Search

Despite much recent research on pollination, we have amassed relatively little hard information about how animal pollinators contribute to angiosperm speciation and species distinctiveness. We usually assume that pollinators make important contributions by specializing on plant species and providing differences upon which plants specialize. But evidence for generalization in many plant-pollinator interactions calls this assumption into question. Generalization makes strong 'floral isolation' unlikely during and after the speciation process, and is seen by some workers as problematical for a scenario of pollinator-mediated disruptive selection leading to species formation. The timeis ripe to initiate several lines of further research. It will be valuable to characterize selection on plants more carefully at various spatial scales to understand whether different suites of pollinators can affect consistent selection in different directions on floral characteristics. Similarly, it will be valuable to explore how reproductive isolation arises within and among populations, races, and higher taxa of plants. Even if they are not strong agents of floral isolation, pollinators may contribute to the evolution of reproductive isolation in other ways, which largely remain to be discovered. One possibility is that reproductive isolation evolves more or less independently of diversification in reproductive characteristics. If so, floral adaptation may not be tightly coupled with the formation of biologicalspecies in angiosperms, as is often supposed. [ABSTRACT FROM AUTHOR]

Background: This study aimed to estimate the relative efficacy of neoadjuvant nivolumab in combination with chemotherapy (neoNIVO + CT) compared to relevant treatments amongst resectable non-metastatic non-small-cell lung cancer (rNSCLC) patients., Methods: Treatment comparisons were based on a network meta-analysis (NMA) using randomized clinical trial data identified via systematic literature review (SLR). The outcomes of interest were event-free survival (EFS) and pathological complete response (pCR). NeoNIVO + CT was compared to neoadjuvant chemotherapy (neoCT), neoadjuvant chemoradiotherapy (neoCRT), adjuvant chemotherapy (adjCT), and surgery alone (S). Due to the potential for effect modification by stage, all-stage and stage-specific networks were considered. Fixed-effect (FE) and random-effects Bayesian NMA models were run (EFS = hazard ratios [HR]; pCR = odds ratios [OR]; 95% credible intervals [CrI])., Results: Sixty-one RCTs were identified (base case = 9 RCTs [ n = 1978 patients]). In the all-stages FE model, neoNIVO + CT had statistically significant EFS improvements relative to neoCT (HR = 0.68 [95% CrI: 0.49, 0.94]), S (0.59 [0.42, 0.82]), adjCT (0.66 [0.45, 0.96]), but not relative to neoCRT (HR = 0.77 [0.52, 1.16]). NeoNIVO + CT (5 RCTs) had statistically significant higher odds of pCR relative to neoCT (OR = 12.53 [5.60, 33.82]) and neoCRT (7.15 [2.31, 24.34]). Stage-specific model findings were consistent., Conclusions: This NMA signals improved EFS and/or pCR of neoNIVO + CT relative to comparators among patients with rNSCLC.

(1) Background: Most patients with mycosis fungoides (MF), a form of cutaneous T-cell lymphoma (CTCL), develop relapsed/refractory (R/R) disease following front-line systemic therapy. This report describes treatment patterns and outcomes from the subpopulation with R/R MF. (2) Methods: This observational, retrospective, cohort study analyzed patient records (1984-2016) from 27 clinical sites in Europe. Outcomes included treatments received, response to first-, second- and third-line treatment, overall survival (OS) and progression-free survival (PFS). (3) Results: Of 104 patients with MF, 100 received second-line and 61 received third-line therapy. The median (range) times from the start of first-line therapy to the first R/R MF and from the first to the second R/R MF were 11.2 (0.3-166.5) and 13.5 (0.0-174.6) months, respectively. Second-and third-line treatment options varied and comprised systemic therapies (85% and 79% of patients, respectively), radiotherapy (32% and 34%, respectively) and topical therapies (48% and 36%, respectively). The median (95% confidence interval [CI]) OS from the diagnosis of the first R/R MF was 11.5 (6.5-not reached [NR]) years and was higher with non-chemotherapy (NR) versus chemotherapy (6.5 years); the estimated median PFS (95% CI) from the time of the first R/R MF was 1.3 (1.0-2.1) years. (4) Conclusions: High rates of R/R disease were observed after second- and third-line treatments in this real-world cohort, with longer median OS in patients receiving non-chemotherapy treatment versus chemotherapy. Following the standard management of MF and using recently approved targeted therapies can help improve patient outcomes in advanced-stage MF.

