Agricultural and Food Science, Vol 20(2011) S. 191-268

2011

cience www.mtt.fi/afs 3,

No.

food

20,

and

Vol. Agricultural Economics Agricultural Engineering Animal Science Environmental Science Food Science Horticulture Plant and Soil Science

gricultural A Agricultural Society The Scientific of Finland Agrifood Research Finland MTT

AGRICULTURAL AND FOOD SCIENCE V o l . 20, 3 (2011) 191–268 235 245 262 191 206 127 228

cience

AND FOOD S FOOD AND ISSN electronic edition 1795-1895

gricultural Composition of weed flora in spring cereals in Finland – a fourth flora in spring cereals in survey Composition of weed Saastamoinen, M. and Salopelto, J. Salonen, J., Laitinen, P., crops in Finland species and their control in oilseed The main weed Meat bone meal as fertiliser and oat for barley and Jalli, H. Salonen, J., Hyvönen, T. Seed quality effects on seedling emergence, plant stand establishment and grain yield in two-row barley plant stand establishment and grain yield in two-row Seed quality effects on seedling emergence, Chen,L., Kivelä, J., Helenius, J. and Kangas, A. Casas, C. and Casas, C. Martinez, O., Salmerón, J., Guillén, M.D. Characteristics of dry- flavouring salmon later treated with liquid smoke and brine-salted A., Niskanen, M., Isolahti M. and Peltonen-Sainio, P. Rajala, Salovaara, H. Brinck, O. and Salovaara, H. T., Sontag-Strohm, Kanerva, P., detection in processed foods of prolamins for gluten extraction Improved Huuskonen, A. Huuskonen, A. meal supplementation on perfor- to commercial concentrate or rapeseed grain compared Effects of barley dairymance of growing bulls offered grass silage-based diet Contents A Vol. 20, No. 3, 2011 Agricultural AND FOOD Science

Agricultural and Food Science publishes original reports on agriculture and food research. The papers, which are of international interest but feature a northern perspective, cover a wide range of topics in basic and applied research. Submissions are internationally refereed. Review articles and research notes will also be considered. Readers are welcome to send their comments to the journal (Letter to the Editor).

Publishers The Scientific Agricultural Society of Finland MTT Agrifood Research Finland

Editorial board Juha Helenius - Editor Tapani Alatossava Xavier Irz Mervi Seppänen Eila Turtola Pekka Uimari Kaisa Kuoppala - Associate editor

Address Agricultural and Food Science Editorial Office MTT Agrifood Research Finland FI-31600 Jokioinen, Finland http://www.mtt.fi/afs

From volume 19 the publication of Agricultural and Food Science is available as an Open Access journal, free of charge to the user. The issues and articles published since volume 11 have also been transferred to the Open Access mode. Publishing continues in electronic format only, available on the journal's web-pages (www.mtt.fi/afs), or via database links or where the journal is indexed. The printing of special issues will still be considered should the applicant offer to provide the resources.

Single printed issues up to volume 18 may be obtained from Bookstore Tiedekirja, Kirkkokatu 14, FI-00170 Helsinki, Finland, e-mail: [email protected]. The price of a single issue is 17 + VAT within the EU.

The journal is indexed by the following databases: AGRIS, BIOSIS, CAB Abstracts,Cambridge Scientific Abstracts, Current Awareness in Biological Sciences, Current Contents, FSTA, Pascal, Scopus, Science Citation Index, ISI Web of Science and Ingenta.

Huuskonen, A. Effects of different concentrate feedings on performance of bulls Vol. 20 (2011): 191–205.

Effects of barley grain compared to commercial concentrate or rapeseed meal supplementation on performance of growing dairy bulls offered grass silage-based diet

Arto Huuskonen MTT Agrifood Research Finland, Animal Production Research, FI-92400 Ruukki, Finland e-mail: [email protected]

The objectives of the study with dairy bulls offered grass silage-based diet were to determine the effects on animal performance of (1) concentrate type (barley vs. commercial concentrate) and (2) supplementation of rapeseed meal (RSM) in barley-based concentrate, with data being compared from preweaning to slaughter. The experiment comprised a total of 37 Finnish Ayrshire and 23 Holstein-Friesian bulls. Experimental concentrate treatments were 1) rolled barley (B), 2) rolled barley + rapeseed meal (BRSM) and 3) commercial concentrate (CC). During the preweaning (from 0.5 to 2.5 months) there were no differences in intake, gain or feed conversion. During the postweaning (from 2.5 to 6.0 months) the energy intake and gain of the B bulls were 12–13% lower than those of the BRSM bulls (p < 0.05) and 16% lower than those of the CC bulls (p < 0.01). However, there were no treatment differences in the energy intake or gain of the bulls during the finishing period (from 6.0 to 18.0 months of age) or on average during the experiment. Furthermore, carcass traits of the bulls did not differ between treatments. It is concluded that production traits were unaffected by concentrate type or RSM supplementation when data is compared from preweaning to slaughter. Key words: Beef production, concentrate supplementation, dairy bulls, supplementary protein

Huuskonen, A. Effects of different concentrate feedings on performance of bulls

Introduction bulls (Manninen et al. 2010) were unaffected by concentrate energy source. Manninen et al. (2010) concluded that production and carcass Unlike in many other countries, beef production in traits were unaffected by concentrate type, i.e. Finland is based mainly on raising dairy-breed bulls concentrates of differing energy sources, since born on dairy farms. The decrease in the number of the energy and protein contents were similar in dairy cows has diminished the supply of calves for both concentrates. beef production from dairy herds. Consequently, In Finland, rapeseed meal (RSM) is the most slaughterhouse pricing favours heavy carcasses and important protein feed used in concentrates for the average carcass weights of animals have clearly cattle. Huuskonen et al. (2007, 2008) reported increased during recent years. For example, the that RSM did not affect animal performance of average carcass weight of bulls increased from finishing dairy bulls (from 6.0 to 18.0 months of 275 kg (1996) to 335 kg (2008) in twelve years age), and concluded that there is no reason to use (Karhula and Kässi 2010). protein supplement for finishing dairy bulls when In Finland, the feeding of dairy bulls is main- they are fed with good quality grass silage and ly based on grass silage and grain, typically on barley-based concentrate. However, inclusion of barley and/or oats. Nowadays, some beef pro- RSM in the diet was found to have a positive ef- ducers supplement grass silage-based rations fect on the performance of young bulls and bull with commercial concentrates of lower starch calves in some feeding experiments (Aronen et al. concentration and higher protein and fibre con- 1992, Aronen and Vanhatalo 1992). The growth centration rather than straight grain. Especially and feed efficiency over the whole growth period, young calves are typically fed using commercial including preweaning, postweaning and finishing starter concentrates. However, the price of these periods, are critical also from the economic view- concentrates is high compared with that of grain point. The amount of commercial concentrate and or forage. Based on literature reports the effect RSM strongly affects the production costs, since of energy supplement type on the intake and per- the prices of commercial concentrate and RSM formance of growing cattle is complicated and are high compared to grain and thus it is impor- partly unclear (McGee 2005). Mayne et al. (1995) tant to asses how long the possible growth advan- concluded that starch or fibre supplements had no tage will be maintained after weaning when dairy significant difference on the mean substitution bulls are raised to carcass weights over 300 kg. rate in growing cattle when considered across To my knowledge, there is a paucity of pub- a range of silage compositions, but there were lished information on the relative performance of interactions between supplement type and silage growing and finishing dairy bulls offered grass type. Steen (1993) reported that silage intake was silage-based diets supplemented by just barley, higher for fibre than starch-based concentrate barley plus RSM or commercial concentrates for growing cattle. However, the silage intake with performance data being compared from of growing and finishing cattle was shown not preweaning to slaughter. Therefore the objec- to be differentially affected by starch, fibre or tives of the present experiment with growing sugar-based concentrates (Moloney et al. 1993) dairy bulls raised to a carcass weight of 340 kg or by fibre or starch-based concentrates (O’Kiely were to determine the effects on diet digestibility, and Moloney, 1994). In more recent studies it has feed intake, gain and carcass characteristics of been observed that the intake, performance and (1) concentrate type (barley grain vs. commercial carcass characteristics of Continental crossbred concentrate) and (2) supplementation of RSM in steers (McGee et al. 2006) or finishing Hereford barley-based concentrate.

192 193 AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls Vol. 20 (2011): 191–205.

Materials and methods a straw-bedded lying area and the front was a feeding area with a solid concrete floor. A feeding trough was situated on the front side of the pen, Animals and housing and there was 0.8 m of feeding space/bull at the feeding trough. There were heated water bowls The feeding experiment was conducted in the between the pens offering water for bulls. experimental barn of the North Ostrobothnia Re- search Station of MTT Agrifood Research Finland (Ruukki, 64°44'N, 25°15'E) and included two Feeding and experimental design trials. The first trial started in November 2007, ended in May 2009 and carried out 540 days in total. The second trial started in January 2009, The three concentrate feeding treatments used ended in July 2010 and carried out 546 days in in the experiment were: 1) barley grain (B), 2) total. The experimental procedures were evalu- barley grain + rapeseed meal (BRSM) and 3) ated and approved by the Animal Care and Use commercial concentrate (CC). The calves were Committee of MTT Agrifood Research Finland. randomly (balanced for breed) allotted to pens (5 The first trial comprised 18 Finnish Ayrshire calves/pen) which were then randomly allotted to bulls and 12 Holstein-Friesian bulls. The second three experimental treatments. trial comprised 19 Finnish Ayrshire bulls and 11 During the preweaning period the calves Holstein-Friesian bulls. received a milk replacer (MR) [at a dilution of All animals, initial live weight (LW) 53±2.5 11.9% dry matter (DM)] supplied by Valio Ltd. kg and age 15±6.3 days, on average, were pur- (Helsinki, Finland). The MR included (g kg-1 chased from local dairy farms. During the DM) skim milk powder (558), whey powder preweaning (from 0.5 to 2.5 months of age) and (245), lard (152), wheat starch (23), rapeseed

postweaning (from 2.5 to 6.0 months of age) peri- oil (9), lecithin (4), CaCl2 (4), NaCl (3) and vi- ods the animals were housed in an insulated barn tamin-mineral premix (2). In both trials and all on peat bedding in six pens (3.0 × 3.5 m, 5 calves treatments the MR was served by a computer- in each) providing 2.1 m2/calf. The air tempera- controlled feeder (two pens/feeder; Stand Alone ture in the insulated barn varied between 11 and 2 Plus, Förster, Engen, Germany; programme: 20 °C in winter (October–April) and between 15 Kalbmanager 4.2). The feeding temperature of and 23 °C in summer (May–September). the MR was 37 ºC. The calves were allocated to For determination of diet digestibility all ani- treatments at 15 days of age, and from days 15 mals were placed in an insulated barn in adjacent to 57 the highest possible MR allowance of the tie-stalls from 6.0 to 7.0 months of age. The width calves was 8.5 l. All calves were weaned gradu- of the stalls was 70–90 cm and the bulls were tied ally from days 57 to 70 with the MR allowance with a collar around the neck attached by a 50 being cut by reducing the number of MR por- cm chain to a horizontal bar 40–55 cm above the tions per day. During the preweaning period the floor. The floor surface was solid concrete under animals received water, concentrate, grass silage the forelegs and metal grid under the hind legs. and hay ad libitum (proportionate refusals as No bedding was used on the floor. Each bull had 5%). Concentrate offered for three concentrate its own water bowl. treatments were 1) rolled barley, 2) mixture of From 7.0 to 18.0 months of age the bulls were rolled barley (800 g kg-1 DM) and RSM (200 g placed in an uninsulated barn in adjacent pens (4 kg-1 DM) and 3) commercial concentrate (Primo × 8 m, 6.4 m2/bull, 5 bulls in each pen). The barn I) produced by Suomen Rehu Ltd. (Hyvinkää, was covered with a roof and had solid wooden Finland). Forage and concentrates were offered walls on all sides except for the front side that separately from a box feeder during the pre- and was left open. The rear half of the pen area was postweaning periods. 192 193 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls

The amounts of the included ingredients of Rehu Ltd., Hyvinkää, Finland) was given at 50 g the commercial concentrates varied slightly be- per animal weekly. The commercial concentrate tween trials. The commercial concentrate (Primo included sufficient vitamins and minerals and I) used during the pre- and postweaning periods therefore separate mineral or vitamin mixtures comprised (g kg-1 DM, shown as mean values were not used in the CC treatment. over the trials) rapeseed cake (150), barley The grass silages in both trials were primary (150), wheat bran (127), oats (100), wheat (80), growth from a timothy (Phleum pratense) and molassed sugar-beet pulp (80), naked oats (60), meadow fescue (Festuca pratensis) sward and rapeseed meal (50), soybean meal (50), molasses ensiled in bunker silos with a formic acid-based (50), wheat feed meal (30), barley malt feed (30), additive (AIV-2 Plus: 760 g formic acid kg-1, 55 -1 CaCO3 (16), brewery yeast (Progut®, patent: g ammonium formate kg , supplied by Kemira FI109759) (10), vegetable oil mix (5), vitamin, Ltd., Oulu, Finland) applied at a rate of 5 litres mineral and trace element premix (4), salt (4), t-1 of fresh grass.

Na2CO3 (2) and MgO (2). During the postweaning period the animals received grass silage, hay and water ad libitum, but the amount of concen- Procedures and sample analyses trate was restricted to 3 kg (air dry)/animal/d. Concentrate feeding treatments and feeds were the same as during the preweaning period. Silage sub-samples for chemical analyses were During the finishing period (including the de- taken twice a week, pooled over periods of four termination of diet digestibility) the bulls were weeks and stored at –20ºC. Thawed samples were fed total mixed ration (TMR) ad libitum. The tar- analysed for DM, ash, crude protein (CP), crude fat get concentrate proportion for all treatments was (CF), neutral detergent fibre (NDF), indigestible 500 g kg-1 DM. The TMR for treatment B includ- NDF (INDF), starch, silage fermentation quality ed grass silage (500 g kg-1 DM) and barley grain (pH, water-soluble carbohydrates (WSC), lactic (500), for treatment BRSM grass silage (500), and formic acids, volatile fatty acids, soluble and barley grain (450) and RSM (50), and for treat- ammonia N content of N) and digestible organic ment CC grass silage (500) and commercial con- matter (DOM) in DM (D value). Concentrate, centrate (500). The commercial concentrate used MR and hay sub-samples were collected weekly, during the finishing period was Primo II (Suomen pooled over periods of eight weeks and analysed Rehu Ltd) which comprised (g kg-1 DM, shown for DM, ash, CP, CF, NDF, INDF and starch (hay as mean values over the trials) barley (264), oats also for D value). The analyses of DM, ash, CP, CF, (220), wheat bran (127), mash feed meal (100), NDF, INDF and starch were made as described by rapeseed meal (89), barley malt feed (80), molas- Huuskonen (2009) and Huuskonen & Joki-Tokola ses (55), CaCo3 (20), oat husk meal (20), brewery (2010). The silage was analysed for fermentation yeast (Progut®, patent: FI109759) (8), salt (7), quality by electrometric titration as described by vegetable oil mix (6), vitamin, mineral and trace Moisio and Heikonen (1989) and for D value by element premix (2), and MgO (2). The animals the method described by Nousiainen et al. (2003). were fed three times per day (at 0800, 1200 and Feed and faecal samples were collected twice a 1800 hours). Refused feed was collected and day (at 0700 and 1500 hours) during the collection measured at 0700 hours daily. The daily ration for period (5 d) and stored frozen prior to analyses. B and BRSM bulls included also 150 g of a min- The samples were analyzed for DM, ash, CP and eral mixture (KasvuApeKivennäinen delivered NDF as described above. The diet digestibility was by A-Rehu Ltd., Seinäjoki, Finland). A vitamin determined using acid-insoluble ash (AIA) as an mixture (Xylitol ADE-Vita delivered by Suomen internal marker (Van Keulen and Young 1977).

194 195 AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls Vol. 20 (2011): 191–205.

