weight; 0.1 mg of metoclopramide (MET)/kg; and 0.07 mg of bethanechol and 0.1 mg of metoclopramide (BETMET)/kg. All treatments were administered SC during the early part of phase I of the migrating myoelectric complex Myoelectric signals were recorded for 4 hours after administration of the treatments from 1 electrode in the antrum and 3 electrodes in the duodenum. For the antral spike rate (ASR), there was no significant difference among treatments during the first hour, but the ASR was significantly (P < 0.05) greater during hours 2 to 4 after treatment with BETMET, compared with ASR for MET alone. The duodenal spike rate (DSR) was significantly (P < 0.05) greater during the first hour after administration of BETMET than after the other treatments. After administration of BET, DSR was significantly (P < 0.05) greater than after MET or NACL. There was no difference in DSR after MET, compared with DSR after NACL. There was no significant difference in DSR among treatments during the second and third hours. The total antegrade propagating spike (TAPS) count was greater after administration of BETMET in all hours, compared with the other treatments. The ratio of TAPS to total spikes on the orad-most duodenal electrode was significantly (P < 0.05) greater after BETMET during hours 1 and 2.

The mydriatic effect of 3 curare-like neuromuscular blocking agents was tested in European kestrels (Falco tinnunculus) after topical application. Alcuronium chloride (5 mg/ml) was found to be effective at a dose of 1 drop (20 drops = 1 ml) administered twice at a 15-minute interval. Mydriasis was achieved at t = 26 +/- 11 minutes, maximal effect was reached at t = 60 +/- 39 minutes, and sufficient mydriasis ended at t = 364 +/- 134 minutes. Nevertheless, side effects, including temporary full paralysis in 1 bird, indicated that this drug should not be used. Pancuronium bromide (2 mg/ml) had an inconsistent effect on each bird at a dose of 2 drops administered twice at 15-minute intervals, and total mydriasis was not reached in 5 of 8 birds. Mydriasis was achieved at t = 34 +/- 11 minutes, maximal effect was reduced and reached at t = 43 +/- 13 minutes, and sufficient mydriasis ended at t = 90 +/- 39 minutes. Vecuronium bromide (4 mg/ml) was administered at a dose of 2 drops, 3 times, at 15-minute intervals. Mydriasis was achieved at t = 23 +/- 8 minutes, maximal effect was reached at t = 65 +/- 12 minutes, and sufficient mydriasis ended at t = 253 +/- 65 minutes. Side effects were not detected. Vecuronium bromide should be used in raptorial birds whenever retinal examination requires fundoscopy.

The mydriatic effect of 3 curare-like neuromuscular blocking agents was tested in European kestrels (Falco tinnunculus) after topical application. Alcuronium chloride (5 mg/ml) was found to be effective at a dose of 1 drop (20 drops = 1 ml) administered twice at a 15-minute interval. Mydriasis was achieved at t = 26 +/- 11 minutes, maximal effect was reached at t = 60 +/- 39 minutes, and sufficient mydriasis ended at t = 364 +/- 134 minutes. Nevertheless, side effects, including temporary full paralysis in 1 bird, indicated that this drug should not be used. Pancuronium bromide (2 mg/ml) had an inconsistent effect on each bird at a dose of 2 drops administered twice at 15-minute intervals, and total mydriasis was not reached in 5 of 8 birds. Mydriasis was achieved at t = 34 +/- 11 minutes, maximal effect was reduced and reached at t = 43 +/- 13 minutes, and sufficient mydriasis ended at t = 90 +/- 39 minutes. Vecuronium bromide (4 mg/ml) was administered at a dose of 2 drops, 3 times, at 15-minute intervals. Mydriasis was achieved at t = 23 +/- 8 minutes, maximal effect was reached at t = 65 +/- 12 minutes, and sufficient mydriasis ended at t = 253 +/- 65 minutes. Side effects were not detected. Vecuronium bromide should be used in raptorial birds whenever retinal examination requires fundoscopy.

Dispositions of caffeine and antipyrine were compared as indicators of decreasing hepatic function in dogs with experimentally induced progressive liver disease. Dimethylnitrosamine, a hepatospecific toxin, was administered orally to 16 dogs; 6 dogs served as controls (group 1). Three classes of liver disease were defined by histologic features: mild (group 2; n = 5), moderate (group 3; n = 6), and severe (group 4; n = 5). Disposition of antipyrine, and 24 hours later, caffeine was studied 3 weeks after the last dose of toxin in each dog. For both drugs, rapid IV administration of 20 mg/kg of body weight was administered and serum samples were obtained at intervals for determination of at least 5 terminal-phase drug half-lives. For both drugs, clearance and mean residence time differed among groups (P less than or equal to 0.01). Clearance of antipyrine and caffeine was decreased in groups 3 and 4, compared with groups 1 and 2. Antipyrine and caffeine mean residence times were longer in group-3 dogs, compared with dogs of groups 1 and 2. Correction of caffeine and antipyrine clearances for hepatic weight increased discrimination between groups 3 and 4. The clearance and mean residence time ratios of antipyrine to caffeine were calculated for each group and, when compared with values for group-1 dogs, were used to test for differences between the 2 drugs in response to disease. Ratios did not differ among groups. These results indicate that the disposition of antipyrine and caffeine may change similarly with progression of dimethylnitrosamine-induced liver disease.

