我国严重威胁性外来入侵植物入侵与扩散历史过程重建及其潜在分布区的预测

外来生物入侵已经在世界范围内造成严重的危害，它不仅导致生物多样性的减少和丧失而且威胁着全球的生态环境和经济发展。入侵种在入侵区域的种群扩散是入侵种带来风险的最根本问题，也是其造成危害的重要原因。研究入侵物种的扩散规律可以了解其在入侵地的入侵状况和扩散趋势，这对制定合理的控制措施具有重要意义。本研究选取14种在我国具有较强威胁性的外来入侵植物，通过对其入侵历史动态的分析，判断其在我国的入侵与扩散阶段，进而预测其在我国的潜在分布区。这14种外来植物为紫茎泽兰（Ageratina adenophora）、土荆芥 （Chenopodium ambrosioides.）、喜旱莲子草（Alternanthera philoxeroides）、反枝苋（Amaranthus retroflexus）、刺苋（Amaranthus spinosus）、皱果苋（Amaranthus viridis）、北美独行菜（Lepidium virginicum）、藿香蓟（Ageratum conyzoides）、钻形紫菀（Aster subulatus）、小蓬草（Conyza canadensis）、一年蓬（Erigeron annuus）、牛膝菊（Galinsoga paviflora）、飞机草（Eupatorium odorata）和北美商陆（Phytolacca americana）。 　　紫茎泽兰最早于20世纪40年代入侵我国，经过1940—1960年长达20年的时滞期，紫茎泽兰开始在云南及其临近的省份如四川、贵州和广西迅速扩散。其中，紫茎泽兰在南亚热带和中亚热带气候条件下的扩散速度为20公里/年，而在垂直地带性北亚热带地区的扩散速度为6.8公里/年。紫茎泽兰仍没有入侵到垂直地带性暖温带地区。尽管1990年后，紫茎泽兰在云南基本停止扩散，但其在邻近省（市）的快速传播表明紫茎泽兰在我国仍处在扩散阶段，还没有达到饱和阶段。生态位模型预测结果和紫茎泽兰在已经入侵地区扩散的地理生态式样基本相符。我国南部及中南部地区的气候条件十分适合紫茎泽兰生长。而在华中地区，由于环境条件不太适宜，其扩散速度会相对较慢。我国北部和西北部的气候条件完全不适合紫茎泽兰生存。因此，我国南部和中南部广大未入侵地区将会受到紫茎泽兰入侵和快速扩散的严重威胁，对此应该立即采取紧急措施。 　　自从紫茎泽兰于1978年首次入侵四川省以来，已经在四川、重庆和湖北大面积扩散。紫茎泽兰在此新近入侵区域主要沿河流、公路、铁路扩散。其中沿安宁河、108国道和成昆铁路向北扩散的平均速度为19公里/年，而沿金沙江－长江向东北扩散的平均速度为33公里/年。特别是2000年以后，紫茎泽兰沿金沙江－长江的扩散速度达到了88公里/年。紫茎泽兰得以如此迅速扩散是其本身生物学特性和当地地理生态特点多种因素之间一系列耦合关系共同作用的结果。人类活动和1998年金沙江－长江特大洪水也促进了紫茎泽兰在该地区的扩散。另一方面，紫茎泽兰在新近入侵地区的垂直分布范围说明随海拔高度上升所引起的垂直地带性气候变冷是紫茎泽兰入侵和扩散的自然限制条件。 　　此外，通过对其余13种外来入侵植物入侵历史过程重建发现，除了牛膝菊、反枝苋和皱果苋最早入侵地位于内陆地区以外，其余10种外来入侵植物的最早入侵地都位于我国的东部、南部沿海和西南边境地区。这13种外来植物通常不是通过单一途径传入，而是通过两种或多种途径传入不同的地点。相应地，入侵后也呈现不同的扩散模式。这13种外来植物在我国大体上包括5种入侵和扩散式样即1）从南部沿海和西南边境分别向内陆扩散，如霍香蓟、飞机草，北美商陆;2）从东部沿海向内陆扩散，例如钻形紫菀，小蓬草、北美独行菜和喜旱莲子草;3）从东部沿海、南部沿海和西南边境分别向内陆扩散，如刺苋;4）从南部沿海向内陆扩散，如土荆芥;5）从最早入侵的内陆地区向周边地区扩散，如反枝苋、牛膝菊和皱果苋。分布区动态分析显示当前这13种外来入侵植物在我国都还处在扩散阶段，而且有些外来植物还处在快速扩散阶段，如飞机草。 　　在上述研究基础上，我们对这13种外来入侵植物开展了生态位模型适生区预测和比较分析。并且结合其扩散动态和当前分布现状，我们对其潜在的分布区及扩散趋势进行了系统分析和预测。飞机草、土荆芥、霍香蓟、牛膝菊和北美商陆，这5种外来植物的现有分布区明显小于其在我国的适生区，因此它们在我国的潜在分布区很广，应给予足够的重视。喜旱莲子草、反枝苋、皱果苋、北美独行菜、一年蓬、钻行紫菀和北美商陆在我国的入侵范围与其适生区域之间基本重合，但在其入侵范围内仍存在大面积的未入侵区域。这些未入侵地区通常被已入侵区域所包围，因此也应当给予足够的重视。刺苋和小蓬草在我国的潜在的分布区相对较小，这些植物已经基本上完全入侵了其适生区。因此，这些植物在我国向临近区域继续扩散的可能性不大。 　　