Calculations Dressing proportions were calculated from the ratio of cold carcass weight to final LW. The carcasses The metabolizable energy (ME) contents of the were classified for conformation and fatness using feeds were calculated according to the Finnish feed the EUROP quality classification (Comission of the tables (MTT 2006). The ME value of the silage was European Communities, 1982). For conformation, calculated as 0.016 × D value (MTT 2006). The development of carcass profiles, in particular the ME value of the hay was calculated as 0.0169 × D essential parts (round, back, shoulder), was taken value – 1.05 (MTT 2006). The ME values of the into consideration according to the EUROP clas- concentrates were calculated based on concentra- sification (E: excellent, U: very good, R: good, O: tions of digestible crude fibre, CP, crude fat and fair, P: poor), and for fat cover degree the amount of nitrogen-free extract described by MAFF (1984). fat on the outside of the carcass and in the thoracic The digestibility coefficients of the concentrates cavity was taken into account using a classification were taken from the Finnish feed tables (MTT range from 1 to 5 (1: low, 2: slight, 3: average, 4: 2006). The supply of amino acids absorbed from high, 5: very high). Each level of the conformation the small intestine (AAT) and the protein balance scale was subdivided into three sub-classes (O+, O, in the rumen (PBV) were calculated according to O-) to produce a transformed scale ranging from 1 the Finnish feed tables (MTT 2006). to 15, with 15 being the best conformation. The animals were weighed on two consecutive days at the beginning of the experiment and thereafter every 14 days during the preweaning period. During the postweaning and finishing periods the animals were weighed approximately Statistical methods every 28 days. Before slaughter they were weighed on two consecutive days. The target for average The results were calculated across the two trials carcass weight in the experiment was 340 kg. The and are shown as least squares means. Normality LWG was calculated as the difference between the of residuals was checked using graphical methods: means of initial and final live weights divided by box-plot and scatter plot of residuals and fitted the number of growing days. The estimated rate values. The pen (a group of five animals) was used of carcass gain was calculated as the difference as an experimental unit for testing feed intake and between the final carcass weight and the carcass feed conversion data. There were 4 pens/treatment weight in the beginning of the experiment divided (20 animal/treatment). The average group feed dry by the number of growing days. Carcass weight in matter intake (DMI) and feed conversion data were the beginning of the experiment was assumed to be subjected to analysis of variance using the SAS 0.40 × initial LW as the same value is used by Atria general linear models procedure. The statistical Ltd. (a Finnish slaughterhouse) in daily extension model (1) used was work (Herva et al. 2009). The LWG and feed dry

matter intakes of the bulls are presented separately yjkl = m + βk + αj + (β × α)jk + ejkl (1) for preweaning, postweaning and finishing periods. where μ is the overall mean, ejkl is the random error term and yjkl is the mean of five animals Carcass measurements penned together (4 pens/treatment; l=1,…,4). α and β are the effects of treatment and trial. The gain and carcass characteristics variables After slaughter in a commercial meat plant the car- were measured individually and were subjected to casses were weighed hot. The cold carcass weight analysis of variance using the SAS MIXED model was estimated as 0.98 of the hot carcass weight. procedure. The following statistical model (2) was

194 195 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls used to analyse the gain and carcass characteristics chemical compositions and feeding values are also data given separately for the two silages in Table 1.

Yijkl = m + βj + α i + (β × α)ij + cijk+ eijkl (2) However, the compositions of the silages differed only slightly from each other. The silages used were

where m is the overall mean and eijkl is the ran- of good nutritional quality as indicated by the D dom error term. α and β are the effects of treat- value as well as the AAT and CP contents (Table 1). ment and trial. cijk is the effect of group within The fermentation characteristics of the silages were treatment-by-trial combination and it was used as also good as indicated by the pH value and the low an error term when differences between treatments concentration of ammonia N and total acids. The were tested. silages used were restricted fermented with high For diet apparent digestibility coefficients residual WSC concentration and low lactic acid animal was used as an experimental unit. During concentration. Because the chemical compositions the digestibility determinations the animals were and feeding values of the hay and concentrates placed in the insulated barn in adjacent tie-stalls, were very uniform throughout the experiment, only and the digestibility data were subjected to analysis mean values over the trials are given for hay, barley, of variance using the SAS MIXED model proce- commercial concentrate, RSM and MR in Table dure. The statistical model (3) used was 1. The calculated ME value of the barley was 6% higher than that of the CC used during the pre- and yjkl = m + βk + αj + (β × α)jk + ejkl (3) postweaning periods and 10% higher than that of the CC used during the finishing period. However, where m is the overall mean and ejkl is the ran- the commercial concentrate contained 17 (finishing) dom error term. βk and αj are the fixed effects of and 49% (pre- and postweaning) more CP than the treatment and trial. barley grain. Furthermore, CC contained clearly Differences between the treatments were tested more crude fat and NDF and less starch than the by making two orthogonal contrasts: B vs. BRSM barley grain (Table 1). and B vs. CC. The first contrast described the ef- The average chemical compositions of the total fects of RSM supplementation and the second con- mixed rations used during the finishing period are trast the effects of concentrate type. presented in Table 2. Because of the higher energy content of the barley grain the CC ration contained 5% less ME than the B and BRSM rations. The B ration contained 7% less CP than the BRSM and CC rations. The CC ration contained 14% more Results NDF and 41% less starch than the B and BRSM rations. Furthermore, the B ration contained 29 and Feeds 39% less crude fat than the BRSM and CC rations, respectively. Because the grass silages used in the feeding experiment came from two different harvests, the

196 197 AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls Vol. 20 (2011): 191–205. - 3±0.8 10 91±2.1 47±1.0 102±0.1 145±4.2 306±4.3 926±2.5 870±2.2 360±5.5 11.9±0.1 period) (finishing concentrate Commercial - 6 - periods) 46±1.9 53±0.9 29±0.5 111±0.1 185±3.6 242±3.4 927±2.1 875±2.0 327±4.9 centrate (pre- 12.4±0.1 and postweaning Commercial con - 14 22±0.2 40±0.7 111±1.1 151±0.1 351±2.5 318±1.4 133±2.1 914±2.8 881±1.7 11.7±0.1 Rapeseed meal 14 - Barley 43±2.3 22±1.3 -38±0.9 104±0.1 124±6.4 204±11.1 971±4.6 874±3.8 525±5.6 13.1±0.2 4 - - 2±0.1 16±0.5 48±0.2 180±0.2 206±2.1 919±1.9 965±2.2 170±5.6 19.9±0.1 Milk replacer c 12 Hay 74±0.2 62±4.9 21±0.6 8.8±0.2 -63±0.3 ND 685±10.5 947±3.2 888±5.7 550±10.6 ). ND a 14 7±0.2 18±9.5 86±3.2 46±60.3 53±18.7 56±16.1 14±11.7 51±5.4 43±0.9 4.0±0.3 161±14.2 550±18.8 504±124.8 936±6.3 260±90.8 695±30.3 11.1±0.5 Silage trial 2 14 6±3.7 85±3.8 87±50.0 35±10.3 49±23.7 27±9.6 12±0.1 60±6.1 39±0.8 4.3±0.3 173±15.6 536±14.6 459±169.2 922±7.2 324±95.3 675±30.8 10.8±0.5 Silage trial 1 DM -1 DM -1 DM -1 DM -1 DM -1 DM feed -1 -1 DM DM -1 -1 DM DM -1 -1 -1 DM DM DM -1 -1 -1 , g kg d , g kg , g kg f e N 4 b Number of feed samples. Other Silage: feeds: values only of mean two values trials over are the given trials separately. are given because the chemical compositions and feeding values were Amino acids absorbed from small intestine. Digestible organic matter in dry matter. Digestible organic Standard deviation. Not determined. Protein balance in the rumen. Lactic + formic acid, g kg PBV PBV In total N, g kg Neutral detergent fibre (NDF), g kg Neutral detergent Indigestible NDF, g kg Indigestible NDF, Crude fat, g kg N Water-soluble carbohydrates, g kg Water-soluble Soluble N Crude protein, g kg Organic matter, g kg matter, Organic D value MJ kg Metabolizable energy, Starch, g kg Table 1. Chemical composition and nutritional values of feeds (mean±S.D. Table NH a b c d e f very uniform throughout the experiment. Dry matter (DM), g kg AAT Fermentation quality of silage pH Volatile fatty acids, g kg Volatile 196 197 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls

Table 2. Chemical compositions and nutritional values of total mixed rations used in finishing period (from six months of age to slaughter). Silage + barley Silage + barley + Silage + commercial rapeseed meal concentrate Dry matter (DM), g kg-1 473 484 472 Organic matter, g kg-1 DM 947 945 924 Crude protein, g kg-1 DM 149 159 159 Neutral detergent fibre (NDF), g kg-1 DM 370 367 421 Indigestible NDF, g kg-1 DM 52 56 76 Crude fat, g kg-1 DM 31 40 43 Starch, g kg-1 DM 269 256 186 Metabolizable energy, MJ kg-1 DM 11.9 11.9 11.3 AAT a, g kg-1 DM 95 97 94 PBV b, g kg-1 DM -6 0 15 a Amino acids absorbed from small intestine. b Protein balance in the rumen.

Diet digestibility and feed intake protein intake (Table 4). However, roughage intake tended to be higher with BRSM than that with B Diet apparent DM digestibility (DMD) and organic diet (p = 0.08) and CP intake tended to be higher matter digestibility (OMD) were both 10% higher in both BRSM (p = 0.07) and CC (p = 0.08) diets with B diet than with CC diet (p < 0.001), but there than that in B diet. In addition, NDF intake tended were no differences between B and BRSM diets in to be 25% higher in BRSM diet than in B diet (p DMD or OMD (Table 3). Diet CP digestibility was = 0.07). Unlike the preweaning period, there were 11% lower with B than that with BRSM diet (p < many treatment differences in intake parameters 0.001), but there were no differences between B during the postweaning period (Table 4). The DM and CC diets in CP digestibility. Furthermore, diet (p < 0.05), ME (p < 0.05) CP (p < 0.001), AAT NDF digestibility (NDFD) was 12% higher with B (p < 0.01) and NDF (p < 0.01) intakes were re- than with CC diet (p < 0.001). In NDFD there was spectively 14, 12, 34, 19 and 21% higher for the no difference between B and BRSM treatments. BRSM animals than for the B animals. Further, During the preweaning period there were no the DM (p < 0.01), ME (p < 0.01) CP (p < 0.001), significant treatment differences in feed, energy or AAT (p < 0.01) and NDF (p < 0.001) intakes were

Table 3. Effects of concentrate type on apparent diet digestibility of growing dairy bulls. Concentrate type a Statistical significance (p value) c B BRSM CC SEM b F T F × T C1 C2 Digestibility coefficients dry matter 0.775 0.779 0.725 0.0038 <0.0001 <0.0001 0.002 0.224 <0.0001 organic matter 0.786 0.794 0.740 0.0038 <0.0001 <0.0001 0.0006 0.335 <0.0001 crude protein 0.722 0.799 0.708 0.0043 <0.0001 <0.0001 0.007 0.043 0.653 neutral detergent fibre 0.679 0.686 0.608 0.0069 <0.0001 <0.0001 0.329 0.269 0.013 a B = barley grain as concentrate supplement; BRSM = barley grain + rapeseed meal as concentrate supplement; CC = commercial concentrate mixture as concentrate supplement. b Standard error of mean. c F = feeding treatment, T = trial, F × T = feeding treatment and trial interaction. Differences between feedings were tested by making two orthogonal contrasts: C1 = B vs. BRSM and C2 = B vs. CC.

198 199 AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls Vol. 20 (2011): 191–205. C2 0.314 0.084 0.291 0.195 0.002 0.003 0.111 0.340 0.007 0.081 0.682 0.030 0.181 0.007 0.586 0.006 0.361 <0.001 <0.001 <0.001 <0.001 c C1 0.951 0.073 0.194 0.070 0.003 0.008 0.059 0.408 0.692 0.241 0.023 0.957 0.953 0.071 0.177 0.015 0.084 0.015 0.192 <0.001 <0.001 (p value) F × T 0.475 0.706 0.701 0.414 0.533 0.467 0.490 0.556 0.501 0.524 0.747 0.406 0.547 0.512 0.679 0.998 0.440 0.166 0.439 0.701 <0.0001 T 0.410 0.140 0.155 0.011 0.139 0.394 0.239 0.018 0.907 0.0004 0.507 0.076 0.425 0.778 0.126 0.062 0.994 0.041 0.001 0.347 0.112 Statistical significance F 0.472 0.125 0.367 0.162 0.0003 0.004 0.003 0.009 0.252 0.0003 0.440 0.016 0.131 0.869 0.066 0.288 0.014 0.083 0.013 0.385 <0.0001 b 2.334 7.9 3.9 0.823 1.413 0.200 0.013 0.047 0.043 0.089 0.014 0.118 0.063 SEM 11.3 11.2 14.7 20.1 51.0 31.1 18.7 80.0 CC 9.78 0.83 2.54 0.38 2.14 0.19 4.68 1.40 a 19.85 57.67 113.61 292 235 214 848 474 1753 1487 912 221 4242 9.23 0.82 2.50 0.38 2.06 0.24 4.56 1.44 BRSM 20.15 55.60 109.78 294 238 230 805 468 1591 1377 886 -94 3376 Concentrate type B 9.20 0.82 2.36 0.29 1.63 0.20 3.99 1.31 18.64 49.56 109.99 259 222 184 601 393 1311 1274 862 -160 3329 -1 -1 -1 -1 -1 -1 -1