Background: Cladribine tablets and fingolimod have similar marketing authorisations in Europe for the treatment of patients with highly active relapsing multiple sclerosis (HA-RMS). In the absence of direct head-to-head studies, real-world data are important to assess the comparative effectiveness of these oral disease-modifying therapies (DMTs). The primary objective of the present study was to compare relapse rates between patients who received either cladribine tablets or fingolimod., Methods: This multicentre retrospective study conducted in the United Kingdom and Germany assessed non-inferiority in relapse rates of cladribine tablets versus fingolimod in HA-RMS patients over a 12-month period. Eligible patients who initiated treatment with cladribine tablets or fingolimod at least 12 months prior to the screening date were sampled consecutively until the target sample size was reached. Patients were censored at discontinuation of study treatment, commencement of another DMT, death, loss to follow-up, or at 12 months post-baseline, whichever happened earliest. The primary analytic timeframe for physician-confirmed relapse outcomes was the study effectiveness period (nine months of follow-up after an initial 12 weeks of treatment). Propensity score analysis was applied based on the inverse probability of treatment weighting approach., Results: The primary analytic cohort consisted of 1,095 HA-RMS patients: 610 (55.7%) receiving cladribine tablets and 485 (44.3%) receiving fingolimod. Fewer patients discontinued cladribine tablets and/or switched to another DMT compared with fingolimod (0.2% versus 3.5%, respectively). The primary endpoint, adjusted annualised relapse rate (ARR), was 0.10 (95% confidence interval [CI]: 0.07-0.14) for cladribine tablets and 0.14 (95% CI: 0.10-0.20) for fingolimod. The adjusted ARR ratio of cladribine tablets versus fingolimod was 0.68 (95% CI: 0.42-1.11). Given the entire 95% CI was less than the non-inferiority margin of 1.2, cladribine tablets was non-inferior to fingolimod., Conclusions: In this real-world retrospective cohort study, cladribine tablets demonstrated comparable effectiveness to fingolimod at one year following treatment initiation. The full treatment dosage of cladribine tablets is completed over two years and so these results may be conservative., Competing Interests: Declaration of Competing Interest WJB, AH, NW, SK, ZK, JD, and SM received consulting fees for model development from EMD Serono Research & Development Institute, Inc., Billerica, MA, USA, an affiliate of Merck KGaA. SK was an employee of ICON plc, Ontario, Canada at the time the study was conducted, and is currently an employee of AstraZeneca, Mississauga, Canada. ZK was employed by ICON plc at the time this study was conducted. BH was an employee of EMD Serono Research & Development Institute, Inc., Billerica, MA, USA, an affiliate of Merck KGaA, at the time the study was conducted. GTH is an employee of Merck Healthcare KGaA, Darmstadt, Germany. SM was an employee of ICON plc, Blue Bell, PA, USA at the time the study was conducted, and is currently an employee of Unlearn.AI, San Francisco, CA, USA, (Copyright © 2023. Published by Elsevier B.V.)

