Amorphophallus johnsonii (N. E. Brown) flowers during April in the main rainy season in Ghana. Anthesis starts at dusk with fluid oozing from the upper spadix accompanied by a strong aminoid odour. Just after dark large numbers of carrion beetles (Phaeochrous amplus) and occasional dung fly species (Hemigymnochaeta unicolor and Paryphodes tigrinus) visit the inflorescences. The beetles become trapped in the lower spathe overnight and remain in the spadix until the following evening. Between 1630 and 1645 h the following day, the anthers produce long threads of sticky pollen. The trapped beetles escape just after dark by crawling up the spadix, past the dehisced anthers and fly away from the spadix tip. Marked beetles were seen to transfer pollen from male phase to female phase inflorescences. Successful fertilisation was only effected if pollen was transferred on the same night from a male inflorescence 30 m or less away. Pollen is psilate and typical of beetle pollinated Araceae. Berries ripen approximately 70 d after fertilization and ripen basisetally in the infructescence.

The life history of Cora marina was followed for one year in two permanent streams at 600 m elevation in Guanacaste National Park; Costa Rica. The water temperature was c. 21°C year-round, but there was a distinct wet-dry seasonality, very little rain falling from January to May. In both streams, C. marina was univoltine. Adults first appeared at the beginning of May and the flight period coincided with the wet season. The availability of water-soaked logs as oviposition sites probably restricts reproduction to the wet season. Eggs hatched from mid-July to December. Recruitment to subsequent larval instars was slow during the wet season but increased at the start of the dry season. Final-instar larvae were collected from March to October. Oviposition in logs above the stream and the ability to live in the low oxygen conditions of the hyporheic zone probably allow eggs and small larvae to survive wet season spates.

El ciclo biológico de una libelula tropical: Cora marina (Odonata: Polythoridae) en Guanacaste, Costa Rica.

Two formicarial ant-plants endemic to Panama and adjacent areas of Costa Rica co-occur in mid-elevation cloud forest yet have distinctive modifications to house and feed small, retiring ants. Besleria formicaria (Gesneriaceae) hosts only Pheidole sp. which harvest the oil-rich seeds of its host and other plants and insect parts. Hoffmannia vesciculifera (Rubiaceae) hosts only Solenopsis sp. ants which obtain lipid and possibly other nutrients from vescicular tissues. Judging by their abundance in refuse deposits, fern sporangia appear to be a plentiful and evenly distributed resource exploited by both ant species.

Changes in the composition and structure of mid-altitude, semi-deciduous tropical forest in Mpanga Research Forest Reserve, Uganda are described for a 25-year period between 1968 and 1993. Three surveys of a 0.64 ha (80 m × 80 m) permanent plot were carried out in 1968, 1982 and 1993, during which a total of 397 trees with a dbh≥9.5 cm were identified. Forty-nine species were identified in total, representing 19 families. Eight individuals remain unidentified. Dbh measurements were recorded for 359 non-buttressed trees, whilst the equivalent measurement for 38 buttressed trees was the diameter of the trunk immediately above the buttress. Basal area, diversity and density of trees increased within the plot during the survey period by, respectively, 8% (from 39.2 to 42.2 m2 ha−1). 7% (from 44 to 47 species) and 11% (459 to 508 trees ha−1). The main compositional changes were increases in understorey trees and a decline in serai taxa. Growth rates (productivity) and turnover were lower during the period 1982 to 1993 than the period 1968 to 1982. Rates of growth and mortality were generally highest in serai species and lowest in main canopy taxa. Mortality rates were also highest amongst the smallest trees enumerated (dbh <30 cm). Changes in composition and structure over the survey period are believed to reflect forest recovery after low intensity pit-sawing was curtailed from 1951. Forest recovery also may have caused the reduced growth rates and turnover recorded for the most recent survey period, by restricting the opportunities for light-demanding, faster-growing and relatively productive serai taxa, and to have outweighed any effects of externally-driven processes, such as changes in atmospheric conditions.

Clusters of four circular forest plots, 0.2 ha in total area, were inventoried for canopy tree (≥30 cm gbh) and terrestrial herb floristic composition at 46 sites in the Central Catchment Nature Reserve, Singapore. The Nature Reserve covers a complex mosaic of lowland tropical forest of different successional stages, with much of the area covered in secondary forest 50–100 years old developed on exhausted agricultural soils. A total of 449 species of tree and 59 species of terrestrial herb were recorded from the clusters. Facultatively-terrestrial epiphytes were excluded from the analysis. Possible reasons for the low herb diversity are discussed. There was a marked similarity between ordinations (detrended correspondence analysis) of the clusters based on tree canopy composition weighted by species relative contribution to cluster total basal area and the herb flora composition, with a highly significant correlation between first axis scores of the two ordinations. Both ordinations showed the three clusters from freshwater swamp forest to be highly distinctive from the rest. The two forest strata exhibited a parallel response to the successional gradient, though the low diversity and patchy distribution of the herb flora blurred the distinction between primary and secondary communities more clearly seen in canopy composition.

The relationship between sepal and nut size of fruits was studied in fruit of 394 species of Dipterocarpaceae in Malesia and Sri Lanka. The fruits of many dipterocarps have long twisted sepals which have a role as blades for dispersal by gyration. But long sepals decrease resource allocation to other parts, and increase the chance of fruit being trapped in the canopy. The frequency distribution of the ratio of sepal length to nut length was bimodal. The gyrationdispersed group with longer sepals comprised 74% of Dipterocarpaceae studied. The group with short sepal mainly comprised understorey trees, but there were a few emergents. In both groups, sepal areas were expressed as a power function of nut volumes'. Because fruit weight per sepal area increases with fruit size, larger fruits fall more quickly than smaller fruits, and sepals of the former can not serve so effectively as propeller blades. This is consistent with the fact that a few emergent trees with large nuts have short sepals and are not dispersed by gyration. Therefore, the re-evolution of short fruit sepals appears to have been mainly due to two reasons: reduction of tree height and enlargement of nut size.

Vitt L. J., Zani P. A. & Caldwell J. P. Behaviour ecology of Tropidurus hispidus on isolated rock outcrops in Amazonia Vol. 12 Part 1: 81–101.

On page 86 the phrase: ‘Wyvern to set equation’

should read:

Cambridge University Press would like to apologise to the authors and readers for this oversight. Pages 85 and 86 will be reprinted and circulated.