-1 -1 -1 -1 -1 -1 -1 -1 -1 intake, g d , g d e d Milk replacer Concentrate Preweaning period Dry matter (DM) intake (DMI), kg DM d Crude protein intake, g d NDF intake, g d AAT fibre (NDF) intake, g d Neutral detergent Postweaning period DMI, kg DM d NDF intake, g d intake, g d AAT PBV AAT intake, g d AAT Finishing period DMI, kg DM d Total Crude protein intake, g d Crude protein intake, g d ME intake, MJ d Concentrate Roughage Roughage DMI Total Total DMI Total ME intake, MJ d Metabolizable energy (ME) intake, MJ d Metabolizable energy Standard error of mean. Amino acids absorbed from small intestine. B = barley grain as concentrate supplement; BRSM + rapeseed meal Protein balance in the rumen. F = feeding treatment, T = trial, F × T = feeding treatment and trial interaction. Differences between feedings were tested by making two orthogonal contrasts: C1 = B vs. BRSM and C2 = C1 = B vs. BRSM and contrasts: two orthogonal by making tested were feedings between Differences interaction. trial and treatment = feeding T F × = trial, T treatment, F = feeding CC = commercial concentrate mixture as supplement. B vs. CC. Table 4. Effects of concentrate type on daily feed intakes of growing dairy bulls during preweaning (from 0.5 to 2.5 months), postweaning 6.0 months) type on daily of concentrate 4. Effects Table and finishing (from 6.0 to 18.0 months) periods on the average during experiment. a b c d e 198 199 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls respectively 17, 16, 41, 21 and 34% higher for the during the whole experiment (g kg-1 carcass gain) CC animals than for the B animals. was better for the B bulls than for the BRSM (p During the finishing period the DMI tended to < 0.05) and CC (p < 0.01) bulls. However, there be 6% higher for the CC bulls than for the B bulls were no treatment differences in DM (Kg DM kg-1 (p = 0.08), but there was no difference in DMI carcass gain) or ME (MJ kg-1 carcass gain) conver- between B and BRSM bulls (Table 4). However, sions during the experiment (Table 6). there were no treatment differences in the ME and AAT intakes between treatments. Instead, the CP (p < 0.01), PBV (p < 0.001) and NDF (p < 0.001) intakes were clearly higher for the CC animals than for the B animals during the finishing period. For Discussion the BRSM bulls the PBV intake was (p < 0.001) and the CP intake tended to be (p = 0.06) higher Diet apparent DMD, OMD and NDFD were higher than for the B bulls. with the B diet than with the CC diet which was possibly due to differences in the sources of both carbohydrates and protein between these two diets. Growth rate, feed conversion and slaugh- Besides grain, the commercial pelleted concentrate also included various by-product fractions, e.g. ter parameters wheat bran and oat husk meal. Therefore the CC included more cell wall fractions than the barley During the preweaning period there were no treat- grain and the NDFD of these by-product fractions ment differences in LWG or feed conversion pa- is generally lower than the NDFD of barley grain rameters (MJ or CP conversion) (Table 5). Instead, (MTT 2006). Also Huuskonen et al. (2009) found during the postweaning period the LWG of the B that the commercial concentrate with more cell wall bulls was 13% lower than the LWG of the BRSM fractions decreased the OMD and NDFD of the diet bulls (p < 0.05) and 16% lower than that of the CC compared to rolled barley grain in grass silage-based bulls (p < 0.01). The improved gain of the BRSM diets for growing dairy heifers. Similarly to what was and CC bulls during the postweaning period also reported by Huuskonen et al. (2008) and Huuskonen emerges from the live weights of the animals (Table (2009), RSM supplementation had no effect on diet 5). The energy conversion rate (MJ kg-1 LWG) did apparent OMD or NDFD when barley was partly not differ between treatments during the postweaning replaced by RSM. In accordance with earlier studies period, but the CP conversion rate (g kg-1 LWG) was (Aronen et al. 1992, Huuskonen et al. 2007, 2008, better with B bulls than with BRSM and CC bulls Huuskonen 2009), the apparent CP digestibility (p < 0.01). There were no treatment differences in increased with protein supplementation. Some of LWG of the bulls during the finishing period or on the increased apparent digestibility of the CP in average during the experiment, but CP conversion the RSM-supplemented diets may have reflected rate was better with B bulls than with BRSM and the better digestibility of RSM protein compared CC bulls. Energy conversion rate did not differ to barley grain protein (MTT 2006). Most of this significantly between treatments during the finish- increase was, probably, only apparent, related to the ing period or on average during the experiment. decreased proportion of faecal metabolic nitrogen The average (all treatments) carcass weight of recovered in faeces when the CP content increased. the animals was 345 kg and very close to the pre- This hypothesis is supported by Minson (1982). planned. There were no treatment differences in During the preweaning period there were no carcass gain or carcass weight of the bulls (Table notable differences in intake parameters between 6). Furthermore, the dressing proportion, carcass the treatments. This is a logical result because the conformation or carcass fat score of the bulls did MR allowance of the calves was 8.5 l d-1 and MR not differ between treatments. The CP conversion was the most important energy and protein source 200 201 AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls Vol. 20 (2011): 191–205. C2 0.119 0.227 0.039 0.476 0.599 0.003 0.003 0.205 0.007 0.186 0.003 0.982 0.622 0.019 0.498 0.922 <0.001 c C1 0.445 0.693 0.170 0.005 0.635 0.400 0.003 0.020 0.371 0.014 0.095 0.006 0.803 0.352 0.027 0.530 0.982 (p value) F × T 0.311 0.445 0.374 0.417 0.540 0.715 0.607 0.626 0.632 0.594 0.866 0.634 0.833 0.790 0.190 0.969 0.957 T 0.152 0.358 0.043 0.004 0.737 0.597 0.411 0.012 0.878 0.123 0.562 0.016 0.035 0.812 0.842 0.411 0.225 Statistical significance F 0.267 0.438 0.098 0.002 0.752 0.677 0.004 0.008 0.406 0.014 0.200 0.006 0.952 0.624 0.034 0.736 0.994 b 9.2 6.1 1.323 3.8 2.341 2.7 1.064 1.2 1.430 SEM 17.7 17.2 20.66 29.9 32.4 15.7 45.1 18.5 CC 81.81 92 41.99 55 30.24 673 451 242 101.21 617 686 445 1138 1126 1089 1324 1384 a 81.33 99.71 94 41.57 55 30.40 660 441 241 602 717 443 BRSM 1115 1116 1044 1252 1355 Concentrate type B 80.39 96.51 86 41.96 50 30.45 650 428 940 216 509 653 423 1119 1105 1142 1198 -1 LWG -1 LWG -1 Live weight gain (LWG), g d Live weight gain (LWG), Feed conversion MJ kg final, at the age of 18.0 months Crude protein g kg average during the experiment at the age of 12.0 months average during the experiment average during the experiment finishing period at the end of postweaning finishing period finishing period postweaning period at the end of preweaning postweaning period postweaning period preweaning period Live weight, kg initial, at age of 0.5 months preweaning period preweaning period Standard error of mean. B = barley grain as concentrate supplement; BRSM + rapeseed meal F = feeding treatment, T = trial, F × T = feeding treatment and trial interaction. Differences between feedings were tested by making two orthogonal contrasts: C1 = B vs. BRSM and C2 = C1 = B vs. BRSM and contrasts: two orthogonal by making tested were feedings between Differences interaction. trial and treatment = feeding T F × = trial, T treatment, F = feeding CC = commercial concentrate mixture as supplement. B vs. CC. Table 5. Effects of concentrate type on live weight, daily weight gain and feed conversion of growing dairy bulls during preweaning (from 0.5 to 2.5 months), of concentrate 5. Effects Table postweaning (from 2.5 to 6.0 months) and finishing 18.0 periods on the average during experiment. a b c 200 201 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls C2 0.237 0.488 0.792 0.565 0.291 0.005 0.791 0.168 c C1 0.604 0.656 0.433 0.420 0.794 0.023 0.860 0.471 (p value) F × T 0.737 0.863 0.837 0.210 0.786 0.614 0.748 0.648 T 0.687 0.020 0.699 0.096 0.654 0.064 0.217 0.992 Statistical significance F 0.466 0.766 0.544 0.690 0.519 0.013 0.961 0.357 b 3.8 0.267 0.150 0.263 3.372 6.8 12.9 43.0 SEM 4.85 2.52 CC 13.02 609 524 153.00 351 a 2035 4.43 2.47 12.69 BRSM 595 523 152.55 345 1956 Concentrate type B 4.75 2.65 12.59 585 520 151.67 338 1772 d e -1 carcass gain -1 -1 carcass gain -1 carcass gain -1 Slaughter data carcass gain, g d carcass weight, kg conformation score carcass EUROP fat score carcass EUROP Feed conversion during the experiment Kg DM kg MJ kg Crude protein g kg dressing proportion, g kg Standard error of mean. Conformation: (1 = poorest, 15 excellent). B = barley grain as concentrate supplement; BRSM = barley grain + rapeseed meal as concentrate supplement; CC = commercial concentrate mixture as concen - mixture concentrate CC = commercial supplement; as concentrate meal + rapeseed grain BRSM = barley supplement; as concentrate B = barley grain F = feeding treatment, T = trial, F × T = feeding treatment and trial interaction. Differences between feedings were tested by making two orthogonal contrasts: C1 contrasts: two orthogonal by making tested were feedings between Differences interaction. trial and treatment = feeding T F × = trial, T treatment, F = feeding Fat cover: (1 = leanest, 5 fattest). trate supplement. = B vs. BRSM and C2 CC. Table 6. Effects of concentrate type on carcass characteristics and feed conversion growing dairy bulls. 6. Effects Table a b c d e 202 203 AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls Vol. 20 (2011): 191–205.

for the calves during the preweaning period. The ance of finishing dairy bulls (from 6 to 18 months) absence of any differences between treatments for with grass silage-barley-based feedings (Huusko- LWG or feed conversion during the same period nen et al. 2007, 2008, Huuskonen 2009). As in the was a reflection of the similar ME and protein present experiment, also in studies by Huhtanen et intakes (Table 4). The average intake, LWG and al. (1989) and Aronen (1990) the positive effect of feed conversion parameters of the calves were on protein supplementation was restricted to only the the same level as in earlier studies with dairy bull early phase of the growth (i.e., LW below 300 kg). calves fed MR-grass silage-grain-based diets in a Similarly, calculations by Titgemeyer and Löest similar housing environment (e.g. Huuskonen et al. (2001) showed that while amino acids were the 2005, 2011, Huuskonen and Khalili 2008). limiting factor with lighter weight calves offered During the postweaning period both the BRSM grass silage, energy availability was the limiting and CC animals ate more both roughage and con- factor with heavier steers. In addition, often much centrate than the B animals. This difference in to- of the advantage of protein supplementation of tal intake together with differences in the chemical young cattle was lost during the finishing period contents of the concentrates led to the increasing due to compensatory growth (McGee 2005). energy and protein intakes and, finally, to the in- As in the present experiment, McGee et al. creasing LWG of the BRSM and CC bulls com- (2006) and Manninen et al. (2010) reported that pared to the B bulls during the postweaning pe- the dressing proportion, carcass conformation and riod. There are many potential reasons which could carcass fat score were unaffected by the concen- cause intake differences between treatments. One trate energy source. Similarly, in accordance with possible reason is the superior palatability of the CC many earlier studies (Huhtanen et al. 1989, Aro- compared to the barley grain. The CC was pelleted nen 1990, Huuskonen et al. 2007, 2008, Huusko- unlike the barley grain which might have affected nen 2009), protein supplementation had no effects the intake of concentrate. According to Spörndly on the dressing proportion, carcass conformation and Åsberg (2006) and Manninen et al. (2010), pel- score or carcass fat score of growing dairy bulls. lets which include small amounts of molasses have In conclusion, there were no notable differenc- good palatability. However, this explanation does es in intake and gain parameters between the treat- not explain the intake and gain differences between ments during the preweaning period but during the the B and BRSM diets in which the concentrates postweaning period commercial concentrate and were not pelleted. Some experiments have shown RSM supplementation clearly increased the dry a positive response of LWG and the hay (Aronen matter and energy intakes as well as the gain of 1990) or grass silage (Aronen 1990, Aronen et al. the calves compared with barley grain. However, 1992) intake of young dairy bulls to RSM supple- during the finishing and entire period, the treat- mentation. The positive effect of RSM on LWG ments had no effect on the LWG or carcass gain. was often explained by the increased feed intake Furthermore, the dressing proportion, carcass con- and thereby higher energy and protein intake. It is formation or carcass fat score of the bulls did not also possible that the B diet without protein sup- differ between treatments. Thus, concentrate with plementation was likely to provide inadequate sup- a higher protein concentration than barley grain is plies of protein or some amino acids for a growing not needed for growing and finishing dairy bulls bull in the early phase of growth. when they are fed high or medium digestibility Although the CP intake of the BRSM and CC and restrictively fermented grass silage and barley- bulls was higher than that of the B bulls during the based concentrate. Because the prices of commer- finishing period, the treatments had no effects on cial concentrates are generally higher in relation to the LWG or carcass gain. This is a logical result, barley and other grains, it is not economical to use because there were no differences in the energy commercial concentrates for feeding growing dairy intakes between treatments during the finishing pe- bulls. Still, this experiment indicates that it is not riod. Similarly, RSM had no effect on the perform- necessary to use commercial starter concentrates 202 203 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls

of dairy-breed bulls with TMR feeding. Livestock Sci- for dairy calves during the pre- and postweaning ence 110: 154–165. periods because much of the advantage of starter Huuskonen, A., Khalili, H. & Joki-Tokola, E. 2008. Need feeds compared with rolled barley was lost during for protein supplementation in the diet of growing dairy bulls fed total mixed ration based on moderate digest- the finishing period. ible grass silage and barley. Agricultural and Food Sci- ence 17: 109–120. Acknowledgements. This study was partially funded by Huuskonen, A., Khalili, H., Kiljala, J., Joki-Tokola, E. & the Employment and Economic Development Centre for Nousiainen, J. 2005. Effects of vegetable fats versus Northern Ostrobothnia. I would like to thank Mr. Lauri lard in milk replacers on feed intake, digestibility, and Jauhiainen for advice on the statistical analyses. I wish growth in Finnish Ayrshire bull calves. Journal of Dairy Science 88: 3575–3581. to express my gratitude also to Mr. Matti Huumonen and Huuskonen, A., Lamminen, P. & Joki-Tokola, E. 2009. The his personnel for technical assistance and their excellent effect of concentrate level and concentrate composition care of the experimental animals. The personnel at Animal on the performance of growing dairy heifers reared and Production Research in Jokioinen are also thanked for the finished for beef production. Acta Agriculturae Scandi- laboratory analyses. navica, Section A, Animal Science 59: 220–229. Huuskonen, A., Tuomisto, L., Kauppinen, R. 2011. Effect of drinking water temperature on water intake and performance of dairy calves. Journal of Dairy Science 94: 2475–2480. Karhula, T. & Kässi, P. 2010. Lihanautatilojen taloudellinen tilanne Suomessa ja vertailumaissa. In: Arto Huusko- References nen (ed.). Kehitystä naudanlihantuotantoon I. [Towards more efficient beef production I]. Tampereen yliopisto- Aronen, I. 1990. Barley protein and rapeseed meal as pro- paino Juvenes Print Ltd; Tampere, Finland. p. 9–34. (In tein supplements for growing cattle. Acta Agriculturae Finnish with English Abstract). Scandinavica 40: 297–307. MAFF, 1984. Energy Allowances and Feeding Systems for Aronen, I. & Vanhatalo, A. 1992. Heat-moisture treatment of Ruminants. ADAS Reference book 433. Ministry of Ag- rapeseed meal: effect on diet digestion, voluntary grass riculture, Fisheries and Food. Her Majesty`s Stationery silage intake and growth of Ay-bulls. Acta Agriculturae Office, London. 85 p. Scandinavica, Section A, Animal Science 42: 157–166. Manninen, M., Jauhiainen, L., Ruusunen, M., Soveri, T., Aronen, I., Toivonen, V., Ketoja, E. & Öfversten, J. 1992. Koho, N. & Pösö, R. 2010. Effects of concentrate type Beef production as influenced by stage of maturity of and level on the performance and health of finishing Her- grass for silage and level and type of supplementary con- eford bulls given a grass silage-based diet and reared centrates. Agricultural Science in Finland 1: 441–460. in cold conditions. Livestock Science 127: 227–237. Commission of the European Communities. 1982. Commis- Mayne, C.S., Agnew, R., Patterson, D.C., Steen, R.W.J., sion of the European Communities (Beef Carcass Clas- Gordon, F.J., Kilpatrick, D.J. & Unsworth, E.F. 1995. An sification) Regulations. Council Regulations 1358/80, examination of possible interactions between silage type 1208/81, 1202/82. Commission Regulations 2938/81, and concentrate composition on the intake characteris- 563/82, 1557/82, Brussels. tics of grass silage offered to growing cattle and dairy Herva, T., Virtala, A-M., Huuskonen, A., Saatkamp & H. cows. Animal Science 60: 514–515 W., Peltoniemi, O. 2009. On-farm welfare and estimated McGee, M. 2005. Recent developments in feeding beef cat- daily carcass gain of slaughtered bulls. Acta Agriculturae tle on grass silage-based diets. In: Ynze van der Hon- Scandinavica, Section A, Animal Science 59: 104–120. ing (Ed.) Book of Abstracts of the 56th Annual Meeting Huhtanen, P., Näsi, M. & Khalili, H. 1989. By-products from of the European Association for Animal Production, Up- integrated starch-ethanol production from barley in the psala, Sweden, 5–8 June 2005. Wageningen: Wagenin- diets of growing cattle. Journal of Agricultural Science gen Academic Publishers. p. 51–64. in Finland 61: 451–462. McGee, M., O’Kiely, P.O. & O’Riordan, E.G. 2006. Intake, Huuskonen, A. 2009. The effect of cereal type (barley ver- growth, carcass traits and plasma metabolites in finish- sus oats) and rapeseed meal supplementation on the ing steers offered contrasting concentrate energy sourc- performance of growing and finishing dairy bulls offered es at two feeding levels. Agricultural Research Forum grass silage-based diets. Livestock Science 122: 53–62. 2006, 32nd Annual Research Meeting of Irish Grassland Huuskonen, A. & Joki-Tokola, E. 2010. Performance of and Animal Production Association, Tullamore 15th and growing dairy bulls offered diets based on silages made 16th March. ISBN: 1-841704512, p. 101. of whole-crop barley, whole-crop wheat, hairy vetch Minson, D.J. 1982. Effect of chemical composition on feed and grass. Agricultural and Food Science 19: 116 –126. digestibility and metabolizable energy. Nutrition Ab- Huuskonen, A. & Khalili, H. 2008. Computer-controlled stracts and Reviews 52: 591–615. milk replacer feeding strategies for group-reared dairy Moisio, T. & Heikonen, M. 1989. A titration method for si- calves. Livestock Science 113: 302–306. lage assessment. Animal Feed Science and Technol- Huuskonen, A., Khalili, H. & Joki-Tokola, E. 2007. Effects ogy 22: 341–353. of three different concentrate proportions and rapeseed Moloney, A.P., McHugh, T.V. & McArthur, A. 1993. Growth meal supplement to grass silage on animal performance and rumen fermentation in steers fed silage and concen- 204 205 AGRICULTURAL AND FOOD SCIENCE

Huuskonen, A. Effects of different concentrate feedings on performance of bulls Vol. 20 (2011): 191–205.

trates differing in energy source. Irish Journal of Agricul- tralised blend of aliphatic organic acids. Irish Journal of tural and Food Research 32: 101. (Abstract). Agricultural and Food Research 33: 25–39. MTT. 2006. Rehutaulukot ja ruokintasuositukset (Feed ta- Spörndly, E. & Åsberg, T. 2006. Eating rate and preference bles and feeding recommendations) [online]. Agrifood of different concentrate components for cattle. Journal Research Finland, Jokioinen. Published 14.2.2006, [cit- of Dairy Science 89: 2188–2199. ed 1.11.2010]. Available at: http://www.agronet.fi/rehu- Steen, R.W.J. 1993. A comparison of supplements to taulukot/. URN:NBN:fi-fe20041449. grass silage for growing beef cattle. In: P. O’Kiely et al. Nousiainen, J., Rinne, M., Hellämäki, M. & Huhtanen, P. (eds.). Proceedings of 10th International Silage Confer- 2003. Prediction of the digestibility of the primary growth ence, Dublin City University, Dublin, Ireland. p. 206 –207. of grass silages harvested at different stages of maturity Titgemeyer, E.C. & Löest, C.A. 2001. Amino acid nutri- from chemical composition and pepsin-cellulase solubil- tion: Demand and supply in forage-fed ruminats. Jour- ity. Animal Feed Science and Technology 103: 97–111. nal of Animal Science 79 (E Supplement): E180–E189. O’Kiely, P. & Moloney, A.P. 1994. Silage characteristics Van Keulen, J. & Young, B.A. 1977. Evaluation of acid-in- and performance of cattle offered grass silage made soluble ash as a marker in ruminant digestibility studies. without additive, with formic acid or with a partially neu- Journal of Animal Science 44: 282–287.