陕北黄土丘陵沟壑区水土保持耕作及施肥下农田土壤种子库特征

为了探明多年免耕下农田恶性杂草发生的机理,提高保护性耕作下作物对农田恶性杂草持久稳定的抑制效果,依据陕西安塞田间4a的定位试验,采用小区调查取样和室内实验相结合的方法,从物种组成、密度特征、多样性以及相似性特征等方面,研究了黄土丘陵旱作农区大豆(Glycine max)、玉米(Zea mays)、红小豆(Semen Phaseoli)、马铃薯(Solanum tuberosum)在翻耕化肥(CF)、翻耕有机肥(CM)、翻耕无肥(CN)、免耕化肥(NF)、免耕有机肥(NM)、免耕无肥(NN)等水平下的农田土壤种子库。结果表明:(1)4种作物24种土样中共萌发出12个物种1965株幼苗,隶属于7科12属。1年生杂草占94%,棒头草(fugax nees ex steud)、苋菜(Acalypha australis)、马唐(Digitaria sanguinalis)、早熟禾(Poasphondylodes)为优势种,占87%。(2)在0～20cm土层不同处理间,土壤种子库的密度变动于(282.9±63.4)～(7482.5±1078.3)粒.m-2,其中,红小豆小区>马铃薯小区>大豆小区>玉米小区;翻耕小区>免...

不同基因型小麦麦茬对杂草的化感抑制作用

采用改进的田间抑制圈生物测定方法,分析了10种不同基因型小麦材料麦茬对杂草生长的影响。结果表明,不同基因型之间以及同一基因型的不同材料之间对杂草的影响均呈现显著差异,说明化感作用基因在稳定遗传的同时存在一定的变异。以现代品种“丰产3号”麦茬地杂草的生物量为对照,其他小麦品种麦茬对杂草均有一定的抑制作用;对10个小麦品种麦茬地杂草的生物量与残茬滞留时间(天)逐步回归分析发现,杂草生长与时间存在明显的正相关关系,其中小麦残茬抑制杂草生长的化感作用有效天数为3～29天,而后杂草生物量逐渐增大,最后达到稳定;染色体组型为 AA、AABB 及 AABBDD 的麦茬化感抑制作用均随残留时间的变化而逐渐减弱,抑制杂草的平均有效期分别为21、24和25天,说明小麦在从二倍体到四倍体到六倍体的染色体组变异与倍体变化的长期进化过程中,小麦的遗传可塑性可能有助于化感作用的增强。

