No abstract

Siljo was prepared by thoroughly cooking powdered horsebean (Vicia faba) in safflower (Carthamus tinctorius) extract into a semi-solid slurry and then adding to it untreated black mustard (Brassica nigra) powder after cooling to 50°C. The black mustard powder was the source of starter microorganisms, with Lactobacillus acidophilus, L. plantarum and L. delbruekii initiating and later dominating the fermentation process. The pH of the fermenting mass dropped to 4.5 within 36 h and reached 4.0 at 168 h. Aerobic mesophilic bacteria and lactic acid bacteria were each present at about 1×10(10) c.f.u./ml after 36 h of fermentation but Enterobacteriaceae were never detected. The dominant aerobic mesophilic flora consisted of Micrococcus, Bacillus and Lactobacillus spp. Crude protein, crude fat and ash increased slightly during the fermentation, with final values of around 28%, 25% and 7%, respectively, but there was a marked increase in protein availability and concentration during the fermentation.

Ersho is a clear, yellow liquid that accumulates on the surface of fermenting teff-flour batter and is collected to serve as an inoculum for the next fermentation. The pH of ersho samples was about 3.5 and titratable acidity ranged between 3.1% and 5.7%. The mean aerobic mesophilic counts from four households varied between 6.9×10(6) and 1.3×10(8) c.f.u./ml and the aerobic bacterial flora consisted of Bacillus spp. Mean yeast counts ranged between 5.2×10(5) and 1.8×10(6) c.f.u./ml and comprised, in order of abundance, Candida milleri, Rhodotorula mucilaginosa, Kluyveromyces marxianus, Pichia naganishii and Debaromyces hansenii. Candida milleri was the most dominant isolate in all samples. About 90% of the teff flour samples had aerobic mesophilic counts ≥10(5) c.f.u./g and Gram-positive bacteria constituted about 71% of the total isolates. About 80% of samples had Enterobacteriaceae counts of 10(4) c.f.u./g.

In fermenting tempeh made from non-acid-soaked horsebean, pea, and soybean Staphylococcus aureus grew rapidly to a final count of 108cfu.g-' or more, but growth was less when chickpea was used. Inoculation of the cooked beans with Lactobacillus plantarum markedly decreased S. aureus growth rate and the final count in non-acidsoaked horsebean and pea tempeh, strongly retarded growth in chickpea-, and completely inhibited growth in soybean-tempeh. Acid-soaking .the beans resulted in lower S. aureus growth, and inoculation with Lb. plantarurn completely inhibited it in soybean and reduced the counts in the other products to less than 104cfu.g-'. Acidity, pH and other substances produced by Lb. plantarum are believed to inhibit S. aureus in fermenting tempeh. Inoculation of beans with Lb. plantarum may be used to control S. aureus growth and enterotoxin production during commercial scale tempeh production.

Salmonella infantis and Escherichia coli multiplied to varying degrees during