204 205 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Kanerva,Vol. P. et 20(2011): al. Extraction 206–216. of prolamins

Improved extraction of prolamins for gluten detection in processed foods

Päivi Kanerva*, Tuula Sontag-Strohm, Outi Brinck and Hannu Salovaara Department of Food and Environmental Sciences, PO Box 66, FI-00014 University of Helsinki, Finland e-mail:[email protected]

A problem in gluten analysis has been inconsistent extractability of prolamins, particularly from processed foods consisting of unknown portions of prolamins from wheat, barley, and rye. This study aimed at improving the extraction of prolamins for immunological analysis, regardless of the cereal species and the production process. The prolamins were extracted with varying concentrations of ethanol, 1-propanol, and 2-propanol. Sodium dodecyl sulphate-polyacrylamide gel electrophoresis and Western blotting were applied to study the protein composition of the extracts and the antibody recognition of the prolamin subgroups. We characterized the affinities of prolamin-specific antibodies that are used in gluten analysis against the prolamin groups that were soluble in 40% 1-propanol. The antibody R5 recognized more abundantly the medium-molecular weight groups, including polymeric proteins, and less the high-molecular weight groups than the anti-ω-gliadin antibody. In the present study, the prolamins were most efficiently extracted by 40% 1-propanol with 1% dithiothreitol at 50 °C . The prolamins were extracted from processed bread samples with efficiency similar to that from untreated meal samples. Key-words: barley; celiac disease; gluten; prolamin; rye; wheat

Introduction which means excluding wheat, barley, or rye protein- containing products from the diet. By avoiding these Gluten intolerance is one of the most common food products, gluten-intolerant people are able to keep intolerances in the Western countries. Currently, its their intestinal mucosa healthy and live normal prevalence is about 1% and has been increasing healthy lives. However, following a gluten-free during recent years (Lohi et al. 2007). People with diet can be very difficult, since these cereals are gluten intolerance must maintain a gluten-free diet, widely used. Many products made of gluten-free

206 207 AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins Vol. 20(2011): 206–216.

cereals, such as maize, rice, and oats, may contain Two important parameters affect the quantita- proteins from wheat, barley, and rye as contaminants, tive analysis of prolamins. The first is the solubility and therefore result in risk for people with gluten of prolamins in the extraction solution, and the sec- intolerance. The safety of gluten-free products for ond is the antibody recognition of the prolamin sub- people with gluten intolerance is controlled by gluten groups. In immunological gluten assays, prolamins enzyme-linked immunosorbent assays (ELISA) (e.g. are commonly extracted with 60% (v/v) ethanol Skerritt and Hill 1990, Valdés et al. 2003, Morón et with or without reduction of disulfide bonds. How- al. 2008). These gluten methods are able to detect ever, previous studies indicate that other alcoholic very low levels of prolamin proteins, and new limits solutions, such as aqueous propanol, extract more for acceptable gluten contents in gluten-free products prolamins from the samples than aqueous ethanol have been set by the European Union (Commission (Lauriére et al. 1976, Shewry et al. 1980, Shewry regulation (EC) No 41/2009). Products that can be et al. 1983). The recommended extraction method included in a gluten-free diet are divided into two with sandwich R5 ELISA analysis for processed groups, one of which is gluten-free products and food samples is the so-called cocktail extraction. the other products with very low levels of gluten. The cocktail extraction solution contains 2-mer- The gluten-free products may contain a maximum captoethanol as a reducing agent and guanidine hy- of 20 mg of gluten per kg of food and the products drochloride as a disaggregating reagent to improve with very low gluten contents may contain between protein solubility. Whether prolamins are extracted 20 and 100 mg of gluten per kg of food. in their native or reduced form has considerable Gluten in the context of celiac disease refers impact on the extraction yield and the composition to proteins that are harmful to people with gluten of the extract. Due to the assumption that wheat intolerance. These are the prolamin proteins of contains about equal halves of the prolamins that wheat, barley, and rye. Prolamin proteins are – by require reduction (glutenins) and those that do not one definition – aqueous-alcohol soluble; however, (gliadins), the results are multiplied by two to ob- some of them form alcohol-insoluble structures tain the total gluten content. Multiplication is nec- (Shewry and Halford 2002). Nevertheless, these essary only if the prolamins are extracted without prolamins become soluble in alcoholic solutions in reduction; however, it is not correct to multiply the reduced form. Wheat prolamins are called gliadins results of the samples that are extracted in the pres- and glutenins, barley prolamins hordeins and rye ence of a reducing agent, since the sample already prolamins secalins. The prolamin group consists contains the total gluten fraction. The degree of of monomeric and polymeric proteins. Monomeric multiplication has also been questioned (Wieser prolamins may contain disulfide bonds in their in- and Koehler 2009). However, there remains the ner structure, whereas polymeric prolamins are question of whether the antibody can recognize all joined together by the disulfide bonds, forming of the prolamin groups that are extracted. large protein complexes (Shewry et al. 1984). The Processing of food alters the properties and the disulfide bonds can be broken by a reducing agent conformation of proteins. Therefore, it is highly such as dithiothreitol (DTT) or 2-mercaptoethanol. important to determine the solubility of proteins The opening of the disulfide bonds of polymeric from products that have undergone different pro- prolamins is needed to enhance the solubility of cesses. The most common process is heating. Heat- polymeric prolamins in aqueous alcohol solutions. ing, e.g. during baking, denatures the proteins and, The prolamins of wheat, barley, and rye have dif- thus, decreases significantly their solubility in com- ferent relative contents of monomeric and poly- mon extraction solutions (Lagrain et al. 2005). Pro- meric prolamins. The ratio between monomeric teins may also be hydrolyzed into smaller peptides and polymeric prolamins in wheat is between 1.5 during the process (e.g. in fermentation processes and 3.1 and in barley between 1.4 and 5.0 (Wieser and brewing), which usually increases their solu- and Koehler 2009). The ratio in rye is between 6.2 bility. Some food products undergo extrusion or and 8.2 (Gellrich et al. 2003). enzymatic treatments during processing, both of 206 207 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins which significantly change the solubility of pro- Equal amounts of the grains of each cultivar were teins. The high pressure denatures proteins and measured, mixed, and milled with a laboratory pin decreases their solubility. The decreased solubility mill (KT-30, Koneteollisuus Oy, Finland) to make of denatured proteins has raised concern and the wheat, barley, and rye meals. The barley meal was reduction of the proteins in heat-processed samples sieved after milling with a 600-µm sieve to remove is recommended during extraction. The reduction the larger seed coat parts. breaks the disulfide bonds and improves the solu- The three bread samples included: 1) a wheat bility of the proteins. However, the heat- and pres- bread that was a yeast-leavened toasting bread sure-induced changes are not the only reason for (Vaasan Arki Vehnäpaahto, Vaasan Oy, Estonia), 2) using reduction during extraction. The polymeric a flat barley bread that was homebaked using barley prolamins contain natural disulfide bonds in their meal (Sunnuntai, Raisio Oyj, Finland), water, and structure. To extract the total prolamin fraction of salt (no yeast added), and 3) a rye loaf bread made the samples, the reduction should also be used for by a sourdough process (Uotilan Aito Pälkäneen samples that are not heat-processed. Maalaislimppu, Uotilan Leipomo Oy, Finland). The aim of this study was to determine the The details of the samples are listed in Table extractability of wheat, barley, and rye prolamins 1. The moisture and ash contents were determined from bread samples that were processed by differ- using the methods of the American Association ent techniques including commercial yeast bread of Cereal Chemists: AACC 44-15A and AACC process and sourdough processes, and baking 08-02. The protein contents were analyzed wit the without yeast. To our knowledge, the optimiza- Dumas combustion method (Vario MAX CN, Ger- tion of prolamin extraction has only been shown many) using N x 5.7 for wheat and N x 6.25 for for meal and baked wheat samples, not for baked barley and rye (AACC method 46-30). rye and barley products. Since gluten analyses are performed in processed samples, we focused on optimization of the extraction of prolamins from Sample extraction the processed bread samples. Furthermore, two prolamin-specific antibodies were compared in this study for their ability to recognize different The wheat, barley, and rye meals were extracted prolamin groups of wheat, barley, and rye. with ethanol, 1-propanol, or 2-propanol in aqueous alcohol concentrations of 20%, 30%, 40%, 50%, 60% and 70% (v/v). In all, 160 mg of meal were extracted with 1.6 ml of alcohol solution for 20 Materials and methods min at 21 ºC or 50 ºC with continuous shaking at

Meal and bread samples Table 1. Moisture, ash, and protein contents of the meal and bread samples (dry weight). The wheat, barley, and rye grains were obtained Sample Moisture Ash Protein from Boreal Plant Breeding Ltd, Finland. The % % % wheat samples consisted of the cultivars Aura, Meal Wheat 11.6 1.55 14.5 Ilves, Kadett, Nisu, Polkka, Runar, Ruso, Satu, Barley 11.4 2.24 13.8 Tjalve, and Tähti. The barley samples consisted of Rye 11.3 1.57 11.9 the cultivars Annabell, Erkki, Inari, Jyvä, Kunnari, Minttu, Polartop, Rolfi, Saana, and Voitto. Ten rye Bread Wheat 29.2 n.d. 12.8 samples consisted of four cultivars: Akusti, Amilo, Barley 41.2 n.d. 9.7 Bor9415, and Picasso, which were grown at six dif- Rye 42.9 n.d. 11.2 ferent locations in Finland between 2001 and 2006. n.d. = not determined 208 209 AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins Vol. 20(2011): 206–216.

approximately 330 rounds per min in an incubator were separated in a NuPage gradient mini (8 cm x 8 (Heidolph Inkubator 1000, Germany). For the re- cm) (4 – 12%) Bis-Tris [Bis(2-hydroxyethyl)imino- duced samples 1%, 2%, 4%, 6% and 8% (w/v) DTT tris(hydroxymethyl)methane-HCl] gel (NP0322, was added to 50 % 1-propanol and the extraction Invitrogen, USA). A NuPage 2-(N-morpholino) was performed at 50 ºC for 20 min. After extraction, ethanesulfonic acid (MES) - SDS running buffer the samples were centrifuged at 15 800 g for 10 min was used with NuPAGE antioxidant (NP0002 and and the supernatants collected. All of the extractions NP0005, Invitrogen, USA). The run was done at were performed in triplicate. The protein contents 150 V and the running time was 70 min. A Novex were analyzed with the Dumas combustion method. Sharp Pre-stained Standard was used as a molecular A sequential extraction following a modified marker (LC5800, Invitrogen, USA). Osborne fractionation was performed for the so- After electrophoretic separation, the pro- dium dodecyl sulfate - polyacrylamide gel elec- teins were transferred to polyvinylidene fluoride trophoresis (SDS-PAGE) analysis (Osborne 1907). (PVDF) membranes (162-0177, BioRad Labo- The meals were first extracted three times with dis- ratories, USA) in 25 mM tris(hydroxymethyl) tilled water and then three times with 0.5 M NaCl aminomethane, 192 mM glycine, and 20% (v/v) at room temperature. Then the precipitates were methanol. The transfer was performed at 30 V in washed briefly with water before extraction three one hour. After blocking with 1% (w/v) bovine se- times with 40% 1-propanol at 50 ºC, three times rum albumin (BSA; A-6793, Sigma-Aldrich Corp., with 40% 1-propanol and 1% DTT at 50 ºC and USA), the membrane was incubated in a solution three times with SDS sample buffer (4% (w/v) SDS, of either a conjugated anti-ω-gliadin antibody (SA 20% (v/v) glycerol, 125 mM tris(hydroxymethyl) GL0301, Diffchamb, France), a conjugated anti- aminomethane, 1% (w/v) DTT, bromophenol blue, body R5 (Ridascreen Gliadin, R7001, R-Biopharm, pH 8.5). Germany) or a polyclonal rabbit anti-gliadin an- The bread samples were extracted with 40% tibody (G9144, Sigma-Aldrich Corp., USA). The 1-propanol containing 1% DTT at 50 ºC and 60 °C conjugated anti-ω-gliadin antibody and the conju- for 30 min or 60 min. Two sample:solution ratios gated antibody R5 were diluted to 1/250 and 1/22, of 1:40 and 1:10, which are commonly used in the respectively, with the dilution buffers provided by extraction protocols before the immunological glu- the assays. Sodium azide and BSA were added to ten analysis, were compared. The efficiency of the the buffers. After incubation, the membranes were extraction was compared with that of the cocktail washed and stained with tetramethylbenzidine extraction (R7006, R-Biopharm, Germany) which (W4121, Promega Corp., USA). The third mem- was recommended for use with the sandwich R5 brane was incubated with the polyclonal anti-glia- ELISA. The protein contents were analyzed with din antibody (1/1000 in a buffer of 50 mM Trisma the Dumas method, and the method of Lowry et al. base, 0.9% NaCl, 0.05 ml Tween, 1% BSA, 0.02%

(1951) was used to compare the extraction protocol NaN3, pH 7.5). After incubation with anti-gliadin with the cocktail extraction. antibody, the membrane was further incubated with anti-rabbit immunoglobulin G (Fc) AP conjugate (S3731, Promega Corp., USA) (diluted to 1/7500 SDS-PAGE and Western blotting with with the same buffer) and stained with BioRad AP conjugate A and B. The polyclonal anti-gliadin an- ω-gliadin, R5, and polyclonal antigliadin tibody was used to compare the results obtained antibodies with the ω-gliadin and the R5 antibodies, due to its ability to recognize prolamins on a wider scale. The samples were concentrated by allowing 100 µl of the meal sample extracts to evaporate to dryness at room temperature. The dried samples were solu- bilized in 40 µl of SDS sample buffer. The proteins 208 209 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins

Results DTT made no difference in the yield when extracting rye prolamins. Aqueous solutions of ethanol, 1-propanol, and Optimizing the extraction temperature, 2-propanol in six different alcohol concentrations reducing conditions, and alcohol concen- were used for the extraction of prolamins. The protein contents of the extracts were measured and tration in the meal samples compared. A 40% concentration of 1-propanol was the most efficient in extraction with all three cereals The extraction temperature, reducing conditions, (Figure 2), extracting about 54% ± 2% of prolamins and alcohol concentration were optimized in the of the total protein of wheat meal, 46% ± 4% of the meal samples. The temperature used during extrac- total protein content of the barley meal and 56% tion had a significant effect on the extraction yield. ± 2% of the total protein content of rye meal as Elevation of the extraction temperature from room measured by the Dumas combustion method. temperature (21 ºC) to 50 ºC increased the protein yields by 20–30%. Heating especially increased the extractability of high-molecular weight (HMW) and Optimizing the extraction temperature polymeric proteins of the cereals. Since heating to 50 ºC significantly increased the yield, this temperature and duration with the bread samples was used for all of the extractions of this study. The effect of disulfide bond reduction on the ex- The extraction temperature and duration of the traction yield was studied, using DTT at concentra- 1-propanol extraction were optimized further with tions of 1%, 2%, 4%, 6% and 8% (w/v) at 50 ºC. The the heat-processed samples. The prolamins from addition of DTT to the 50% 1-propanol extraction the bread samples were most efficiently extracted yielded 31% and 34% higher protein concentra- with 40% 1-propanol and 1% DTT at 50 ºC for 60 tions for wheat and barley, and 23% higher con- min at an extraction ratio of 1:10 (Table 2). Heat- centration for rye compared with the unreduced ing of the samples to 60 °C instead of 50 °C did extracts (Figure 1). Increasing the amount of DTT not significantly increase the extraction yield nor to 8% yielded somewhat higher concentrations for did the sample ratio of 1:40 (data not shown). The wheat and barley, whereas the addition of more aqueous 1-propanol solution was about 25% more

Prolamin content (% of total protein) 80

75 WHEAT

70 RYE 65 BARLEY 60

45 Fig.1. Effect of increasing dithiothreitol (DTT) concentration 40 on the solubility of wheat, bar- 35 ley, and rye prolamins in 50% 1 2 3 4 5 6 1-propanol at 50°C. The amount of added DTT (%)

210 211 AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins Vol. 20(2011): 206–216.