Dispositions of caffeine and antipyrine were compared as indicators of decreasing hepatic function in dogs with experimentally induced progressive liver disease. Dimethylnitrosamine, a hepatospecific toxin, was administered orally to 16 dogs; 6 dogs served as controls (group 1). Three classes of liver disease were defined by histologic features: mild (group 2; n = 5), moderate (group 3; n = 6), and severe (group 4; n = 5). Disposition of antipyrine, and 24 hours later, caffeine was studied 3 weeks after the last dose of toxin in each dog. For both drugs, rapid IV administration of 20 mg/kg of body weight was administered and serum samples were obtained at intervals for determination of at least 5 terminal-phase drug half-lives. For both drugs, clearance and mean residence time differed among groups (P less than or equal to 0.01). Clearance of antipyrine and caffeine was decreased in groups 3 and 4, compared with groups 1 and 2. Antipyrine and caffeine mean residence times were longer in group-3 dogs, compared with dogs of groups 1 and 2. Correction of caffeine and antipyrine clearances for hepatic weight increased discrimination between groups 3 and 4. The clearance and mean residence time ratios of antipyrine to caffeine were calculated for each group and, when compared with values for group-1 dogs, were used to test for differences between the 2 drugs in response to disease. Ratios did not differ among groups. These results indicate that the disposition of antipyrine and caffeine may change similarly with progression of dimethylnitrosamine-induced liver disease.

Cardiovascular and at accompany markedly long periods (12 hours) of halothane anesthesia were characterized. Eight spontaneously breathing horses were studied while they were positioned in left lateral recumbency and anesthetized only with halothane in oxygen maintained at a constant end-tidal concentration of 1.06% (equivalent to 1.2 times the minimal alveolar concentration for horses). Results of circulatory and respiratory measurements during the first 5 hours of constant conditions were similar to those previously reported from this laboratory (ie, a time-related significant increase in systemic arterial blood pressure, cardiac output, stroke volume, left ventricular work, PCV, plasma total solids concentration, and little change in respiratory system function). Beyond 5 hours of anesthesia, arterial blood pressure did not further increase, but remained above baseline. Cardiac output continued to increase, because heart rate significantly (P < 0.05) increased. Peak inspiratory gas flow increased significantly (P < 0.05) in later stages of anesthesia. There was a significant decrease in inspiratory time beginning at 4 hours. Although PaO2, and PaCO2, did not significantly change during the 12 hours of study, PVO2 increased significantly P < 0.05) and progressively with time, beginning 6 hours after the beginning of constant conditions. Metabolic acidosis increased with time significantly [P < 0.05] starting at 9 hours), despite supplemental IV administered NaHCO3. Plasma concentrations of eicosanoids: 6-ketoprostaglandin F1 alpha (PGF1 alpha, a stable metabolite of PGI2), PGF2 alpha, PGE, and thromboxane (TxB2, a stable metabolite of TxA2) were measured in 5 of the 8 horses before and during anesthesia. Significant changes from preanesthetic values were not Significant changes from preanesthetic values were not detected. Dynamic thoracic wall and lung compliances decreased with time.

Cardiovascular and at accompany markedly long periods (12 hours) of halothane anesthesia were characterized. Eight spontaneously breathing horses were studied while they were positioned in left lateral recumbency and anesthetized only with halothane in oxygen maintained at a constant end-tidal concentration of 1.06% (equivalent to 1.2 times the minimal alveolar concentration for horses). Results of circulatory and respiratory measurements during the first 5 hours of constant conditions were similar to those previously reported from this laboratory (ie, a time-related significant increase in systemic arterial blood pressure, cardiac output, stroke volume, left ventricular work, PCV, plasma total solids concentration, and little change in respiratory system function). Beyond 5 hours of anesthesia, arterial blood pressure did not further increase, but remained above baseline. Cardiac output continued to increase, because heart rate significantly (P < 0.05) increased. Peak inspiratory gas flow increased significantly (P < 0.05) in later stages of anesthesia. There was a significant decrease in inspiratory time beginning at 4 hours. Although PaO2, and PaCO2, did not significantly change during the 12 hours of study, PVO2 increased significantly P < 0.05) and progressively with time, beginning 6 hours after the beginning of constant conditions. Metabolic acidosis increased with time significantly [P < 0.05] starting at 9 hours), despite supplemental IV administered NaHCO3. Plasma concentrations of eicosanoids: 6-ketoprostaglandin F1 alpha (PGF1 alpha, a stable metabolite of PGI2), PGF2 alpha, PGE, and thromboxane (TxB2, a stable metabolite of TxA2) were measured in 5 of the 8 horses before and during anesthesia. Significant changes from preanesthetic values were not Significant changes from preanesthetic values were not detected. Dynamic thoracic wall and lung compliances decreased with time.