Objectives: To quantify the long-term comparative efficacy and safety of nivolumab in combination with ipilimumab (NIVO + IPI) relative to other immunotherapy (IO)-based regimens and chemotherapy in patients with first-line advanced non-small cell lung cancer (aNSCLC)., Methods: Phase 3 randomized controlled-trials (RCTs) with minimum 3-year follow-up evaluating IO-based regimens approved for first-line aNSCLC were identified via systematic literature review. Analytic populations were defined by levels of PD-L1 expression and histology. Due to presence of proportional hazards violations, time-varying hazard ratios (HRs) of overall survival (OS) and progression-free survival (PFS) were estimated via Bayesian fractional polynomial network meta-analysis. For safety endpoints, odds ratios (ORs) were estimated using indirect treatment comparisons (ITCs)., Results: CheckMate 227, KEYNOTE-189, KEYNOTE-407, KEYNOTE-024, KEYNOTE-042, and IMpower150 were included in the base case analysis. For OS and PFS, HRs of NIVO + IPI relative to other IO-based regimens trended downward over time across analytic populations. The 36-month OS HRs of NIVO + IPI versus comparators were: 0.69 (95 % credible interval: 0.47, 1.00) versus pembrolizumab + chemotherapy and 0.65 (0.45, 0.93) versus atezolizumab + bevacizumab + chemotherapy in the non-squamous and PD-L1 all-comers population; 0.73 (0.53, 1.02) versus pembrolizumab + chemotherapy in the squamous and PD-L1 all-comers population; and 1.05 (0.83, 1.32) versus pembrolizumab in the mixed histology and PD-L1 ≥ 50 % population. For PFS, 36-month HR point estimates ranged from 0.46 to 0.85 (only statistically significant versus pembrolizumab + chemotherapy in the squamous population; 0.46 [0.31, 0.69]). Adverse events (AEs) leading to discontinuation were not statistically significantly different between NIVO + IPI and pembrolizumab + chemotherapy, nor between NIVO + IPI and pembrolizumab monotherapy, although treatment-related grade ≥ 3 AEs were higher with NIVO + IPI than pembrolizumab monotherapy (OR = 2.21 [1.30, 3.75])., Conclusions: This study indicates trends towards long-term benefit with NIVO + IPI compared with other IO-based combinations, with manageable toxicities., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 The Author(s). Published by Elsevier B.V. All rights reserved.)

Introduction: The study objective was to estimate the relationship between objective response and survival-based endpoints by drug class, in first-line advanced non-small cell lung cancer (aNSCLC)., Materials and Methods: A systematic literature review identified randomized controlled trials (RCTs) of first-line aNSCLC therapies reporting overall survival (OS), progression-free survival (PFS), and/or objective response rate (ORR). Trial-level and arm-level linear regression models were fit, accounting for inclusion of immunotherapy (IO)-based or chemotherapy-only RCT arms. Weighted least squares-based R 2 were calculated along with 95% confidence intervals (CIs). For the main trial-level analysis of OS vs. ORR, the surrogate threshold effect was estimated. Exploratory analyses involved further stratification by: IO monotherapy vs. chemotherapy, dual-IO therapy vs. chemotherapy, and IO + chemotherapy vs. chemotherapy., Results: From 17,040 records, 57 RCTs were included. In the main analysis, trial-level associations between OS and ORR were statistically significant in both the IO-based and chemotherapy-only strata, with R 2 estimates of 0.54 (95% CI: 0.26-0.81) and 0.34 (0.05-0.63), respectively. OS gains associated with a given ORR benefit were statistically significantly larger within IO vs. chemotherapy comparisons compared to chemotherapy vs. chemotherapy comparisons (p < 0.001). Exploratory analysis suggested a trend by IO type: for a given change in ORR, 'pure' IO (IO monotherapy and dual-IO) vs. chemotherapy RCTs tended to have a larger OS benefit than IO + chemotherapy vs. chemotherapy RCTs. For ORR vs. PFS, trial-level correlations were strong in the IO-based vs. chemotherapy (R 2  = 0.84; 0.72-0.95), and chemotherapy vs. chemotherapy strata (R 2  = 0.69; 0.49-0.88). For OS vs. PFS, correlations were moderate in both strata (R 2  = 0.49; 0.20-0.78 and R 2  = 0.49; 0.23-0.76)., Conclusion: The larger OS benefit per unit of ORR benefit in IO-based RCTs compared to chemotherapy-only RCTs provides an important addition to the established knowledge regarding the durability and depth of response in IO-based treatments., (Copyright © 2022 Elsevier B.V. All rights reserved.)