Prolamins (% of total protein of the meal) contains two steps. The Lowry method was used for 60 comparison of the extractions with the 1-propanol WHEAT 50 and the cocktail solution, because the extra nitrogen of the cocktail solution interferes with the Dumas 40 method. The cocktail solution contains nitrogen 30 in the form of guanidine hydrochloride. However, it should be noted that the protein contents by the 20 Lowry method are somewhat different from the 10 quantities obtained by Dumas, probably due to globulin standard of the Lowry method. The results 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 were used only for the comparative analysis of the

60 1-propanol and cocktail extraction. BARLEY 50 40 SDS-PAGE and Western blot analysis 30

20 The prolamins of wheat, barley, and rye were sepa-

10 rated by gel electrophoresis and blotted against three prolamin-specific antibodies. The anti-ω-gliadin 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 and R5 antibodies are commonly used in gluten analysis assays and, therefore, their abilities to 70 recognize prolamins were compared. The polyclonal RYE anti-gliadin antibody was used as a reference, due 60 to its known ability to recognize prolamins on a 50 wider scale. 40 Gel electrophoresis showed that the C hordeins 30 of barley and a proportion of the B hordeins were

20 dissolved in aqueous 1-propanol without reduc-

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Table 2. Proportion of prolamins of the total protein content alcohol % (dry weight) of the meal and the bread samples extracted by 40% 1-propanol with 1% dithiothreitol (DTT) at 50 °C or Fig. 2. Prolamins of the total protein content of barley, bythe cocktail solution, and analyzed by the Lowry method. rye, and wheat meal when extracted with six concentrations of ethanol (♦), 1-propanol (■), and 2-propanol (▲). 40% 1-propa- Cocktail The protein contents were analyzed by the Dumas method. Sample nol with 1% DTT solution (%) (%) Meal Wheat 57.6 ± 2.5 54.4 ± 3.0 efficient a solvent for the wheat prolamins from Barley 31.6 ± 0.6* 34.7 ± 1.3* the bread sample than the cocktail solution (Table Rye 31.9 ± 0.7 32.7 ± 0.8 2). All the other samples were extracted with equal Bread Wheat 81.4 ± 5.7* 56.9 ± 6.0* efficiency with both solutions. However, extraction Barley 38.4 ± 7.1 36.8 ± 3.3 with 1-propanol is much faster to perform than Rye 32.4 ± 5.2 29.4 ± 3.9 the cocktail extraction, since it contains only one Statistically significant differences are indicated with an as- extraction step, whereas the cocktail extraction terisk (p < 0.05). 210 211 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins tion (Figure 3 barley gel: lane 4). The remaining HMW secalins were extracted. Wheat prolamins polymeric B hordeins required reduction before clearly differed from barley and rye prolamins. they were dissolved and a proportion remained Wheat contained a large number of prolamins that insoluble until the SDS buffer was used for the required reduction (Figure 3 wheat gel: lanes 5 and extraction. The polymeric D hordeins dissolved in 6). All of the protein groups, including a proportion 1-propanol when reduction was used. As with the B of the HMW glutenin subunits were, however, pre- hordeins, some of the D hordeins required extrac- sent in the alcohol extract even without reduction. tion with the SDS buffer. Most of the rye prolamins In Western blot analysis, the monoclonal were extracted with aqueous 1-propanol without anti-ω-gliadin antibody, the monoclonal R5 an- reduction. After reduction, more of the γ-75 and tibody and the polyclonal anti-gliadin antibody

Fig. 3. SDS-PAGE (gel) and Western blot of protein fractions of barley, rye, and wheat extracted by 1) water, 2) 0.5M NaCl, 3) water, 4) 1-propanol, 5) 1-propanol+DTT, and 6) the SDS buffer. Proteins were blotted, using the polyclonal anti- gliadin antibody (P), the anti- ω-gliadin antibody (ω), and the monoclonal antibody R5 (R5).

212 213 AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins Vol. 20(2011): 206–216.

were compared with respect to their recognition samples. We also showed that the prolamin-specific of different prolamin subgroups (Figure 3). The antibodies differ substantially in their specificity for polyclonal anti-gliadin antibody recognized all different prolamin subgroups. of the prolamin groups of wheat, barley, and rye, Wheat, barley, and rye prolamins cause inflam- although, the recognition of ω-secalins was weak mation of the small intestine in people with gluten (Figure 3: lanes P4). The anti-gliadin antibody also intolerance. Prolamins consist of various proteins recognized some of the water- and salt-soluble that show wide recognition variability among ce- proteins (Figure 3: lanes P1–3). This indicates that liac patients (Vader et al. 2002). Wheat prolamins some prolamins may be slightly soluble in water and their role in celiac disease have been studied and saline. Therefore, the preextraction of water- extensively, whereas less focus has been on barley soluble albumins and salt-soluble globulins should and rye prolamins. However, these cereals, rye in not be performed to avoid the loss of prolamins. particular, are widely used in Northern and East- The anti-ω-gliadin antibody strongly detected the ern Europe and constitute a major source of fiber C and D hordeins but weakly the B hordeins of intake. Barley and rye are also grown in the same barley (Figure 3 barley ω: lanes 4–6). The antibody areas as oats, which are considered safe for most recognized the HMW and γ-75 secalins, whereas celiac patients (Salovaara et al. 2009). In com- the ω-secalins appeared only faintly in the blot mercial farming and trade practices, some mixing (Figure 3 rye ω: lanes 4–6). In the wheat extracts, of wheat, barley, or rye seeds may occur with the the anti-ω-gliadin antibody strongly detected the crops that are grown in the same areas, i.e. buck- ω-gliadins and HMW glutenin subunits (Figure 3 wheat, maize, and oats. Gluten analysis methods wheat ω: lanes 4–6). The anti-ω-gliadin antibody are based on the characteristics of wheat prolamins did not cross-react with the water- and salt-soluble and, therefore, it has been questioned how accu- proteins of wheat, barley, and rye. The antibody R5 rate are the results obtained for products contain- reacted strongly with the B hordeins, but weakly ing barley or rye (Kanerva et al. 2006). The total with the C hordeins and HMW proteins of barley extraction of prolamins of each cereal species must (Figure 3 barley R5: lanes 4–6). Moreover, it de- be well known to obtain reliable analysis results tected all of the secalin fractions of rye (Figure 3 for products containing any of the cereals that are rye R5: lanes 4–6). In the wheat extracts, antibody harmful for people with celiac disease. R5 reacted mainly with α-, β-, and γ-gliadins and Three alcohols of increasing aqueous alcohol low-molecular weight (LMW) glutenins (Figure 3 concentrations were studied for their extraction wheat R5: lanes 4–6). It also recognized ω-gliadins efficiency of wheat, barley, and rye prolamins. and HMW glutenin subunits, but less intensively. A Of these, 40% 1-propanol was the most efficient. antibody R5 also reacted with some of the water- Extraction with 60% ethanol, which is commonly and salt-soluble proteins of wheat and rye (Figure used in gluten ELISAs, showed similar results for 3 rye and wheat R5: lanes 1–3). wheat as for 40% 1-propanol; however, it was less efficient for barley and rye prolamins. Heating the samples to 50 ºC instead of extracting at room temperature significantly increased the extraction yields of the prolamins. However, no significant in- Discussion crease in the yields was obtained when the samples were extracted at 60 °C. Heating during extraction The extraction characteristics are major factors increased especially the yield of the polymeric and affecting the reliability of the analysis of gluten higher molecular weight prolamins. Reduction of contamination in gluten-free products. In the present the proteins with different concentrations of DTT study, extraction with 40% 1-propanol with reduc- indicated that the increase in DTT concentration tion at 50 °C for 1 h was the most efficient method increased the extraction yields. Although the high for prolamin extraction from commercial bread DTT concentrations slightly increased the prolamin 212 213 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins yield in wheat and barley, a 1% DTT concentration on processed breads, since food processing may was considered sufficient, since high concentra- alter the solubility of prolamins, e.g. the denatura- tions of reducing agents interfere with antibody- tion caused by heating decreases the solubility of antigen reactions (Doña et al. 2008). This also the proteins, whereas the depolymerization of glu- illustrates why SDS and urea were avoided in tenins increases their solubility in aqueous alcohols the extraction, regardless of their known extrac- (Loponen et al. 2004). In addition, the formation of tion power. In addition, extraction power could gluten network during mixing and baking induces be increased by increasing the temperature and, changes in the extractability of wheat prolamins. therefore, the amount of extractable protein could The gluten network is formed when wheat flour is also be increased. However, high temperatures mixed with water. Mixing induces changes, such alter protein structure and, as a consequence, the as new disulfide bonds between prolamins, that immunoreactivity of the proteins may be lost. In influence their extractability in different solvents the present study, the samples were extracted by a (Kuktaite et al. 2004). Heating also causes such one-step procedure. Another alternative would be strong bonds between the prolamins that even re- to use more extraction steps, which are effective in duction cannot result in the same solubility of the prolamin extractions (Bean et al. 1998). However, prolamins that they had before the heat treatment we considered a simple one-step extraction to be (Lagrain et al. 2011, Rombouts et al. 2011). All of more practical when combining the extraction pro- these challenges quantitative gluten analysis, since tocol with immunological gluten analysis. almost inevitably some of the prolamins in proc- Estimation of the efficacy and the yield of the essed foods remain unextracted. Another obstacle extraction, with respect to the total prolamin con- is caused by the nature of immunological analysis, tent of the sample, is difficult, especially because since very small changes caused by the extraction the relative proportion of prolamin proteins in the conditions in the protein structure may abolish the total protein content varies among crops and culti- reactivity of protein with the antibodies and leave vars. In seeds, the relative proportion of prolamins it thus undetected. The extraction conditions must increases when the total protein content increases be such that the affinity of prolamins is not lost. (Kirkman et al. 1982). If the prolamin content of the Comparison of the three antibodies and their sample is high, a more efficient extraction protocol cross-reactivity with the prolamins of wheat, bar- is needed than for the sample with low prolamin ley, and rye revealed differences among the anti- content. However, extracting the prolamins in their bodies. The polyclonal antibody recognised all of reduced form makes it possible to extract proteins the prolamin groups of wheat, barley and rye. The from all prolamin groups and, thus, to increase the anti-ω-gliadin antibody recognized mainly the high yield. The yield is also improved by selecting more molecular weight proteins of wheat, barley and rye. hydrophobic solvents, since prolamins are hydro- Antibody R5, on the other hand, recognized mainly phobic due to their high content of hydrophobic the medium-molecular weight prolamin groups and amino acids. The hydrophobicity of 1-propanol is less intensively the high molecular weight groups. slightly higher than that of ethanol, which increases The polyclonal anti-gliadin antibody and the R5 its efficiency in prolamin extraction. In the present antibody also recognized some of the water- and study, the solubility of the prolamins of wheat salt-soluble proteins of wheat and rye. Although bread increased by 25% with 40% 1-propanol, prolamins in the native state are characterized as compared with the yield with the cocktail solution. insoluble in water and saline solutions, some of Whether prolamins are extracted with or with- them can be water-saline-soluble (Fu et al. 1996). out reduction has a significant impact on the extrac- Since sample extractions prior to ELISA analysis tion yield and prolamin composition. The prolamin are performed directly from the meal or food sam- subgroups of wheat, barley, and rye meal could be ples, some of the water- and salt-soluble protein collected in an alcohol solution in the presence of a are co-extracted. reducing agent. In the present study, the focus was 214 215 AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins Vol. 20(2011): 206–216.

This study showed that different prolamin sub- low gliadin concentrations raises questions as to groups have special affinities with the antibodies whether the limit of 20 mg kg-1 of gluten is appro- used, which is in agreement with the findings of priate or set too low with respect to the capacity Wieser and Seilmeier (1999). At low gluten con- of the analytical methods. The relative error of the centrations, the characteristics of prolamins play a gluten assay with antibody R5 is 30% (Méndez et crucial role in gluten analysis since the composi- al. 2005). The gliadin concentration of 20 mg kg-1 tion of different prolamin subgroups in the sample would lead to results between 14 and 26 mg kg-1, extract plays a key role in quantification. Immu- which is a relatively wide range when compared nological methods are based on a chosen standard with the regulatory limit of 20 mg kg-1 for gluten- that contains certain amounts of different prolamin free products. subgroups. Therefore, the affinity of antibody to- In conclusion, this study showed that the wards the standard is actually a sum of the various prolamins of wheat, barley, and rye from processed affinities towards the prolamin subgroups repre- bread samples are most efficiently extracted with sented in the standard. If the composition of the 40% 1-propanol containing 1% DTT at 50 °C. The prolamin groups in the sample distinctly differs use of such an efficient method in the total extrac- from that of the standard, the analysis results may tion of prolamins would improve the accuracy of be unreliable. Moreover, the extraction method has current analysis, since 1) the need for multiplica- a major effect on the prolamin composition and, tion of the results by two is excluded because all therefore, plays an important role in the quanti- of the prolamin groups are extracted and detected fication of gluten. The prolamin composition of by the prolamin-specific R5 antibody, and 2) the the sample is different if the proteins are extracted method uniformly extracts the prolamins of wheat, under reducing conditions. The current practice is barley, and rye, which is an advantage when ana- to multiply the prolamin results in the gluten as- lyzing unknown samples and processed samples. says by two to obtain the total gluten content. This is done for two reasons: first, assuming that only Acknowledgements.The authors are grateful for the monomeric prolamins are extracted with aqueous financial support of the Finnish Funding Agency for alcohol solutions (without reduction) and secondly, Technology and Innovation (TEKES) and The Finn- assuming that the prolamin-specific antibodies do ish Graduate School on Applied Bioscience (ABS). not recognize glutenins (Valdés et al. 2003). How- ever, significant amounts of polymeric glutenins are extracted in propanolic solutions without reduction (Fu and Sapirstein 1996), and if the proteins are extracted with reduction, the amount of References polymeric proteins increases further. The Western Bean, S.R., Lyne, R.K., Tilley, K.A., Chung, O.K., & blot analysis of this study confirmed that both the Lookhart G.L. 1998. A rapid method for quantitation anti-ω-gliadin and the R5 antibody recognized the of insoluble polymeric proteins in flour. Cereal Chem- polymeric glutenins, which were extracted with istry 75:374-379. Doña, V. V., Fossati, C. A., & Chirdo, F. G. 2008. Inter- aqueous alcohols in reduced form and were visible ference of denaturing and reducing agents on the an- in Figure 3 lanes 5 and 6 of the wheat blots. These tigen/antibody interaction. Impact on the performance characteristics enable the direct quantification of of quantitative immunoassays in gliadin analysis. Eu- ropean Food Research and Technology 226: 591-602. prolamins without multiplication. Gellrich, C., Schieberle, P., & Wieser, H. 2003. Biochem- Current immunological methods for gluten ical characterization and quantification of the storage quantification are able to detect prolamins at very protein (secalin) types in rye meal. Cereal Chemistry 80: 102-109. low concentrations. Low regulatory gluten limits Fu, B.X., & Sapirstein, H.D. 1996. Procedure for isolating demand high accuracy from the analysis method monomeric proteins and polymeric glutenin of wheat that is used for gluten quantification. The variabil- flour. Cereal Chemistry 73:143-152. ity in the results obtained in ELISA methods at Fu, B. X., Sapirstein, H. D., & Bushuk, W. 1996. Salt-in- 214 215 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Kanerva, P. et al. Extraction of prolamins