Middle carpal cartilage explants from 4 horses with mild osteoarthritis involving that joint were maintained in tissue culture to test the effects of a polysulfated glycosaminoglycan (PSGAG) on proteoglycan synthesis and degradation. Cultures were exposed to 0.025 or 25 mg of PSGAG/ml for 48 hours, after which the medium was replaced with medium containing similar doses of PSGAG and 35S. Subsequently, the sulfated proteoglycan content of the medium and extracts of the explants was measured. Gel filtration chromatography was used to estimate the size and to purify the principal, large proteoglycan monomer, which was further characterized by digestion, using glycosidic enzymes. In a second experiment, explants were incubated with 35S for 48 hours, and were subsequently exposed to the same concentrations of the PSGAG for an additional 48 hours. The amount of remaining labeled proteoglycan was determined for culture medium and cartilage extracts. Gel filtration chromatography was used to assess the hydrodynamic size of the large proteoglycan monomer. Aliquots of proteoglycans from the second experiment were incubated in high-molecular weight hyaluronate and chromatographed to assess reaggregation. Polysulfated glycosaminoglycan caused a significant (P < 0.04) decrease in sulfated proteoglycan synthesis by cartilage explants. Radioactive proteoglycan content in explants labeled prior to exposure to PSGAG were similar. Large proteoglycan monomer size was similar in both experiments (median partition coefficient [K(AV)] = 0.40), and was not influenced by PSGAG treatment. Prelabeled explants exposed to hyaluronate and chromatographed under associative conditions had similar proportions of the radiolabel eluting as proteoglycan aggregate. Enzymatic digestion of newly synthesized large monomer revealed a mild dose-dependent increase in the proportion of keratan sulfate substitution on core protein. It was concluded that PSGAG in vitro, at the dosages evaluated, caused a decre.

Middle carpal cartilage explants from 4 horses with mild osteoarthritis involving that joint were maintained in tissue culture to test the effects of a polysulfated glycosaminoglycan (PSGAG) on proteoglycan synthesis and degradation. Cultures were exposed to 0.025 or 25 mg of PSGAG/ml for 48 hours, after which the medium was replaced with medium containing similar doses of PSGAG and 35S. Subsequently, the sulfated proteoglycan content of the medium and extracts of the explants was measured. Gel filtration chromatography was used to estimate the size and to purify the principal, large proteoglycan monomer, which was further characterized by digestion, using glycosidic enzymes. In a second experiment, explants were incubated with 35S for 48 hours, and were subsequently exposed to the same concentrations of the PSGAG for an additional 48 hours. The amount of remaining labeled proteoglycan was determined for culture medium and cartilage extracts. Gel filtration chromatography was used to assess the hydrodynamic size of the large proteoglycan monomer. Aliquots of proteoglycans from the second experiment were incubated in high-molecular weight hyaluronate and chromatographed to assess reaggregation. Polysulfated glycosaminoglycan caused a significant (P < 0.04) decrease in sulfated proteoglycan synthesis by cartilage explants. Radioactive proteoglycan content in explants labeled prior to exposure to PSGAG were similar. Large proteoglycan monomer size was similar in both experiments (median partition coefficient [K(AV)] = 0.40), and was not influenced by PSGAG treatment. Prelabeled explants exposed to hyaluronate and chromatographed under associative conditions had similar proportions of the radiolabel eluting as proteoglycan aggregate. Enzymatic digestion of newly synthesized large monomer revealed a mild dose-dependent increase in the proportion of keratan sulfate substitution on core protein. It was concluded that PSGAG in vitro, at the dosages evaluated, caused a decrease in proteoglycan synthesis, had little effect on labeled proteoglycan degradation, did not influence the size of large monomer, and caused a modest increase in the concentration of keratan sulfate in proteoglycans synthesized by osteoarthritic equine chondrocytes.

Fifteen isolates of Escherichia coli obtained from the blood and tissues of septic foals had plasmid DNA of size ranging from 2.5 to 93 megadaltons. These isolates grew in normal equine serum (serum resistant), a trait previously documented to be expressed by isolates obtained from blood and tissues of septic foals, but not by isolates obtained from the feces of clinically normal horses. Of these isolates, 3 contained conjugal plasmids that encoded resistance to multiple antimicrobial agents linked to serum resistance and, in 1 isolate, to production of aerobactin as well. Serum resistance and production of aerobactin are related to virulence of septicemic E coli from non-equine sources.

Fifteen isolates of Escherichia coli obtained from the blood and tissues of septic foals had plasmid DNA of size ranging from 2.5 to 93 megadaltons. These isolates grew in normal equine serum (serum resistant), a trait previously documented to be expressed by isolates obtained from blood and tissues of septic foals, but not by isolates obtained from the feces of clinically normal horses. Of these isolates, 3 contained conjugal plasmids that encoded resistance to multiple antimicrobial agents linked to serum resistance and, in 1 isolate, to production of aerobactin as well. Serum resistance and production of aerobactin are related to virulence of septicemic E coli from non-equine sources.