Aim: The aim of this systematic literature review was to describe treatment patterns in nonmetastatic non-small-cell lung cancer. Methods: A search was conducted in MEDLINE and EMBASE. Eligible studies were multicentered (>50 patients) and conducted after 2000 in North America, Europe and Asia. Results: Twenty studies met the eligibility criteria. Based on US and Canadian studies in the resectable population, the proportion of patients who received neoadjuvant chemotherapy/chemoradiotherapy and adjuvant chemotherapy/chemoradiotherapy increased with increasing stage (i.e., from <3% in stage I to about 40% in stage III and from 15% in stage I to 30% in stage III, respectively). Within the resectable population, the breakdown between bimodal and trimodal therapy was variable, suggesting that clinical practice is not uniform. Conclusion: Overall, studies were heterogeneous, precluding data extrapolation across regions. Despite heterogeneity and limited evidence, this review suggested an increase in neoadjuvant and adjuvant chemotherapy with increasing stage, generally in line with treatment guidelines.

The treatment pattern of cutaneous T-cell lymphoma (CTCL) remains diverse and patient-tailored. The objective of this study was to describe the treatment patterns and outcomes in CTCL patients who were refractory or had relapsed (R/R) after a systemic therapy. A retrospective chart review study was conducted at 27 sites in France, Germany, Italy, Spain and the United Kingdom (UK) of patients who received a first course of systemic therapy and relapsed or were refractory. Data were collected longitudinally from diagnosis to first-, second- and third-line therapy. The study included 157 patients, with a median follow-up of 3.2 years. In total, 151 proceeded to second-line and 90 to third-line therapy. In the first line ( n = 147), patients were treated with diverse therapies, including single- and multi-agent chemotherapy in 67 (46%), retinoids in 39 (27%), interferon in 31 (21%), ECP in 4 (3%), corticosteroids in 3 (2%) and new biological agents in 3 (2%). In the second line, the use of chemotherapy and retinoids remained similar to the first line, while the use of new biologics increased slightly. In sharp contrast to the first line, combination chemotherapy was extremely diverse. In the third line, the use of chemotherapy remained high and diverse as in the second line. From the time of first R/R, the median PFS was 1.2 years and the median OS was 11.5 years. The presented real-world data on the current treatments used in the management of R/R CTCL in Europe demonstrate the significant heterogeneity of systemic therapies and combination therapies, as expected from the European guidelines.

Introduction: Secondary immunodeficiency diseases (SID) caused by hematological malignancies (HMs), stem cell transplant (SCT), and associated therapies are mainly characterized by the presence of hypogammaglobulinemia or antibody production deficits., Areas Covered: The authors summarized the scientific literature on disease burden of SIDs in patients who had HMs or SCT. Systematic searches were conducted to identify English-language articles from 1994-2020, reporting on clinical, humanistic, and economic burdens of SID due to HMs or SCT. Definitions of SID and serum immunoglobulin G thresholds varied across 24 eligible studies. In most (n = 16) studies, patients received immunoglobulin replacement therapy (IGRT). Several studies found IGRT was associated with significant reductions in rates of infection and antimicrobial use. However, 1 study found no statistically significant difference in antibiotic use with IGRT. Only 3 studies reported on quality of life, and no economic studies were identified., Expert Opinion: Overall, the findings show several beneficial effects of IGRT on clinical outcomes and quality of life; however, disparate definitions, infrequent reporting of statistical significance, and scarcity of clinical trial data after the 1990s present areas for further investigation. This paucity indicates an unmet need of current evidence to assess the benefits of IGRT in SID.

