In the current context of global warming, an analysis is required of spatially-extensive and long-term blooming data in fruit trees to make up for insufficient information on regional-scale blooming changes and determinisms that are key to the phenological adaptation of these species. We therefore analysed blooming dates over long periods at climate-contrasted sites in Western Europe, focusing mainly on the Golden Delicious apple that is grown worldwide. On average, blooming advances were more pronounced in northern continental (10 days) than in western oceanic (6-7 days) regions, while the shortest advance was found on the Mediterranean coastline. Temporal trends toward blooming phase shortenings were also observed in continental regions. These regional differences in temporal variability across Western Europe resulted in a decrease in spatial variability, i.e. shorter time intervals between blooming dates in contrasted regions (8-10-day decrease for full bloom between Mediterranean and continental regions). Fitted sequential models were used to reproduce phenological changes. Marked trends toward shorter simulated durations of forcing period (bud growth from dormancy release to blooming) and high positive correlations between these durations and observed blooming dates support the notion that blooming advances and shortenings are mainly due to faster satisfaction of the heating requirement. However, trends toward later dormancy releases were also noted in oceanic and Mediterranean regions. This could tend toward blooming delays and explain the shorter advances in these regions despite similar or greater warming. The regional differences in simulated chilling and forcing periods were consistent with the regional differences in temperature increases.

No abstract

The occurrence of Fusarium species in winter wheat in southern Belgium (Wallonia) and the deoxynivalenol content in 692 samples collected in commercial fields in the region's main cereal growing area were investigated. The main Fusarium species associated with head blight in wheat were identified at levels that varied from year to year. Interactions between fungal species causing head blight in wheat were detected, most of them positive. The years 2007 and 2008 were very conducive to the disease and there was a strong correlation between mean annual deoxynivalenol content and number of days with a mean relative humidity above 80% over a period starting from 7 days before the mean flowering date and ending 16 days after this date. A two-stage approach, based on type of year (at risk or not) and agricultural practices during risk years has been developed to help cereal storage companies reduce the risk of mixing sound and deoxynivalenol-contaminated lots at harvest and to limit the number of analyses.

-Evaluation of BactoScan FC for the enumeration of bacteria in raw milk. The evaluation of BactoScan FC for routine enumeration of bacteria in raw milk was carried out according to the model proposed by the International Dairy Federation. The linearity of response of BactoScan FC was verified between 6 000 and 27 000 000 pulses . mL -1 without reaching the limits of detection which have not been determined. The transmission error was estimated at 0.09% and the repeatability at 0.084 log 10 CFU . mL -1 for a 60 000 CFU . mL -1 mean level of contamination. The stability of the equipment over time was verified; in both the short term and the long term, reproducibility is in the same order of magnitude as repeatability. The relationship between BactoScan FC and the reference method is linear over the whole range analysed (700 to 26 000 000 CFU . mL -1 ), with a determination threshold estimated at under 4 000 CFU . mL -1 . The regression equation between the reference method and the BactoScan FC is: log 10 (CFU . mL -1 ) = 0.839 × log 10 (pulses . mL -1 ) + 0.095. The accuracy of estimation of the reference value, expressed by the residual standard deviation of the regression, is 0.287 log 10 CFU . mL -1 . Finally, the independence of this estimate compared to the somatic cell level of the samples was verified. BactoScan FC / bacteriological quality / raw milkRésumé -L'évaluation du BactoScan FC pour la numération en routine des bactéries du lait cru a été réalisée selon le modèle proposé par la Fédération Internationale de Laiterie. La linéarité de réponse du BactoScan FC a été vérifiée entre 6 000 et 27 000 000 impulsions . mL -1 sans toutefois atteindre les limites de détection qui restent indéterminées. L'erreur de transfert a été estimée

No abstract