duced disaggregation/solubilization of gliadin and glu- Osborne, T.B. 1907. The proteins of the wheat kernel. tenin proteins in water. Journal of Cereal Science 24: Publication No. 84. Carnegie Institution of Washington, 241-246. Press of Judd & Detweiler, Inc., Washington D.C., USA. Kanerva, P. M., Sontag-Strohm, T. S., Ryöppy, P. H, Alho- Rombouts, I., Lagrain, B., Brunnbauer, M., Koehler, P., Bri- Lehto, P., & Salovaara, H.O. 2006. Analysis of barley js, K. and Delcour, J.A. 2011. Identification of isopep- contamination in oats using R5 and ω-gliadin antibod- tide bonds in heat-treated wheat gluten peptides. Jour- ies. Journal of Cereal Science 44: 347-352. nal of Agricultural and Food Chemistry 59:1236-1243. Kirkman, M. A., Shewry, P. R., & Miflin, B. J. 1982. The ef- Salovaara, H., Kanerva, P., Kaukinen, K., & Sontag-Strohm, fect of nitrogen nutrition on the lysine content and pro- T. 2009. Oats - an overview from a celiac disease point of tein composition of barley seeds. Journal of Science of view. In: Arendt, E. K. & Dal Bello, F. (Eds.). The Science Food and Agriculture 33: 115 -127. of Gluten-Free Foods and Beverages. St. Paul, Minne- Kuktaite, R., Larsson, H., & Johansson, E. 2004. Varia- sota, USA: AACC International Inc. p. 69-81. tion in protein composition of wheat flour and its rela- Shewry, P. R., Field, J. M., Kirkman, M. A., Faulks, A. J. & tionship to dough mixing behaviour. Journal of Cereal Miflin, B. J. 1980. The extraction, solubility, and charac- Science 40:31-39. terization of two groups of barley storage polypeptides. Lagrain, B., Brijs, K., Veraverbeke, W.S. and Delcour, J.A. Journal of Experimental Botany 31:393-407. 2005. The impact of heating and cooling on the physic- Shewry, P. R., & Halford, N. G. 2002. Cereal seed stor- chemical properties of wheat gluten-water suspensions. age proteins: structures, properties and role in grain Journal of Cereal Science 42:327-333. utilization. Journal of Experimental Botany 53: 947-958. Lagrain, B., Rombouts, I., Brijs K., & Delcour, J.A. 2011. Shewry, P. R., Miflin, B. J., & Kasarda, D. D. 1984. The Kinetics of heat-induced polymerization of gliadin. Jour- structural and evolutionary relationships of the prola- nal of Agricultural and Food Chemistry 59:2034-2039. min storage proteins of barley, rye and wheat. Philo- Lauriére, M., Charbonnier, L., & Mossé, J. 1976. Nature sophical Transactions of the Royal Society of London et fractionnement des protéines de l’Orge extraites par B 304: 297-308. l’éthanol, l’isopropanol et le n-propanol à des titres dif- Shewry, P. R., Parmar, S., & Miflin, B. J. 1983. Extraction, férents. Biochimie 58: 1235-1245. separation, and polymorphism of the prolamin storage Lohi, S., Mustalahti, K., Kaukinen, K., Laurila, K., Collin, proteins (secalins) of rye. Cereal Chemistry 60: 1-6. P., Rissanen, H., Lohi, O., Bravi, E., Gasparin, M., Reu- Skerritt, J. H., & Hill, A. S. 1990. Monoclonal antibody sand- nanen, A., & Mäki, M. 2007. Increasing prevalence of wich enzyme immunoassays for determination of glu- coeliac disease over time. Alimentary Pharmacology & ten in foods. Journal of Agricultural and Food Chem- Therapeutics 26: 1217-1225. istry 38: 1771-1778. Loponen, J., Mikola, M., Katina, K., Sontag-Strohm, T., Vader, W., Kooy, Y., Van Veelen, P., De Ru, A., Harris, D., & Salovaara, H. 2004. Degradation of HMW glutenins Benckhuijsen, W., Peña, S., Mearin, L., Drijfhout, J. W., during wheat sourdough fermentations. Cereal Chem- & Koning, F. 2002. The gluten response in children with istry 81: 87-93. celiac disease is directed toward multiple gliadin and Lowry, O., Rosebrough, N., Farr, A. & Randall, R. 1951. glutenin peptides. Gastroenterology 122: 1729-1737. Protein measurement with the folin phenol reagent. The Valdés, I., García, E., Llorente, M., & Méndez, E. 2003. In- Journal of Biological Chemistry 193: 265-275. novative approach to low-level gluten determination in Méndez, E., Vela, C., Immer, U., & Janssen, F. W. 2005. foods using a novel sandwich enzyme-linked immuno- Report of a collaborative trial to investigate the perfor- sorbent assay protocol. European Journal of Gastroen- mance of the R5 enzyme linked immunoassay to deter- terology & Hepatology 15: 465-474. mine gliadin in gluten-free food. European Journal of Wieser, H., & Koehler, P. 2009. Is the calculation of the Gastroenterology & Hepatology 17: 1053-1063. gluten content by multiplying the prolamin content by a Morón, B., Cebolla, Á., Manyani, H., Álvarez-Maqueda, factor of 2 valid? European Food Research and Tech- M., Megías, M., del Carmen Thomas, M., López, M.C., nology 229: 9-13. & Sousa, C. 2008. Sensitive detection of cereal fractions Wieser, H., & Seilmeier, W. 1999. Reactivity of gliadin frac- that are toxic to celiac disease patients by using mono- tions and components from different wheat species in clonal antibodies to a main immunogenic wheat peptide. the Skerritt test. In: Stern, M. (Ed.). Proceedings of the The American Journal of Clinical Nutrition 87: 405-414. 13th Meeting of the Working Group on Prolamin Analy- sis and Toxicity. Tuebingen, Germany. p. 11-18.

216 PB AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method Vol. 20(2011): 217–227.

Characteristics of dry- and brine-salted salmon later treated with liquid smoke flavouring

Olaia Martinez*, Jesús Salmerón, María D. Guillén and Carmen Casas Departamento de Farmacologia y Ciencias de los Alimentos, Facultad de Farmacia, Universidad del País Vasco (UPV) 01006-Vitoria, Spain. *e-mail: [email protected]

The use of smoke flavourings for the processing of salmon has begun to substitute traditional smoking methods. This review examines the quality issues associated with salted salmon ‘smoked’ by this technique along the salting and smoking steps. Firstly, the evidence is examined to determine whether dry or brine salting is better for salmon flesh destined to be treated by liquid smoking. Secondly, influence of liquid smoking on the sensorial, physicochemical and textural characteristics of the flesh are described, as are its effects on potential spoilage organisms.

Key-words: Salmon, liquid smoke flavouring, dry and brine salting.

Introduction The process of cold-smoked salmon includes salting and drying before any type of smoking, and the final product typically contains between Atlantic salmon (Salmo salar) is an economically 2.0–3.9% salt in the water phase (Bannerman and and nutritionally important cultured fish species Horne 2001, Huang et al. 2002; Oenhlenschlager (Dondero et al. 2004). About 40–50% of farm-reared 2007, Bocker, et al., 2008). Salmon thus salted is Atlantic salmon reaches the consumer as traditional categorized as “lightly preserved” (Leroi 2010). cold-smoked gourmet food products (Birkeland Salt functions as a flavour enhancer, increasing and Bjerkeng 2005); indeed, a number of countries the perception of the fullness, thickness and sweet- have long traditions of producing and consuming ness of food (Gallart-Jornet et al. 2007a). It may smoked food products (Birkeland and Skara 2008). be added by injection, by dry salting or by brine

PB 217 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method salting, the latter two being the most commonly transfers during these steps or acting as physical used in the salmon processing industry (Birkeland barrier (Cardinal et al. 2001). and Bjerkeng 2005). When a salt solution or dry salt is used for Smoking impregnates high-protein foods with salting, two main simultaneous fluxes occur: the aromatic components. These lend flavour and col- uptake of salt by the muscle and the loss of water our to the finished product as well as playing a from it (Gallart-Jornet et al. 2007b, Czerner and bacteriostatic and antioxidant role (Kristinsson et Yeannes 2010), a consequence of the differences al. 2008). Along the last decades traditional smok- in the osmotic pressure of the muscle cells and the ing methods have given way to the use of smoke salting agent (Barat et al. 2002). Table 1 shows flavourings. Treatment with these flavourings is several salting protocols and their effect in the flesh much cheaper and has a much less toxic effect quality. In dry salting, a water interface is created on the environment (Pszczola 1995, Underwood by the extraction of the intercellular water to the and Shoop 2007). In addition, their use avoids the surface of the flesh. In brine salting, however, the formation of polycyclic aromatic hydrocarbons, a products are soaked in a solution which reduces the drawback of the traditional process (Guillen et al. outward diffusion of water (Rora et al. 2004, Bel- 1996). Smoke flavourings of different composition lagha et al. 2007). The use of dry salt causes fiber are available and can be combined to obtain prod- shrinkage (Sigurgisladottir et al. 2000, Bocker et al. ucts with different sensorial qualities (Abu-Ali and 2008), the cross-sectional area of the muscle fibers Barringer 2007, Gedela et al. 2007). becoming smaller after such treatment than after The optimisation of the individual operations brine salting. Brine-salted salmon can, however, involved in the processing of smoked salmon, i.e., show a wide variation in muscle fiber size, with salting, drying and smoking, is important in the some fibers expanding, others shrinking and others economic success of this sector which has small still showing no change. This might be an effect profit margins (Birkeland et al. 2004). This paper of an uneven salt distribution within the muscle. summarises the available information on the salting Montero et al. (2003) reported that differences procedures employed and tries to determine wheth- in protein solubility (sarcoplasmic and myofibrillar er dry salting or brine salting is better for use with fractions) between salmon processed using differ- salmon to be smoked with liquid smoke flavour- ent salting methods were not significant, and that ings. Moreover, the most outstanding features im- protein solubilisation was not influenced by the salt parted by liquid smoke flavouring to salmon flesh level. However, salt is well known to affect the after salting by different methods are discussed. structural proteins of muscle (Larsen et al. 2008) and the changes in rates of salt penetration closely follow changes in the amount of extractable protein Salting (Minh et al. 2011). Brine salting causes the solubilisation of some of the muscle proteins, which are released into The uptake and distribution of salt in fish fillets the brine (Unlusaying et al. 2010). Martinez et al. depends on the salting method used, the species (2005) and Martinez-Alvarez and Gómez-Guillén in question, the thickness of the fillet, the fish/salt (2006) indicate that the loss of nitrogenous com- weight ratio (Gallart-Jornet et al. 2007b,2007c, Bras ponents during brine salting may be due to the and Costa 2010), and intrinsic flesh factors such as enhancement of protein solubility, a consequence the composition and structure of the muscle and the of the increasing salt content of the tissues. Salt-in- rigor condition (Aursand et al. 2009, 2010). Lipids duced protein destruction would be followed by the in flesh play an important role as a limiting factor leaching of protein components into the brine. At during the salting and drying steps, either replac- the same time, the cations introduced into the flesh ing the aqueous phase that serves as a vector for might also induce conformational changes and affect protein/water interactions, and hence protein 218 219 AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method Vol. 20(2011): 217–227. texture- shear force (N) - - - - - 30-32 36-37 27-28 33-35 - - - - 62-65 70-74 69-71 65-76 WHC (%) - - - - - 62.3 + 1.7 63.9 + 1.2 66.0 + 1.6 64.4 + 0.9 ------

2 (%) - - After smoking total yield - - 77.58 - - - - 90.9-91.9* 92.7-93.1 94.15-94.6* 94.9-95.0 91.1-92.2 92.7-93.1 94.2-94.6 95.0 - - T (ºC) T - - 70 + 1 20 + 1 30 + 1 20 + 1 30 + 1 20 + 1 30 + 1 20 + 1 30 + 1 20 + 1 30 + 1 20 + 1 30 + 1 5 5 time (h) 5 5 0.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 5 5 5 5 liquid-smoking1, 5% (DS) liquid-smoking1, 10% (DS) method liquid-smoking1, 5% (BS) Smoking process liquid-smoking1, 10% (BS) Hot-smoking (%50 smoke) cold-smoking (DS) cold-smoking (DS) cold-smoking (BS) cold-smoking (BS) cold-smoking, beech (DS) cold-smoking, beech (DS) cold-smoking, beech (BS) cold-smoking, beech (BS) cold-smoking, beech (DS) cold-smoking, beech (DS) cold-smoking, beech (BS) cold-smoking, beech (BS) Total yield, water holding capacity (WHC) and texture are shown for some of them. Total acids, 11–12%; benzo(a)pyrene, <10 ppm; *Preferred from a sensorial point of view; acids, 11–12%; - - T (ºC) T 8 + 1 12 + 1 12 + 1 12 + 1 12 + 1 12 + 1 12 + 1 WB, Warner-Bratzler; DS, Dry-smoked; BS, Brine-smoke Warner-Bratzler; WB, 24 24 time (h) 2 12 6 6 6 6 6 Dry-salting Brine -salting (30%) method Salting process Brine-salting (125 g/l, ratio 1:1 w/v) Dry-salting Brine-salting (360g/l, ratio 50/50 w/w) Dry-salting Brine-salting (360 g/l, ratio 50/50 w/w) Dry-salting Brine-salting (360 g/l) Xiphias gladius fish species

Scomber japonicus Salmo salar Salmo salar Salmo salar Composition: pH, 2.7; phenol level, 12–24 mg/g; carbonyl 14–25 mg/100 ml; Muratore et al. (2007) Author(s)

Table 1. Different smoking protocols proposed for smoking of various fish species: salting and smoking methods are described. As different objectives were stated for As different smoking protocols proposed for of various fish species: salting and methods are described. 1. Different Table each work, it is difficult to find measurements of the same quality parameters. Goulas and Kotominas (2005) Montero et al. (2003) Cardinal et al. (2001) Sigurgisladottir Sigurgisladottir et al. (2000) Calculated taking the first clean fillet weight as initial measurement; 1 2 218 219 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method solubility. Hultmann and Rustad (2004) and Carton ones to show significantly greater firmness and less et al. (2009) suggest that the conformation of the elasticity than low concentrated brine-salted fillets. myofibrillar proteins may change as a result of the The same tendency is observed when comparing increased salt content of the muscle rendering the early and last salting stages with saturated brines. proteins more susceptible to attack by endogenous Finally, several authors (Aubourg and Ugliano proteases. 2002, Guillen et al. 2004, Goulas and Kontomi- Birkeland et al. (2004) and Gallart-Jornet et al. nas 2005, 2007, Yanar et al. 2006, Alfonso and (2007b, 2007c) report that fillets subjected to dry Sant´Ana 2008) report that the contact of the salt salting show a significantly higher total liquid loss with the flesh enhances lipid oxidation. Some stud- than fillets subjected to brine salting. However, ies have shown that the stimulation of lipid oxida- brine salting shows great water losses and smaller tion occurs through the activation or iron. Sodium water phase in protein matrix when saturated brine ions displace the iron from macromolecules such as (25% w/w) is used (Barat et al. 2002). This lead to myoglobin, increasing the availability of the latter a lower water holding capacity, comparing with for the catalysis of lipid oxidation (Thiansilakul less saturated brines (20% w/w) and important et al. 2010, Kashiri et al. 2011). Several authors changes in texture (Sannaveerappa et al. 2004, indicate that brine salting offers the major advan- Thorarinsdottir et al. 2004). In fact, Instrumentally tages of preventing rancidity by precluding contact measured texture parameters (Force, Area) show an with the air and of affording a higher weight yield exponential increase with respect to the NaCl con- caused by the uptake of water (Andrés et al. 2005, centration in the liquid phase (Barat et al. 2002). Gallart-Jornet et al. 2007b,c). However, Muratore Sigurgisladottir et al. (2000) report the shear force et al. (2007) indicates brine-salted liquid-smoked required for cutting dry salted fillets to be lower salmon to have poorer sensorial qualities, as judged than that required to cut brine-processed fillets, for by a tasting panel. Moreover, the soluble compo- certain fish species and in a season dependent way. nent of the fillets can leach out during brine salt- Nevertheless, most authors (Birkeland et al. 2004, ing (Larsen et al. 2007). Thus, brining may have Gallart-Jornet et al. 2007b) describe harder textures economic advantages such as increasing the yield in fish treated with saturated brines and even harder but nutritional components are lost from the flesh for dry-salted fish fillets. Texture of fish is related to during the process (Larsen et al. 2008). the diameter of the muscle fibers: smaller diameters Salting has a noticeable preservative effect. and higher number of fibers give harder textures This is mainly due to the reduction in water activ- and larger diameters but less number of fibers, ity and the consequent prevention of the growth softer (Hatae et al. 1990). The greater reduction of of many spoilage microorganisms, along with the the cross sectional area in dry-salting processes has formation of a more membranous surface which been suggested as the reason for their firmer tex- further inhibits microorganism growth (Muratore ture when comparing them with the not saturated et al. 2007, Frangos et al. 2010). brine-salted ones. However, fish texture in known to be a multifactorial quality in which fiber size is not necessarily the major determinant (Johnston et Smoking al. 2006, Morkore et al. 2009). Cardinal et al. (2001) reported that dry and brine salting have little effect on the sensorial properties Smoking is a centuries-old food preservation of salmon. However, Birkeland et al. (2003) indi- technique, but nowadays fish products are mainly cate dry salting to be a ‘more gentle’ curing method smoked because of the attractive flavour and colour that helps retain the intactness of the fillet surface this process confers upon them (Guillen and Man- and its colour - important consumer quality criteria. zanos 1996). It is well known that smoking food Moreover, Gallart-Jornet et al. (2007b) report dry- enriches it in the aromatic compounds contained in salted fillets and the more saturated brines salted the smoke. Processes of adhesion, condensation, 220 221 AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method Vol. 20(2011): 217–227.