Aim: Randomized controlled trials (RCTs) with clinical outcomes are considered the gold standard for regulatory approval. However, by design they are only able to answer a small number of clinical questions. Other high-quality studies are required for clinical decision-making. The EVOLVE was the largest RCT, evaluating the effects of cinacalcet on clinical outcomes among adult patients receiving maintenance dialysis suffering from secondary hyperparathyroidism. While the EVOLVE trial did not reach its primary end point, imbalance in subjects' age at randomization and discontinuation rates are two of the reasons that the lack of mortality benefit is in question. We undertook a systematic literature review and Bayesian meta-analysis combining randomized and observational studies on the estimated effects of the oral calcimimetic cinacalcet on clinical outcomes including all-cause mortality, cardiovascular-related mortality, hospitalization for cardiovascular events, fracture and parathyroidectomy among patients on maintenance dialysis., Methods: Data sources included MEDLINE, EMBASE and Cochrane Central Register of Controlled Trials databases. RCTs and observational studies were included. Data extraction was completed by two authors independently and in duplicate determined the methodological quality of the studies and extracted data., Results: Of 564 unique citations identified, 16 studies were included: six observational studies and ten RCTs. Four high-quality studies (two observational and two RCTs) were deemed suitable for meta-analysis. Results indicated a statistically significant reduction in the risk of death associated with cinacalcet (hazard ratio: 0.83; 95% credible interval: 0.78-0.89)., Conclusion: The results of this meta-analysis indicate that treatment of secondary hyperparathyroidism with calcimimetic therapy may in fact reduce mortality among patients receiving maintenance dialysis. This finding provides justification for a well-designed and adequately powered randomized trial to definitively address the question.

Hepatitis C virus (HCV) infection is associated with substantial clinical and economic burden and is an important public health issue in Asia. The objective of this review was to characterize HCV epidemiology and related complications in China, Japan, South Korea and Taiwan. A search of electronic databases and conference abstracts identified 71 potentially relevant articles. Of those, 55 were included in the epidemiology review and 9 in the review of HCV-related complications. HCV prevalence in the general population was 1.6% in China, 0.6-0.9% in Japan, 0.6-1.1% in South Korea and 1.8-5.5% in Taiwan. Prevalence was higher for injecting drug users (48-90%) and those with human immunodeficiency virus coinfection (32-85%) and was lower for blood donors (<1%). Annual incidence of HCV in China was 6.01 per 100,000. HCV genotype 1b was associated with the highest incidence of hepatocellular carcinoma (HCC). Five-year survival for patients with liver cirrhosis was 73.8%, decreasing to 39.2% following liver transplantation; the majority of deaths were attributable to HCC. Limitations were that the majority of studies included in the epidemiology review were small, regional studies conducted in specific populations, and there was an absence of large population-based studies. Thus, estimates may not be representative of the epidemiology of HCV for each country. The prevalence HCV in China and HCV incidence in the Asian region remain largely unknown, and they are likely underestimated. Further epidemiologic and clinical data are needed to provide more precise estimates for use by public health agencies.

Importance: Surgical site infection (SSI) complicates 2-5% of surgeries in the United States. Severity of SSI ranges from superficial skin infection to life-threatening conditions such as severe sepsis, and SSIs are responsible for increased morbidity, mortality, and economic burden associated with surgery. Staphylococcus aureus (S. aureus) is a commonly-isolated organism for SSI, and methicillin-resistant S. aureus SSI incidence is increasing globally., Objective: The objective of this systematic review was to characterize risk factors for SSI within observational studies describing incidence of SSI in a real-world setting., Evidence Review: An initial search identified 328 titles published in 2002-2012; 57 were identified as relevant for data extraction. Extracted information included study design and methodology, reported cumulative incidence and post-surgical time until onset of SSI, and odds ratios and associated variability for all factors considered in univariate and/or multivariable analyses., Findings: Median SSI incidence was 3.7%, ranging from 0.1% to 50.4%. Incidence of overall SSI and S. aureus SSI were both highest in tumor-related and transplant surgeries. Median time until SSI onset was 17.0 days, with longer time-to-onset for orthopedic and transplant surgeries. Risk factors consistently identified as associated with SSI included co-morbidities, advanced age, risk indices, patient frailty, and surgery complexity. Thirteen studies considered diabetes as a risk factor in multivariable analysis; 85% found a significant association with SSI, with odds ratios ranging from 1.5-24.3. Longer surgeries were associated with increased SSI risk, with a median odds ratio of 2.3 across 11 studies reporting significant results., Conclusions and Relevance: In a broad review of published literature, risk factors for SSI were characterized as describing reduced fitness, patient frailty, surgery duration, and complexity. Recognition of risk factors frequently associated with SSI allows for identification of such patients with the greatest need for optimal preventive measures to be identified and pre-treatment prior to surgery.

