Test of the temperature difference model predicting masting behavior

The differential temperature (ΔT) model, based on the assumption that masting plants respond to the difference in the temperatures during the growing seasons 1 and 2 years prior to seed production, has recently been proposed to explain the proximate factor of masting. In this study, we used a 28-year series of data on pollen cone and seed production in Cryptomeria japonica D. Don in Japan and compared several models based on temperatures and resources to test whether ΔT acts as a cue or is a proxy for resource limitation. Of all the models tested, models including ΔT, previous summer absolute temperature [T.sub.n-1], and reproduction in the previous year provided the best fit. The number of pollen cones and seeds produced was proportional to the difference in the mean daily maximum temperature during June to August between the preceding two years. In addition, to test whether the double mast events in consecutive years was less common than consecutive warm summers, we used our dataset and 12 datasets of pollen dispersal of 17 years or longer of C. japonica in Japan. Although consecutive warm summers occurred in 4.4% of pairs, double mast events occurred in 1.0%. The ΔT model was a considerably better predictor of the rare phenomenon of double mast events, which occur only after a specific sequence of cold-moderate-hot absolute summer temperatures. Thus, ΔT acts as cue for masting in C. japonica. Key words: climate cue, Cryptomeria japonica, ΔT model, Japanese cedar, masting. Le modele de temperature differentielle (ΔT), fonde sur l'hypothese que les plantes reagissent a la difference de temperature entre les deux saisons de croissance qui ont precede une annee semenciere, a recemment ete propose pour expliquer la cause directe de ce phenomene. Dans cette etude, nous avons utilise une serie de donnees portant sur 28 annees de production de cones males et de graines chez Cryptomeria japonica D. Don au Japon et nous avons compare plusieurs modeles fondes sur les temperatures ainsi que les ressources pour verifier si le modele ΔT agit comme un signal ou est un indice de ressources limitees. De tous les modeles testes, les modeles qui incluaient ΔT, la temperature absolue de l'ete precedent [T.sub.n-1] et la reproduction au cours de l'annee precedente donnaient le meilleur ajustement. Le nombre de cones males et de graines produits etait proportionnel a la difference de temperature journaliere maximum moyenne durant les mois de juin et aout entre les deux annees precedentes. De plus, pour determiner si les annees semencieres doubles consecutives etaient moins frequentes que les etes chauds consecutifs, nous avons utilise notre base de donnees et 12 bases de donnees sur la dispersion du pollen sur 17 annees ouplus pour C. japonica au Japon. Bienque des etes chauds consecutifs soient survenus dans 4,4% des paires, les annees semencieres doubles sont survenues dans 1% des cas. Le modele ΔT predisait beaucoup mieux le phenomene rare des annees semencieres doubles qui survenaient seulement a la suite d'une sequence specifique de temperatures estivales absolues: froides-moderees-chaudes. Par consequent, ΔT agit comme un signal pour les annees semencieres chez C. japonica. [Traduit par la Redaction] Mots-cles: signal climatique, Cryptomeria japonica, modele ΔT, cedre du Japon, annee semenciere.
Introduction Population-level fluctuations of large crops of seeds or flowers, known as masting or mass flowering, are common and important phenomena in plants (Silvertown 1980, Kelly and Sork 2002). Masting [...]