diffusion and/or absorption occur between the and coniferaldehyde, which are strongly retained food and the smoke. As a consequence, generally by food and intensify the aroma of smoked fish desirable changes occur in the colour, taste, tex- very well (Varlet et al. 2007b). Their adhesion also ture and composition of the food. These changes prevents them from being lost over time by evapo- vary depending on the composition of the smoke ration (Guillén and Manzanos 1996) or smoke flavouring used, the composition of the The aroma of liquid-smoked salmon depends food itself, and the conditions under which either on the type of liquid smoke used, the surface com- type of smoking is undertaken (Toth and Potthast position of the fish, and the conditions under which 1984) (Table 1). the treatment is performed (Sérot et al. 2004). The The treatment of salmon with liquid smoke volatile aldehydes responsible for the aroma of leads to changes in its physicochemical and sen- smoked fish are generated by two main pathways: sorial attributes, depending on the composition of the Maillard reaction (i.e., interactions between the liquid flavouring (and therefore on the type of smoke carbonyl groups and food amino groups), wood originally used to produce it) (Toth and Pot- which provides smoke aromas, and lipid oxidation, thast 1984). Guillén et al. (2005) reported liquid which provides fishy aromas (Varlet et al. 2007b). beech and oak flavours to generate different fish The compounds making up the aromatic profile headspace compositions and, therefore, different of liquid-smoked salmon are different to those as- flavours in salmon flesh. The headspace of salmon sociated with traditional smoking; derivatives of smoked with liquid oak smoke is richer in phenol pyridine have been detected, along with the prod- derivatives than samples smoked with liquid beech ucts of lipid oxidation (Varlet et al. 2007c, 2007d). smoke. Cardinal et al. (1997) reported that, depending on Little attention has been paid to the physi- the type of liquid smoke used, the products of li- cochemical attributes of fish treated with liquid pid oxidation may be so abundant as to mask the smoke. Bugueño et al. (2003) reported that salmon product’s own fishy odour. treated in this way showed no significant changes Guaiacol and its derivatives have been de- in any of the chemical or physical variables they scribed responsible for the ‘smoked’ taste and investigated (pH, total volatile nitrogenous bases, syringol and its derivatives for the smoked smell. etc.). However, Martinez et al. (2007a, 2009) re- Other compounds mainly associated with this ported changes in non-protein nitrogen, proteins, smoke aroma and flavour include phenol, p-cresol, and in the fat and moisture contents. These authors o-cresol, guaiacol, 4-memethyguaiacol, 4-ethyl used two liquid smokes of different composition to guaiacol, eugenol, 4-propylguaiacol, and isoeug- smoke salmon, one rich in phenolic compounds, enol (Sérot and Lafficher 2003, Varlet et al. 2006, the other rich in carbonyl compounds. The fish Varlet et al. 2007a). The final aroma of the salmon treated with the latter showed larger quantities of is, however, a product of a much more complex non-protein nitrogen and smaller quantities of pro- mix of compounds (Daun 1972, Maga 1987). tein (both sarcoplasmic and myofibrillar), fat and The taste of liquid-smoked salmon is also moisture. very different to that produced by the traditional It should be remembered that the sensorial at- technique, especially its in terms of the descrip- tributes of a food are vital for consumer accept- tor “salty” (Varlet et al. 2007c). This could mean ance, and this may be no more true than with that liquid smoke, as well as generating aromas, smoked foods. The following lines indicate the is involved in physical interactions that influence changes these attributes may undergo in salmon saltiness. treated with liquid smoke. The conspicuous pink coloration of salmonid fish The aroma and flavour of smoked foods are flesh is owed to carotenoid pigments, notably astax- mainly owed to volatile compounds in the liquid anthin (3,3´-dihydroxi-β,β-carotene-4,4´-dione) and smoke adhering to the flesh. Of special importance canthaxantin (β,β-carotene-4,4´-dione), that the fish are phenolic compounds such as syringaldehyde accumulate through their diet (Choubert et al. 2005, 220 221 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method

Rora et al. 2005, Webb et al. 2007). After smoking, is greater when the temperature reaches closer to the colour of fish flesh can vary depending on the 30 ºC than 20 ºC (Gómez-Guillén et al. 2000, Hult- pigments present, the quantity and composition of mann et al. 2004). smoke deposits, and the interactions between these The water content of fish flesh strongly influ- and the flesh (Sirkoski et al. 1998). However, the ences its texture (Rongrong et al. 1998), with lower colour of the product is mainly due to Maillard- water contents producing firmer products. Montero type reactions (Ruiter 1979). Oily fish commonly et al. (2003) showed the water content of salmon go browner because of a reaction between proteins fillets to be reduced after traditional cold smok- or amino acids and the carbonyl derivatives of lip- ing. Martinez et al. (2007a), working with salmon ids. This reaction is influenced by the same factors smoked with liquid smoke, reported a reduction that influence lipid oxidation (temperature, oxygen in the water content and an increase in its hard- water activity, and the presence of oxidants and ness, fracturability and cohesiveness, especially if antioxidants). In salmon, non-enzymatic brown- the liquid smoke was rich in carbonyl compounds. ing caused by Maillard-type reactions typically Thus, the components of the smoking liquid also occurs at high water activities, while at low water influence the textural quality of smoked foods activities browning is usually caused by the above (Daun 1972, Sink and Hsu 1981). Hassan (1988) protein-lipid derivative interactions. The two types showed that smoking causes a reduction in flesh of reaction can, in fact, inhibit one another. pH, perhaps due to smoke acid absorption, a loss of Toth and Potthast (1984) and Cardinal et al. moisture and the reaction of phenols, polyphenols (2004) indicate colour development to depend and carbonyl compounds with proteins, SH groups mainly on carbonyls, while the flavour in their and amino groups respectively. samples was largely owed to the type and quan- Finally, it should be remembered that the one tity of phenolic compounds present. Martinez et the major problems faced in the marketing and dis- al. (2007a) reported a strong colour intensity in tribution of smoked seafoods is their perishability salmon smoked with liquid smoke rich in carbonyl and hygiene, i.e., contamination by spoilage and compounds. However, Daun (1972) indicate that pathogenic microorganisms. When foods are re- certain high molecular weight phenols in the smoke jected on the grounds of their sensorial properties vapour phase contribute to the colour of the prod- it is reasonable to assume them to be contaminated uct. These phenolic compounds would appear to by both types. Nevertheless, it is no always true. have sufficient hydroxyl groups to be able to enter Sometimes developing of off-odours or off-fla- into lattices (via hydrogen bonding) with collagen vours occurs when hygienic quality is still between at multiple sites. Volatile aldehydes such as conife- acceptable margins (Gomez-Guillén et al. 2009). raldehyde and syringaldehyde give rise to orange Some other times, a product is accepted, attending colours in liquid-smoked fish (Clifford et al. 1980). to sensorial criteria even when spoilage bacterial One of the most important quality characteris- growth is already too high (Jorgensen et al. 2001) tics of fish flesh is its muscle texture. In traditional or the product is not acceptable from a safety per- hot-smoked foods, texture changes are mainly spective (Gram 2001a, 2001b, Dressler 2005). owed to the denaturing of proteins by heat (Gill et The initial bacterial load of fish living in tem- al. 1992), while in liquid-smoked fish it is mainly a perate waters is dominated by Gram negative or- consequence of the action of endogenous proteases ganisms – psycrophilic bacilli of the genera Pseu- hydrolysing the proteins more easily given the lat- domonas, Alteromonas, Moraxella, Acinetobacter, ters’ salt-induced changes in conformation (Hansen Flavobacterium and Vibrio. However, after capture et al. 1996, Lund and Nielsen 2001, Hultmann et and handling, qualitative and quantitative changes al. 2004). in this initial flora can occur, and great variation Some authors report an increase in the firmness in microorganisms and their numbers have been of traditional cold-smoked fish (Sigurgisladottir et reported in salmon from different smoking facili- al. 2000; Birkeland et al. 2004) – an increase that ties. This variability is due to differences in the 222 223 AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method Vol. 20(2011): 217–227.

raw material and processing variables (Hansen et the greatest antimicrobial potential against both al. 1998). Gram negative and positive organisms (Milly et al. Generally, the vacuum packaging of smoked 2005). Martinez et al. (2005) reported that a liquid salmon inhibits the growth of Pseudomonas and smoke flavouring rich in carbonyl compounds had favours the proliferation of Shewanella putre- greater antimicrobial activity than another rich in faciens plus organisms such as Photobacterium phenolic compounds.

phosphoreum, both of which are tolerant to CO2 With respect to pathogenic bacteria, the great- and psychrotrophs (Gram et al. 2002). Lactic acid est risk is posed by the growth of Listeria mono- bacteria and members of Enterobacteriaceae are cytogenes (Gram 2001b, Ward 2001) since it can the largest group spoilage organisms affecting grow at low temperature and in high salt concentra- traditional cold-smoked salmon (Rachmana et al tions. However, good manufacturing practices and 2004, Gram and Huss 1996). adequate storage can eliminate this risk (Autio et al. Authors working with liquid-smoked salmon 1999, González-Rodríguez et al. 2002). Spoilage that was later vacuum packed and stored refriger- organisms can act as a barrier to the entry of pos- ated reported smoke concentrates to have a certain sible pathogenic species (Giménez and Dalgaard antibacterial action. A shelf-life of 25–45 days 2004, Tome et al. 2006), and strains of Carnobac- has been estimated for salmon processed this way terium are being studied for their possible use as (Table 2). Indeed, both the phenolic and carbonyl bioprotectors against Listeria monocytogenes in fractions show antimicrobial properties (Maga smoked salmon (Connil et al. 2002, Brillet 2004, 1987, Painter 1998, Suñén 1998, Suñén et al. 2001, Nilsson et al. 2004, Brillet et al. 2005). Suñén et al. 2003). However, liquid smokes with a low pH and a high carbonyl content are those with

Table 2. Shelf-life of smoked Atlantic salmon (Salmo salar) treated with liquid smoke flavourings.

Authors Smoking method Storage coditions Criteria Shelf life Martinez et al. Liquid smoking: vacuum-packed sensory 30 days (2010) Scansmoke, Broste A/S refrigerated storage (high phenolic content) Bikerland and Skara Liquid smoking (Smokez 5096 vacuum-packed sensory < 31 days (2008) Salmon Smoke) refrigerated storage Martinez et al. Liquid smoking: vacuum-packed sensory a) 32 days in vacuum- (2007a,b) a) Scansmoke, Broste A/S refrigerated storage packed refrigerated (high phenolic content) storage b) AFS-10 SOL, Amcan Ingredientes b) 45 days in vacu- (high carbonyl content) um-packed refrigerated storage Bugueño et al. Liquid smoking: a) vacuum-packed microbial 25 days (2003) 5 BF 4046 refrigerated storage growth (Taberner, S.A., Spain) b) modified atmosphere, refrigerated storage Leroi and Joffraud Cold smoked:a) 2% NaCl (w/w) and vacuum-packed sensory 28 days (2000) 0.8 mg phenols/100g or, refrigerated storage b) 3% NaCl (w/w) and 0.45 mg phenols/100g 222 223 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method

Bannerman A. & Horne J. 2001. Recommendations for the Conclusion production of smoked salmon. Aberdeen, Scotland: Tor- ry Research Station, Ministry of Agriculture. Barat J.M., Rodriguez-Varona S., Andres A. and Fito P. 2002. Influence of increasing brine concentration in Dry salting is the best method for salting salmon to the cod salting process. Journal of Food Science 65: be liquid smoked; this technique reduces economic 1922–1925. and nutritional losses. Dry salting and treatment Bellagha S., Sahli A., Farhat A., Kechaou N. and Glenza A. 2007. Studies on salting and drying of sardine (Sardinel- with liquid smoke flavouring can help improve la aurita): experimental kinetics and modelling. Journal the physiochemical and sensorial characteristics of Food Engineering 78: 947–952. of salmon better than dry or brine salting followed Birkeland S., Skara T., Bjerkeng B. and Rora A.M.B. 2003. Product yield and gaping in cold-smoked Atlan- by traditional smoking. Autolytic deterioration and tic salmon (Salmo salar) fillets as influenced by differ- the growth of spoilage bacteria would appear to be ent injection-salting techniques. Journal of Food Sci- similar in both types of product. The microbiological ence 68:1743–1748. Birkeland S., Rora A.M., Skara T. and Bjerkeng, B. 2004. quality of salmon liquid-smoked with flavourings Effects of cold smoking procedures and raw material rich in carbonyl compounds would appear to better characteristics on product-yield and quality parameters than that achieved with flavourings rich in phenolic of cold smoked Atlantic salmon (Salmo salar L.) fillets. Food Research International 27: 273–286. compound. Birkeland, S. & Bjerkeng B. 2005. The quality of cold- smoked Atlantic salmon (Salmo salar) as affected by salting method, time and temperature. International Journal of Food Science and Technology 40:963–976. Birkeland S. & Skara T. 2008. Cold smoking of Atlantic salmon (Salmo salar) fillets with smoke condensate - References an alternative processing technology for the production of smoked salmon. Journal of Food Science 73: S326–S332. Abu-Ali J.M. & Barringer S. 2007. Colour and texture de- Bocker, U.; Kohler, A.; Aursand, IG and Ofstad, R. 2008. velopment of potato cylinders with liquid smoke during Effects of brine salting with regard to raw material varia- baking, frying and microwaving. Journal of food Process- tion of Atlantic salmon (Salmo salar) muscle investigated ing and Preservation 31: 334–344. by Fourier transform infrared microspectroscopy. Jour- Alfonso M.D. & Sant´Ana. 2008. Effects of pretreatment nal of Agricultural and Food Chemistry 56: 5129–5137. with rosemary (Rosmarinus officinalis L.) in the preven- Bras A. & Costa R. 2010. Influence of brine salting prior to tion of lipid oxidation in salted tilapia fillets. Journal of pickle salting in the manufacturing of various salted-dried Food Quality 31: 586–595. fish species.Journal of Food Engineering 100: 490–495. Andrés A., Rodriguez-Barona S., Barat J.M. and Fito P. Brillet A. 2004. Biodiversity of Listeria monocytogenes sen- 2005. Salted cod manufacturing: influence of salting sitivity to bacteriocin-producing Carnobacterium strains procedure on process yield and product characteristics. and application in sterile cold-smoked salmon. Journal Journal of Food Engineering 69: 467–471. of Applied Microbiology 97: 1029–1037. Aubourg S. P. & Ugliano M. 2002. Effect of brine pre-treat- Brillet A., Pilet M.F., Prevost H., Cardinal M. and Leroi, F. ment on lipid stability of frozen horse mackerel (Trachu- 2005. Effect of inoculation of Carnobacterium divergens rus trachurus). European Food Research and Technol- V41, a biopreservative strain against Listeria monocy- ogy 215: 91–95. togenes risk, on the microbiological, chemical and sen- Aursand I.G., Veliyulen E. and Bocker U. 2009. Water sory quality of cold-smoke salmon. International Jour- and salt distribution in Atlantic salmon (Salmo salar) nal of Food Microbiology 104: 309–324. studied by low-field H-1 NMR, H-1 and Na-23 MRI and Bugueño G., Escriche I., Martinez-Navarrete N., Cama- light microscopy: effects of raw material quality and cho M.D.M. and Chiralt, A. 2003. Influence of storage brine salting. Journal of Agriculture and Food Chem- conditions on some physical and chemical properties istry 57:46–54. of smoked salmon (Salmo salar) processed by vacuum Aursand I.G, Erikson U. and Veliyulin E. 2010. Water prop- impregnation techniques. Food Chemistry 81: 85–90. erties and salt uptake in Atlantic salmon fillets as affect- Cardinal M., Berdagué J.L., Dinel V., Knockaert C. and ed by antemorten stress, rigor mortis and brine salting: Vallet J.J. 1997. Effect of various smoking techniques a low-field H-1 NMR and H-1/Na-23 MRI study. Food on the nature of volatile compounds and on the sen- Chemistry 120: 482–489. sory characteristics of salmon meat. Science des Ali- Autio T., Hielm S., Miettinen M., Sjoberg A.M., Aarnisalo ments 17: 679–696. K., Bjorkroth J., Mattila-Sandholm T. and Korkeala H. Cardinal M., Knockaert C., Torrissen O., Sigurgisladottir S., (1999). Sources of Listeria monocytogenes contamina- Morkore T., Thomassen M. and Vallet J.L. 2001. Rela- tion in a cold-smoked rainbow trout processing plant de- tion of smoking parameters to the yield, color and sen- tected by pulsed-field gel electrophoresis typing.Applied sory quality of smoked Atlantic salmon (Salmo salar). and Environmental Microbiology 65: 150–155. Food Research International 34: 537–550.

224 225 AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method Vol. 20(2011): 217–227.