瑞香狼毒（Stellera chamaejasme L.）对高寒草甸系统生长和非生长季节碳氮循环的影响及其花粉化感效应

瑞香狼毒（Stellera chamaejasme L.）是瑞香科（Thymelaeaceae）狼毒属的一种多年生野草，有毒。据调查，从20 世纪60 年代开始至今，狼毒在青藏高原东缘的高寒草甸上不断蔓延、密度不断变大，在一些地段甚至成为优势物种。有关狼毒在高寒草甸蔓延的生态系统效应的研究尚未见报道。本文从系统碳、氮循环的角度，分别研究狼毒在生长和非生长季节对高寒草甸生态系统的影响。同时，从花粉化感的角度，深入研究狼毒对当地同花期物种有性繁殖的影响。系统地研究高寒草甸生态系统物质循环过程，特别是非生长季节微生物和土壤碳氮库的动态变化，有助于揭示狼毒在系统物质循环方面的“物种效应”以及这种效应的季节变化，为丰富有关高海拔生态系统，特别是其非生长季的物质循环的科学理论做出贡献。同时，碳氮循环和花粉化感的研究还有助于深刻地理解狼毒作为一种入侵性很强的杂草的特殊的蔓延机制，从而为狼毒的有效防治、高寒草甸的科学管理提供依据。 针对狼毒在青藏高原高寒草甸上蔓延的生态系统碳氮循环方面的影响，开展以下2 方面的研究：（1）在生长季，研究松潘县尕米寺附近（北纬32°53'，东经103°40'，海拔3190 m）的两种地形（平地和阳坡）条件下狼毒对土壤碳氮循环影响及可能的原因。狼毒和其它几个主要物种（圆穗蓼（Polygonummacrophyllum D. Don var. Macrophyllum），草地早熟禾（Poa pretensis L.），四川嵩草（Kobresia setchwanensis Hand.-Maizz.），鹅绒委陵菜（Potentilla anserina L.var. anserine）和鳞叶龙胆（Gentiana squarrosa Ledeb.）的地上凋落物产量以及地上凋落物和根的化学组成被测量。在有-无狼毒斑块下，各种土壤的库（比如，铵态氮、硝态氮、无机磷和微生物生物量）和周转率（包括净矿化、净硝化、总硝化、反硝化和微生物呼吸速率）被测量和比较。（2）在非生长季节，尤其是春季冻融交替期，选取了两个研究地点——尕米寺和卡卡沟（北纬32°59'，东经103°41'，海拔3400 m），分别测定有狼毒和无狼毒斑块下土壤微生物生物量碳和氮、可溶性有机碳和氮以及铵态氮和硝态氮的动态变化。同时，分别在上述两个地点有-无狼毒的样地上，研究6 个主要物种（狼毒、圆穗蓼、草地早熟禾、四川嵩草、鹅绒委陵菜和鳞叶龙胆）从秋季开始、为期1 年的凋落物分解过程。 针对狼毒花粉化感对同花期其它物种可能的花粉化感作用开展以下工作：在实验室中，用一系列浓度的狼毒花粉水浸提液对与它同花期的其它物种以及自身花粉进行测试，测定花粉萌发率；在野外自然条件下的其它物种的柱头上施用上述浓度的狼毒花粉水浸提液，观测种子结实率，同时，观察狼毒花粉的种间花粉散布数量。 生长季节的研究结果表明，狼毒地上凋落物氮含量比其它几个主要物种更高，而木质素-总氮比更低。狼毒显著地增加其斑块下表层土壤中有机质的含量，而有-无狼毒的亚表层土壤在有机碳和总磷方面没有显著差异。狼毒表土中硝态氮含量在平地和阳坡比无狼毒土壤分别高113%和90%。狼毒表土中微生物生物量碳和氮量显著高于无狼毒表土。无论是平地还是阳坡，狼毒土壤的总硝化和微生物呼吸速率显著高于无狼毒土壤；而它们的反硝化速率只在平地有显著的差异。狼毒与其它物种间地上凋落物的产量和质量的差异可能是导致有-无狼毒土壤碳氮循环差异的原因。我们假设，狼毒可能通过增加贫氮生态系统土壤中的有效氮含量提高其入侵能力。 非生长季的研究结果表明，在青藏高原东缘的高寒草甸上，土壤微生物生3物量在11 月的秋-冬过渡期达到第一个峰值；在春季的冻融交替期，微生物生物量达到第二个峰值后又迅速降低。无机氮以及可溶性有机碳氮与微生物生物量有相似的变化过程。微生物碳氮比呈现显著的季节性变化。隆冬季节的微生物生物量碳氮比显著高于生长旺季的微生物碳氮比。这种变化可能暗示冬、夏季微生物的群落组成和对资源的利用有所不同。有-无狼毒斑块下土壤微生物和土壤碳、氮库一般只在秋-冬过渡期有显著差异，有狼毒土壤微生物生物量和土壤碳、氮库显著高于无狼毒土壤；而在之后的冬季和春季没有显著差异。所有6 个物种凋落物在非生长季分解率为24%-50%，均高于生长季的10%到30%。其中在秋-冬过渡期，凋落物开始埋藏的两周时间内，分解最快，达10%-20%。不同物种凋落物全年的分解率和分解过程有显著差异。圆穗蓼在全年的分解都较缓慢（非生长季26%，生长季15%），草地早熟禾和四川嵩草等全年的分解速率比较均匀（非生长季和生长季均为30%，非生长季略高），而狼毒在非生长季分解较快（约50%），而在接下来的生长季分解变得缓慢（约12%）。所有物种的凋落物氮含量在非生长季下降，而在随后的生长季上升。 实验室的花粉萌发试验证明，狼毒花粉对自身花粉萌发没有自毒作用，而其它受试的所有物种（圆穗蓼，秦艽（Gentiana macrophylla Pall. var. fetissowii），湿生扁蕾（Gentianopsis paludosa (Hook. f.) Ma var. paludosa），鳞叶龙胆，椭圆叶花锚（Halenia elliptica D. Don var. elliptica），蓝钟花（Cyananthus hookeri C. B.Cl. var. grandiflorus Marq.），小米草（Euphrasia pectinata Ten.），川西翠雀花（Delphinium tongolense Franch.），高原毛茛（Ranunculus tanguticus (Maxim.)Ovcz. var. tanguticus）和鹅绒委陵菜）的花粉萌发率随着狼毒花粉浸提液浓度的增加呈显著的非线性降低。大约3 个狼毒花粉的浸提液就可以抑制受试的多数物种的50%的花粉萌发。在鳞叶龙胆和小米草柱头上狼毒花粉的数量分别为5.76 个和3.35 个。狼毒花粉散布数量的差异最可能的原因在于是否有共同的传粉昆虫。花的形状（辐射对称VS 左右对称）、植株或花的密度以及花期重叠性可以部分解释这种差异。在野外试验中，我们发现6 个物种（秦艽、湿生扁蕾、鳞叶龙胆、椭圆叶花锚、蓝钟花和小米草）的种子结实率随狼毒花粉浸提液浓度的增加呈显著的非线性降低。鳞叶龙胆和小米草柱头上狼毒花粉的数量（分别是5.76 个和3.35 个）分别达到了抑制它们63%和55%种子结实率的水平。因此，狼毒对鳞叶龙胆和小米草可能存在明显的花粉化感抑制作用。狼毒周围的物种可能通过花期在季节或昼夜上的分异避免受到狼毒花粉化感的影响或者通过无性繁殖来维持种群繁衍，因此狼毒通过花粉化感作用对其周围物种繁殖的影响程度还需要进一步地研究。如果狼毒的花粉化感抑制作用确实存在，那么它可能成为一种自然选择压力，进而影响物种的进化。 Stellera chamaejasme L., a perennial toxic weed, has emerged and quicklydominated and spread in the high-frigid meadow on the eastern Tibetan Plateau ofChina since the 1960s. In the present study, effects of S. chamaejasme on carbon andnitrogen cycles on the high-frigid meadow on the eastern Qinghai-Tibetan Plateau ingrowing and non-growing season, and its pollen allelopathic effects on the sympatricspecies were determined. The present study that focused on carbon and nitrogencycles, especially on microbial biomass and pools of carbon and nitrogen innon-growing season, could profoundly illuminate plant-species effects on carbon andnutrient cycles and its seasonal pattern and help to understand spread mechanism ofS. chamaejasme as an aggressive weed. The present study also contributed to furtherunderstand carbon and nutrient cycles on alpine regions in non-growing season andprovide a basis on weed control of