Purpose: The aim of of this study was to evaluate prospectively the long-term outcome of mesocolic and pelvic diverticular abscesses of the left colon., Methods: Between October 1986 and October 1997, a total of 465 patients urgently admitted to our hospital with a suspected diagnosis of acute left-sided colonic diverticulitis had a CT scan. Of 76 patients (17 percent) who had an associated mesocolic or pelvic abscess, 3 were lost to follow-up. The remaining 73 patients (45 with a mesocolic abscess and 28 with a pelvic abscess) were followed for a median of 43 months., Results: of the 45 patients with a mesocolic abscess, 7 (15 percent) required surgery during their first hospitalization versus 11 (39 percent) of the 28 patients with a pelvic abscess (P = 0.04). At the end of follow-up, 22 (58 percent) of the 38 patients with a mesocolic abscess who had successful conservative treatment during their first hospitalization did not need surgical treatment vs. 8 (47 percent) of the 17 who had a pelvic abscess. Altogether, 51 percent of the patients with a mesocolic abscess had surgical treatment versus 71 percent of those with a pelvic abscess (P = 0.09)., Conclusions: Considering the poor outcome of pelvic abscess associated with acute left-sided colonic diverticulitis, percutaneous drainage followed by secondary colectomy seems justified. Mesocolic abscess by itself is not an absolute indication for colectomy.

Natural hybrid zones between related species illustrate processes that contribute to genetic differentiation and species formation. A common viewpoint is that hybrids are essentially unfit, but they exist in a stable tension zone where selection against them is balanced by gene flow between the parent species. An alternative idea is that selection depends on the environment, for example, by favoring opposite traits in the two parental habitats or favoring hybrids within a bounded region. To determine whether selection of hybrids is environment dependent, we crossed plants of naturally hybridizing Ipomopsis aggregata and I. tenuituba in the Colorado Rocky Mountains and reciprocally planted the seed offspring into a suite of natural environments across the hybrid zone. All types of crosses produced similar numbers and weights of seeds. However, survival of the offspring after 5 years differed markedly among cross types. On average, the F1 hybrids had survival and growth rates as high as the average for their parents. But hybrid survival depended strongly on the direction of a cross, that is, on which species served as the maternal parent. This fitness difference between reciprocal hybrids appeared only in the parental environments, suggesting cytonuclear gene interactions that are environment specific. These results indicate that complex genotype-by-environment interactions can contribute to the evolutionary outcome of hybridization.

The Ipomopsis aggregata species complex (Polemoniaceae) includes species pairs that hybridize readily in nature as well as pairs that meet along contact zones with no apparent hybridization. Artificial hybrids can be made between I. aggregata and I. arizonica, yet morphological intermediates between these two species have not been observed in natural populations. This apparent lack of hybridization is perplexing given that plants of the two species often grow within a few metres of each other and both species have red flowers visited by the same species of hummingbirds. We used trained hummingbirds to examine pollen transfer within and between species. We also hand-pollinated flowers to examine paternal success of heterospecific and conspecific pollen, testing paternity with electrophoretic examination of seeds. Hummingbirds were not simply better at transferring pollen within than between species. Instead, I. arizonica was a better pollen donor so that considerable pollen transfer was observed from I. arizonica to I. aggregata, but very little in the opposite direction. Conversely, once pollen arrived at stigmas, I. arizonica pollen performed very poorly on I. aggregata pistils. However, pollen from I. aggregata could, in some cases, sire seeds on I. arizonica. We hypothesize that hybrids are scarce in nature, in part, because of asymmetric barriers to reproduction: little pollen transfer in one direction and poor pollen performance in the other.

