The invention of electric light has facilitated a society in which people work, sleep, eat, and play at all hours of the 24-hour day. Although electric light clearly has benefited humankind, exposures to electric light, especially light at night (LAN), may disrupt sleep and biological processes controlled by endogenous circadian clocks, potentially resulting in adverse health outcomes. Many of the studies evaluating adverse health effects have been conducted among night- and rotating-shift workers, because this scenario gives rise to significant exposure to LAN. Because of the complexity of this topic, the National Toxicology Program convened an expert panel at a public workshop entitled "Shift Work at Night, Artificial Light at Night, and Circadian Disruption" to obtain input on conducting literature-based health hazard assessments and to identify data gaps and research needs. The Panel suggested describing light both as a direct effector of endogenous circadian clocks and rhythms and as an enabler of additional activities or behaviors that may lead to circadian disruption, such as night-shift work and atypical and inconsistent sleep-wake patterns that can lead to social jet lag. Future studies should more comprehensively characterize and measure the relevant light-related exposures and link these exposures to both time-independent biomarkers of circadian disruption and biomarkers of adverse health outcomes. This information should lead to improvements in human epidemiological and animal or in vitro models, more rigorous health hazard assessments, and intervention strategies to minimize the occurrence of adverse health outcomes due to these exposures.

[ 1 ] The 5,575-km 2 Acad emy of Sciences Ice Cap is the largest in the Russia n Arctic. A 100-M Hz airbo rne radar, digital Landsat imagery, and satelli te syntheti c apertu re radar (SAR) interfero metry are used to investig ate its form and flow, incl uding the propor tion of mass lost throu gh iceberg calvi ng. The ice cap was covered by a 10-km -space d grid of radar fli ght paths , and the centr al porti on was covered by a grid at 5-km inte rvals: a total of 1,657 km of radar data. Digit al elevation model s (DEM s) of ice surfa ce elevation, ice thickness, and bed elevat ion data sets wer e produce d (cell size 500 m). The DEMs wer e used in the selec tion o f a deep ice core dril l site. To tal ice cap volum e is 2,1 84 km 3 ( $ 5.5 mm sea level equiva lent). The ice cap has a single dome reachi ng 749 m. Maxi mum ice thic kness is 819 m. About 200 km, or 42%, of the ice margin is mari ne. About 50% of the ice cap bed is below sea level. The central divide of the ice cap and severa l maj or drain age basins, in the south an d east of the ice cap and of up to 975 km 2 , are delimited from satel lite ima gery. The re is no eviden ce of past surge ac tivity on the ice cap. SAR interfero metric frin ges and phase-unwrapped velociti es for the whole ice cap indicate slow flow in the interior and much of the margin, punctu ated by four fast flowin g featu res with late ral shear zones and maxi mum veloci ty of 140 m y r À 1 . These ice streams extend back into the slower moving ice to wi thin 5 -10 km of the ice cap crest. They have lengths of 17 -37 km a nd widths of 4 -8 km. Mass flux from these ice streams is $ 0.54 km 3 yr À 1 . Tabular icebergs up to $ 1.7 km long are produce d. Total iceberg flux from the ice cap is $ 0.65 km 3 yr À 1 and probably represe nts $ 40% of the overal l mass loss , with the rema inder coming from surfa ce mel ting. Drivin g stre sses are general ly low est (<40 kP a) close to the ice cap divides and in severa l of the ice stre ams. Ice stream motion is likely to incl ude a signifi cant basal compo nent and may involve deformabl e marine sedim ents.

[1] In this study, 3370 km of 100 MHz ice-penetrating radar data were acquired from Devon Ice Cap, Arctic Canada. Bed returns were obtained from >90% of flight tracks. Mean crossing point errors in ice surface elevation and ice thickness were 7-8 m. Digital elevation models of ice cap surface and bed elevation, and ice thickness, were produced and can be used as boundary conditions in numerical modeling. Devon Ice Cap, including 1960 km 2 of contiguous stagnant ice to its west, is 14,010 km 2 . The ice cap proper is 12,050 km 2 . Its largest drainage basin is 2630 km 2 . The ice cap crest has a maximum measured elevation of 1921 m. Maximum recorded ice thickness is 880 m. Ice cap volume is 3980 km 3 (about 10 mm sea level equivalent). The bed, 8% of which lies below sea level, is an upland plateau dissected by steep-sided valleys that control the locations of the major outlet glaciers which dominate ice cap drainage. About 73 km, 4%, of the ice cap margin ends in tidewater. The margin is not floating. Icebergs of <100 m were observed offshore. Only a single outlet glacier showed signs of past surge activity. All major outlet glaciers along the eastern ice cap margin have retreated 1-3 km since 1960. Synthetic aperture radar-interferometic velocity structure shows slow, undifferentiated flow predominating in the west and center, with fast-flowing outlet glaciers and intervening slow-flowing ridges typical elsewhere. Outlet glacier velocities are 7-10 times higher than in areas of undifferentiated flow. This velocity structure, of fast-flowing units within slower-flowing ice, appears typical of many large (10 3 km 2 ) Arctic ice caps.

Studies of rodent photoperiodism almost without exception have employed fixed day lengths (DLs) and abrupt transitions from long to short DLs. Because the natural progression of changes in DL carries predictive information and may have physiological consequences, we determined seasonal fluctuations in testis size and body weight in Siberian hamsters maintained in a year-long pattern of increasing and decreasing DLs. A cycle of gonadal development, regression, and recrudescence, and corresponding changes in body weight, were observed in hamsters maintained in separate simulated natural photoperiods (SNPs) in which DLs neither fell below nor exceeded the putative critical DL of 13 h. Gradually decreasing DLs as long as 15.3 h induced gonadal regression, and DLs as short as 12.3 h supported ponderal growth, depending on the hamster's prior photoperiodic history. DLs experienced by hamsters during development influenced the adult incidence of responsiveness to short DLs. Photorefractoriness to short DLs occurred earlier in hamsters kept in static 10-h than in 12-h DLs. Increasing DLs in winter had little impact on the rate of gonadal recrudescence and weight gain. These data extend earlier investigations in showing the photoperiodic history determines gonadal responses over a broad range of DLs and influences gonadal responsiveness to short DLs and the triggering of the interval timer underlying recrudescence.