S. chamaejasme and scientific management in thehigh-frigid ecosystem. Effects of S. chamaejasme on carbon and nitrogen cycles on the high-frigidmeadow on the eastern Qinghai-Tibetan Plateau were determined. The study couldbe divided into two parts. (1) In the growing season, we quantified the effects of S.chamaejasme on carbon and nitrogen cycles in two types of topographic habitats, theflat valley and the south-facing slope, where S. chamaejasme was favored to spreadlitter and root were measured to explain the likely effects of S. chamaejasme on soilcarbon and nutrient cycles. The sizes of various soil pools, e.g. nitrite, ammonium,inorganic phosphorus and microbial biomass, and turnover rates including netmineralization, gross nitrification, denitrification and microbial respiration weredetermined. (2) In the non-growing season study, microbial biomass carbon andnitrogen, soluble organic carbon and nitrogen, ammonium and nitrate weredetermined through the non-growing season, especially in the processes offreeze-thaw of spring in two high-frigid sites, i.e. Kaka valley and Gami temple, onthe eastern Qinghai-Tibetan Plateau. Meanwhile, litter decomposition of six commonspecies, including Stellera chamaejasme L., Polygonum macrophyllum D. Don var.Macrophyllum, Poa pretensis L., Kobresia setchwanensis L., Potentilla anserina L.var. anserine and Gentiana squarrosa Ledeb., were also examined under theabove-mentioned experimental design through one whole-year, which began in theautumn in 2006. In the study of pollen allelopathy, several work, including in vitro study oneffects of extract of pollen from S. chamaejasme on sympatric species and pollenfrom itself, field experiments on effects of pollen extract with the same regime ofconcentrations on seed set and field observation on heterospecific pollen transfer ofS. chamaejasme to six of those sympatric species has been done. The results in the growing season showed that aboveground litter of S.chamaejasme had higher tissue nitrogen and lower lignin: nitrogen ratio than thoseco-occurring species. S. chamaejasme significantly increased topsoil organic matter,whereas no significant differences were found for organic C and total P in subsoilbetween under-Stellera and away-Stellera locations. The nitrate in Stellera topsoilwas 113% and 90% higher on the flat valley and on the south-facing slope,respectively. Both microbial biomass C and N were significantly higher in Stelleratopsoil. Gross nitrification and microbial respiration were significantly higher inStellera topsoil both on the flat valley and on the south-facing slope, whereassignificant differences of denitrification were found only on the flat valley. Thedifferences in the quantity and quality of aboveground litter are a likely mechanismresponsible for the changes of soil variables. We assumed that S. chamaejasme couldenhance their spread by increasing nutrient availability in N-deficient ecosystems. The results in the non-growing season showed that microbial biomass achievedthe first summit in late autumn and early winter on the eastern Qinghai-TibetanPlateau. In the stages of freeze-thaw of spring, microbial biomass firstly achieved thesecond summit and subsequently sharply decreased. Inorganic nitrogen, solubleorganic carbon and nitrogen had a similar dynamics with that of microbial biomass.Ratio of microbial biomass carbon and nitrogen had an obviously seasonal pattern.The highest microbial C: N were in the non-growing season, which weresignificantly higher than those in the growing season. The seasonal pattern inmicrobial biomass C: N suggested that large changes in composition of microbialpopulation and in resources those used by microbes between summer and winter.Generally, microbial biomass and pools size of carbon and nitrogen in Stellera soilwere significantly higher than those under adjacent locations in late autumn andearly winter, but there were not significant differences in winter and in spring. Litterof all the focal species (Stellera chamaejasme L., Polygonum macrophyllum D. Donvar. Macrophyllum, Poa pretensis L., Kobresia setchwanensis Hand.