Outbreeding depression in progeny fitness may arise from disruption of local adaptation, disruption of allelic coadaptation, or a combination of these "environmental" and "physiological" mechanisms. Thus the minimum spatial scale over which outbreeding depression arises should depend on the spatial scale of gene dispersal and (with an environmental mechanism) of change in selection regimes. We previously reported substantial outbreeding depression in lifetime fitness of progeny resulting from crosses among parents separated by 100 m in natural populations of the herbaceous plant Ipomopsis aggregata. In this paper we explore the effect of crossing distance on fitness in two additional experiments begun in 1987 and 1990. We planted seed progeny derived from partial diallel crossing designs in randomized blocks in maternal environments and scored emergence of seedlings, survival, and eventual flowering of individuals over the subsequent six to eight years. Nested within each diallel design were crossing distances of 1 m, 10 m, and 100 m. Compared to 1-m and 10-m progeny, 100-m progeny of the 1987 diallel suffered a significant reduction in seedling emergence, and both 1-m and 100-m progeny that survived to flower achieved lower lambda-values on average than 10-m progeny. Total outbreeding depression suffered by 100-m relative to 10-m progeny was approximately 10%, compared to approximately 30% in our earlier study of I. aggregata. Progeny of 10-m crosses also outperformed 1-m and 100-m progeny of the 1990 diallel by approximately 5%, but no difference among crossing distance treatments was significant. Thus, the magnitude of outbreeding depression in 100-m crosses varied among experiments. This is not surprising given likely spatial and temporal variation in gene flow and selection regimes, different population histories, and different parental and progeny environments. Characterizing outbreeding depression on the shortest spatial scales over which it is expressed, as well as its variation and causes, is worthwhile because it promises to shed light on the earliest stages of angiosperm speciation.

Plant populations vary in density both naturally and as a consequence of anthropogenic impacts. Density in turn can influence pollination by animals. For example, plants in dense populations might enjoy more frequent visitation if pollinators forage most efficiently in such populations. We explored effects of plant density on pollination and seed set in the larkspur Delphinium nuttallianum and monkshood Aconitum columbianum. At our site in the Colorado Rocky Mountains, flowers of D. nuttallianum are pollinated primarily by queen bumble bees, solitary bees, and hummingbirds, whereas those of A. columbianum are pollinated primarily by queen and worker bumble bees. We found that the quantity of pollination service to both species (pollinator visitation rate and pollen deposition) was at best weakly related to density. In contrast, seed set declined by approximately one-third in sparse populations relative to nearby dense populations. This decline may stem from the receipt of low-quality pollen, for example, inbred pollen. Alternatively, sparsity may indicate poor environmental conditions that lower seed set for reasons unrelated to pollination. Our results demonstrate the value of simultaneously exploring pollinator behavior, pollen receipt, and seed set in attempting to understand how the population context influences plant reproductive success.

Flowers of many angiosperm species are well adapted to avoid self-pollination, for instance by temporal and spatial separation of pollen and stigma within the same flower. However, such adaptations do not prevent the transfer of pollen between flowers on the same plant (geitonogamy). Recent empirical studies, aided by advances in field techniques, statistical methods and modelling, show that geitonogamy often is substantial and increases with plant size. Selfing by geitonogamy incurs a fitness cost of reduced pollen export, and recent reports show that seed set can suffer as well, even in self-incompatible species. Geitonogamy has important implications for sex-allocation theory, the evolution of dioecy and other issues in evolutionary biology., (Copyright © 1993. Published by Elsevier Ltd.)