-Maizz.,Potentilla anserina L. var. anserine and G. squarrosa Ledeb.) decomposed about24%-50% in the non-growing season, which were higher than those in the growingseason (ranged from 10% to 30%). Litter decomposed 10%-20% within the first twoweeks in late autumn and early winter. Significant differences in the whole-yeardecomposition rate and in the processes of decomposition were found among species.Polygonum macrophyllum decomposed slowly through the whole year (26% and15% in the non-growing season and in the growing season, respectively). Certainspecies, such as P. pretensis L. and K. setchwanensis, decomposed at a similar rate(30% both in the non-growing and in the growing season, slightly higher in the8growing season than those in the growing season), whereas S. chamaejasmedecomposed more rapidly (about 50%) in the non-growing season and subsequentlydecomposition became slow (about 12%) in the growing season. Litter nitrogencontents of all the focal species firstly decreased in the non-growing season and thenincreased in the growing season. In vitro experiments of pollen allelopathy, the results showed that pollen from S.chamaejasme was not autotoxic, whereas pollen germination in all the sympatricspecies (Polygonum macrophyllum D. Don var. Macrophyllum, Gentianamacrophylla Pall. var. fetissowii, Gentianopsis paludosa (Hook. f.) Ma var. paludosa,Gentiana squarrosa Ledeb., Halenia elliptica D. Don var. elliptica, Cyananthushookeri C. B. Cl. var. grandiflorus Marq., Euphrasia pectinata Ten., Delphiniumtongolense Franch., Ranunculus tanguticus (Maxim.) Ovcz. var. tanguticus andPotentilla anserina L. var. anserina) decreased nonlinearly as the increasingconcentrations of extract of pollen from S. chamaejasme. Pollen Extract of threepollens from S. chamaejasme generally inhibited 50% pollen germination of most ofthe focal species. 5.76 and 3.35 pollens from S. chamaejasme were observed in fieldon stigmas of G. squarrosa and E. pectinata, respectively. Differences inheterospecific pollen transfer of S. chamaejasme could be attributed to the primaryreason whether they shared common pollinators. Flower morphology (e.g.zygomorphic or actinomorphic), plant or floral density and concurrence in floweringphonologies could explain, in part, the differences in heterospecific pollen transfer.In field experiments, the results showed that seed set in six sympatric species(Gentiana macrophylla Pall. var. fetissowii, Gentianopsis paludosa (Hook. f.) Mavar. paludosa, Gentiana squarrosa Ledeb., Halenia elliptica D. Don var. elliptica,Cyananthus hookeri C. B. Cl. var. grandiflorus Marq. and Euphrasia pectinata Ten.)decreased nonlinearly as the increasing concentrations of extract of pollen from S.chamaejasme. According to the nonlinear curves, the amounts of pollens from S.chamaejasme on stigmas of G. squarrosa and of E. pectinata (i.e. 5.76 grains and3.35 grains, respectively) could reduce 63% and 55% seed set of G. squarrosa and ofE. pectinata, respectively. Thus, allelopathic effects of S. chamaejasme on G.squarrosa and E. pectinata could be realistic. The sympatric species of S.chamaejasme could avoid pollen allelopathy of S. chamaejasme to sustainthemselves. This highlights the need to study how much pollen allelopathy in S.chamaejasme influences the sympatric species through divergence in seasonal ordiurnal flowering phonologies or through asexual reproduction. If pollen allelopathyin S. chamaejasme was confirmed, it could be as a pressure of natural selection andthus play an important role in species evolution.

半湿润农田杂草及施氮对夏玉米产量及氮素利用的影响

以土垫旱耕人为土为供试土壤,采用大田试验,研究了半湿润农田两种杂草处理方式下(成熟后期清除杂草-A区和苗期开始清除杂草-B区),不同施氮量对夏玉米产量及氮素利用效率的影响。结果表明,当施氮量为0、45、90、1351、80 kg/hm2时,B区玉米子粒产量比A区分别增加了8.7%、12.1%、9.4%、5.0%和12.5%;吸氮量分别增加了1.5、2.9、4.85、.2和4.3 kg/hm2。A区和B区全生育期0—100 cm土层矿质氮(Nmin)累积量变化趋势基本一致,但B区比A区变幅较大。当施氮量为45、90、135和180 kg/hm2时,B区氮肥利用率、氮肥农学利用率、氮肥生理利用率均高于A区。研究还发现,在A区,当施氮量为180 kg/hm2时,杂草干生物量最大,为1518.3 kg/hm2,不施氮时,杂草的生物量最低,为845.7 kg/hm2;杂草的吸氮量随施氮量的增加而增加。可见,清除玉米农田杂草不仅可以提高作物产量和氮肥利用率,而且在减少氮素损失方面具有一定作用。

Productivity and persistence of perennial grass mixtures under competition from annual weeds in the alpine region of the Qinghai-Tibetan Plateau

In the alpine region of the Qinghai-Tibetan Plateau four indigenous perennial grass species Bromus inermis (BI), Elymus sibiricus (ES), Elymus nutans (EN) and Agropyron cristatum (AC) were cultivated as three mixtures with different compositions and seeding rates, BI + EN, BI + ES + AC and BI + ES + EN + AC. From 1998 to 2001 there were three different weeding treatments: never weeded (CK); weeded on three occasions in the first year (1-y) and weeded on three occasions in both the first and second year (2-y) and their effect of grass combination and interactions on sward productivity and persistence was measured. Intense competitive interference by weedy annuals reduced dry matter (DM) yield of the swards. Grass combination significantly affected sward DM yields, leaf area index (LAI) and foliar canopy cover and also species composition DM and LAI, and species plant cover. Interaction between weeding treatments and grass combination was significant for sward DM yield, LAI and canopy cover, but not on species composition for DM, LAI or species plant cover. Grass mixture BI + ES + EN + AC gave the highest sward DM yield and LAI for both weeding and non-weeding treatments. Species ES and EN were competitively superior to the others. Annual weedy forbs must be controlled to obtain productive and stable mixtures of perennial grasses, and germination/emergence is the most important time for removal. Weeding three times (late May, late June and mid-July) in the establishment year is enough to maintain the production and persistence of perennial grass mixtures in the following growing seasons. Extra weeding three times in the second growing year makes only a slight improvement in productivity.