Cardinal M., Gunnlaugsdottir H., Bjoernevik M., Ouisse A., nal of Applied Microbiology, 96, 96–109. Vallet J.L. and Leroi F. 2004. Sensory characteristics of Gómez-Guillén M.C., Montero P., Hurtado O. and Bor- cold smoked Atlantic salmon (Salmo salar) from Euro- derías, A.J. 2000. Biological characteristics affect the pean market and relationships with chemical, physical quality of farmed Atlantic salmon and smoked muscle. and microbiological measurements. Food Research In- Journal of Food Science 65: 53–60. ternational 37: 181–193. Gómez-Guillén M.C., Gómez-Estaca J., Gimenez B. and Carton I., Bocker U., Ofstad R, Sorheim O and Kohler A. Montero P. 2009. Alternative fish species for cold-smok- 2009. Monitoring secondary structural changes in salting process. International Journal of Food Science and ed and smoked salmon muscle myofiber proteins by FT- Tchnology 44: 1525–1535. IR microspectroscopy. Journal of Agriculture and Food González-Rodríguez M.N., Sanz J.J., Santos J.A., and Chemistry 57: 3563–3570. Garcia, M.L. 2002. Numbers and types of microorgan- Clifford M.N., Tang S.L. and Eyo A.A. 1980. Smoking of isms in vacuum-packed cold-smoked freshwater fish at foods. Process Biochemistry (June/July): 8–11. the retail level. International Journal of Food Microbiol- Connil N., Prevost H. and Dousset X. 2002. Production of ogy 77: 161–168. biogenic amines and divercin V41 in cold smoked salm- Goulas A.E. & Kontominas M.G. 2005. Effect of salting and on inoculated with Carnobacterium divergens V41, and smoking-method on the keeping quality of chub mack- specific detection of this strain by multiplex-PCR. Jour- erel (Scomber japonicus): biochemical and sensory at- nal of Applied Microbiology 92: 611–617. tributes. Food Chemistry 93: 511– 520. Choubert G., Dentella E., Atgie C. and Baccaunaud M. Goulas A.E. & Kontominas M.G. 2007. Combined effect of 2005. Effect of light on colour stability of sliced smoked light salting, modified atmosphere packaging and oreg- rainbow trout (Orcorhynchus mykiss) fed astaxanthin. ano essential oil on the shelf-life of sea bream (Spa- Food Research International 38, 949–952. rus aurata): Biochemical and sensory attributes. Food Czerner M. & Yeannes M.I. 2010. Brining kinetics of differ- Chemistry 100: 287–296. ent cuts of anchovy (Engraulis anchoita). International Gram L.; & Huss H.H. 1996. Microbiological spoilage of fish Journal of Food Science and Technology 45: 2001–2007. and fish products. International Journal of Food Micro- Daun, H. 1972. Sensory properties of phenolic compounds biology 33: 121–37. isolated from curing smoke as influenced by its genera- Gram L. 2001a. Potential hazards in cold-smoked fish: tion parameters. Lebensmittel-Wissenschaft und-Tech- Clostridium botulinum type E. Journal of Food Science nologie 5, 102–5. 66: S1082–S1087. Dondero M., Cisternas F., Carvajal L. and Simpson, R. Gram L. 2001b. Potential hazards in cold-smoked fish: 2004. Changes in quality of vacuum-packed cold- Listeria monocytogenes. Journal of Food Science 66: smoked salmon (Salmo salar) as a function of storage S1072–S1081. temperature. Food Chemistry 87: 543–550. Gram L., Ravn L., Rasch M., Bruhn J. B., Christensen A. Dressler D. 2005. Botulism caused by consumption of B. and Givskov M. 2002. Food spoilage-interactions be- smoked salmon. Nervenart 76: 763–766. tween food spoilage bacteria. International Journal of Frangos L., Pyrgotou N., Giatracou V., Ntzimani A. and Food Microbiology 78: 79–97. Savvaidis I.N. 2010. Combined effects of salting, oreg- Guillén, M.D. & Manzanos, M.J. 1996. Study of the com- ano oil and vacuum-packaging on the shelf-life of re- ponents of a solid smoke flavoring preparation. Food frigerated trout fillets. Food Microbiology 27: 115–121. Chemistry, 55, 251–7. Gallart-Jornet L., Rustad T., Barat J.M., Fito P. and Escriche Guillén M.D., Ruiz A. and Cabo N. 2004. Study of the ox- I. 2007a. Effect of superchilled storage on the freshness idative degradation of farmed salmon lipids by means and salting behaviour of Atlantic salmon (Salmo salar) of Fourier transform infrared spectroscopy. Influence fillets.Food Chemistry 103: 1268–1281. of salting. Journal of the Science of Food and Agricul- Gallart-Jornet L., Barat J.M., Rustad T., Erikson U. and ture 84: 1528–1534. Escriche I. 2007b. Influence of brine concentration of Guillén M.D., Carton I., Casas C. and Salmeron J. 2005. Atlantic salmon fillet salting.Journal of Food Engineer- Changes in the headspace of farmed salmon by salting, ing, 80: 267–275. smoking and both combined processes. Proceedings of Gallart-Jornet L., Barat J.M, Rustad T., Erikson U., Escriche Intrafood 2005. Valencia. pp. 105–108. I. and Fito P. 2007c. A comparative study of brine salt- Hansen L.T., Gill T., Røentved S.D. and Huss H.H. 1996. ing of Atlantic cod (Gadus morhua) and Atlantic salmon Importance of autolysis and microbiological activity on (Salmo salar). Journal of Food Engineering 79: 261–270. quality of cold-smoked salmon. Food Research Inter- Gedela S., Gamble R.K., Macwana S., Escoubas J.R. and national 29: 181–188. Muriana P.M. 2007. Effect of inhibitory extracts derived Hansen L.T., Røntved S. and Hans H.H. 1998. Microbio- from liquid smoke combined with postprocess pasteur- logical quality and shelf life of cold-smoked salmon ization for control of Listeria monocytogenes on ready- from three different processing plants. Food Microbiol- to-eat meats. Journal of Food Protection 70: 2749–2756. ogy 15: 137–150. Gill T.A., Chan J.K., Phonchareon K.F. and Paulson A.T. Hassan I.M. 1988. Processing of smoked common carp 1992. Effect of salt concentration and temperature on fish and its relation to some chemical, physical and or- heat-induced aggregation and gelation of fish myosin. ganoleptic properties. Food Chemistry 27: 95–106. Food Research International 25: 333–341. Hatae K.., Sakamoto H., Shimada A., Matsumoto M., Ya- Giménez B. & Dalgaard P. 2004. Modelling and predicting manaka H. and Watabe S. 1990. Physical-properties the simultaneous growth of Listeria monocytogenes and of arai made by applying of thaw-rigor. Nippon Suisan spoilage micro-organisms in cold-smoked salmon. Jour- Gakkaishi 56: 2113 –2118. 224 225 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method

Huang Y., Cavinato A.G., Mayes D.M., Bledsoe G.E. and processes during storage. Food Science and Technol- Rasco B. A. 2002. Nondestructive prediction of mois- ogy International 13: 477–484. ture and sodium chloride in cold smoked Atlantic salmon Martinez O., Salmeron J., Guillen M.D. and Casas C. 2009. (Salmo salar). Journal of Food Science 67: 2543–2547. Textural and physicochemical characteristics of salm- Hultmann L. & Rustad T. 2004. Iced storage of Atlantic on (Salmo salar) treated by different smoking process- salmon (Salmo salar) - effects on endogenous enzymes es during storage. Alimentaria 399: 92–97. and their impact on muscle proteins and texture. Food Martinez O., Salmeron J., Guillen M.D. and Casas C. Chemistry 87: 31–41. 2010. Effect of freezing on the physicochemical, textur- Hultmann L., Røra A.M.B., Steinsland I., Skara T. and Rus- al and sensorial characteristics of salmon (Salmo salar) tad T. 2004. Proteolytic activity and properties of pro- smoked with a liquid smoke flavouring. LWT-Food Sci- teins in smoked salmon (Salmo salar) - effects of smok- ence and Technology, 43: 910–918. ing temperature. Food Chemistry 85: 377–387. Milly P.J., Toledo R.T. and Ramakrishnan S. 2005. Deter- Johnston J.A., Li X., Vieira V.L.A., Nickell D., Dingwall A., mination of minimum inhibitory concentrations of liquid Alderson R., Campbell P. and Bickerdike R. 2006. Mus- smoke fractions. Journal of Food Science 70: 12–17. cle and flesh quality traits in wild and farmed Atlantic Minh V.N, Thorarinsdottir K.A, Gudmundsdottir A., Thor- salmon. Aquaculture 256: 323–336. kelsson G. and Arason S. 2011. The effect of salt con- Jorgensen L.V. Huss H.H. and Dalgaard D. 2001. Signifi- centration on conformational changes in cod (Gadus cance of volatile compound produced by spoilage bacte- morhua) proteins during brine salting. Food Chemistry ria in vacuum-packed cold-smoked salmon (Salmo salar) 125: 1013–1019. analyzed by GC-MS and multivariate regression. Jour- Montero P., Gómez-Guillén M.C. and Borderias A.J. 2003. nal of Agricultural and Food Chemistry 49: 2376–2381. Influence of salmon provenance and smoking process Kashiri H., Haghparast S. and Shabanpour B. 2011. Effects on muscle functional characteristics. Journal of Food of sodium salt solutions (sodium acetate, lactate and cit- Science 68: 1155 –1160. rate) on physicochemical and sensory characteristics of Morkore T., Ruohonen K., and Kiessling A. 2009. Varia- Persian sturgeon (Acipenser persicus) fillets under re- tion in texture of farmed Atlantic salmon (Salmo salar L.). frigerated storage. Journal of Agriculture, Science and Relevance of muscle fiber cross-sectional area.Journal Technology 13: 89–98. of Texture Studies 40: 1–15. Kristinsson H.G., Crynen S. and Yagiz Y. 2008. Effect of Muratore G., Mazzaglia A., Lanza C.M. and Licciardel- filtered wood smoke treatment compared to various gas lo F. 2007. Effects of process variable on the quality of treatments on aerobic bacteria in yellowfin tuna steaks. swordfish fillets flavoured with smoke condensate.Jour - LWT-Food Science and Technology 41: 963–976. nal of Food Processing and Preservation 31: 167–177. Larsen R., Stormo S.K., Dragnes B.T. and Elvevoll E. 2007. Nilsson L., Ng Y.Y., Christiansen J.N., Jorgensen B.L., Losses of taurine, creatine, glycine and alanine from cod Grotinum D. and Gram L. 2004. The contribution of bac- (Gadus morhua) fillet during processing.Journal of Food teriocin to inhibition of Listeria monocytogenes by Car- Composition and Analysis 20: 396–402. nobacterium piscicola strains in cold-smoked salmon Larsen R., Olsen S.H., Kristoffersen S. and Elvevoll, E.O. systems. Journal of Applied Microbiology 96: 133–143. 2008. Low salt brining of pre-rigor filleted farmed cod Oehlenschlager, J. 2007. Water, fat and salt content in (Gadus morhua L.) and the effect on different quality commercially traded 200g consumer packages of cold parameters. LWT-Food Science and Technology 41: smoked salmon (Salmo salar). Deutsche Lebensmittel- 1167–1172. rundschau 103: 569–573. Leroi F. 2010. Occurrence and role of lactic acid bacteria Painter T.J. 1998. Carbohydrate polymers in food preser- in seafood products. Food Microbiology 27: 698–709. vation: an integrated view of the Maillard reaction with Leroi F. & Joffraud J.J. 2000. Salt ad smoke simultaneous- special reference to discoveries of preserved foods in ly affect chemical and sensory quality of cold-smoked Sphagnum-dominated peat bogs. Carbohydrate Poly- salmon during 5ºC storage predicted using factorial de- mers 36: 335–347. sign. Journal of Food Protection 63: 1222–1227. Pszczola D.E. 1995. Tour highlights production and uses Lund K.E. & Nielsen H.H. 2001. Proteolysis in salmon of smoke-based flavours. Food Technology 49: 70–74. (Salmo salar) during cold storage; effects of storage Rachmana C., Fourriera A., Sya A., De La Cochetiereb time and smoking process. Journal of Food Biochem- M.F., Prevosta H. and Xavier D. 2004. Monitoring of istry 25: 379–395. bacterial evolution and molecular identification of lac- Maga J. A. 1987. The flavour chemistry of wood smoke. tic acid bacteria in smoked salmon during storage. Le Food Review International 3, 139–183. Lait 84 : 145–154. Martinez O., Casas C., Guillen M.D. and Salmeron J. 2005. Rongrong L., Carpenter J.A. and Cheney R. 1998. Sensory Microbiological changes of salmon treated with two liq- and instrumental properties of smoked sausage made uid smoke flavourings. 2nd International Congress of with mechanically separated poultry (MSP) meat and Self-Control and Food Safety, Bilbao, Spain. pp. 193. wheat protein. Journal of Food Science 63: 923–929. Martinez O., Salmeron J., Guillen M.D. and Casas C. Rora A.M., Furukang R., Fjaera S. and Skjervold, P. 2004. 2007a. Textural and physicochemical changes in salm- Salt diffusion in pre-rigor filleted Atlantic salmon.Aqua - on (Salmo salar) treated with commercial liquid smoke culture 232: 201–211. flavourings.Food Chemistry 100: 498–503. Rora A.M.B., Birkeland S., Hultmann L., Rustad T., Ska- Martinez O., Salmeron J., Guillen M.D. and Casas C. ra T. and Bjerkeng, B. 2005. Quality characteristics of 2007b. Sensorial and physicochemical characteristics farmed Atlantic salmon (Salmo salar) fed diets high in of salmon (Salmo salar) treated by different smoking soybean or fish oil as affected by cold-smoking tempera- 226 227 AGRICULTURAL AND FOOD SCIENCE

Martinez, O. et al. Liquid smoked salmon as affected by salting method Vol. 20(2011): 217–227.

ture. Food Science and Technology / Lebensmittel-Wis- bergsson K. 2004. The effect of various salt concentra- senschaft und-Technologie 38: 201–211. trions during brine curing of cod. International Journal Ruiter A. (1979). Color of smoked foods. Food Technol- of Food Science and Technology 39: 79–89. ogy 33: 54–63. Tome E., Teixeira P. and Gibbs P. A. 2006. Anti-listerial Sannaveerappa T., Ammu K. and Joseph J. 2004. Pro- inhibitory lactic acid bacteria isolated from commercial tein related changes during salting of milkfish Chanos( cold smoked salmon. Food Microbiology 23: 399–405. chanos). Journal of Science and Food Agriculture 84: Toth L. & Potthast K. 1984. Chemical aspects of the smok- 863–869. ing of meat and meat products. Advances in Food Re- Serot T. & Lafficher C. 2003. Optimisation of solid-phase search 29: 87–158. microextraction coupled to gas chromatography for de- Underwood G. & Shoop J. 2007. “Raw materials for fla- termination of phenolic compounds in smoked herring. vourings: smoke flavourings. Nature, preparation and Food Chemistry 82: 513–519. application”, in: flavourings production, composition, Sérot T., Baron R., Knockaert C. and Vallet J. L. 2004. Ef- application and regulations (2nd edition). Ziegler. Wiley- fect of smoking processes on the contents of 10 major VCH Verlag GmbH & Co. KGaA, Weinheim, Germany. phenolic compounds in smoked fillets of herring (Cuplea p. 289–308. harengus). Food Chemistry 85: 111–120. Unlusayin M., Erdilal R., Gumus B. and Gulyavuz H. 2010. Sigurgisladottir S., Sigurdardottir M.S., Torrissen O., Vallet The effects of different salting methods on extract loss J.L. and Hafsteinsson H. 2000. Effects of different salt- from rainbow trout. Pakistan Veterinary Journal 30: ing and smoking processes on the microstructure, the 131–134. texture and yield of Atlantic salmon (Salmo salar) fillets. Varlet V., Knockaert C., Prost C. and Serot T. 2006. Com- Food Research International 33: 847–855. parison of Odor-Active Volatile Compounds of Fresh Sink J.D. & Hsu, L.A. (1977). Chemical effect of smoke- and Smoked Salmon. Journal of Agricultural and Food processing on frankfurter manufacture and storage char- Chemistry 54: 3391–3401. acteristics. Journal of Food Science 42: 1489. Varlet V., Prost C. and Serot, T. 2007a. New procedure Sirkoski Z., Haard N., Motohiro T. and Sin Pan, B. 1998. for the study of odour representativeness of aromat- Fish drying and smoking. Production and quality. In: ic extracts from smoked salmon. Food Chemistry 100: Quality. J. Doe (Ed.). CRC Press. Miami. pp. 250. 820–829. Smith G. & Hole M. 1991. Browning of salted sun-dried Varlet V., Prost C. and Serot, T. 2007b. Volatile aldehydes fish. Journal of the Science of Food and Agriculture in smoked fish: Analysis methods, occurrence and mech- 55: 291–301. anisms of formation. Food Chemistry 105: 1536–1556. Suñén E. 1998. Minimum inhibitory concentration of smoke Varlet V., Serot T., Knockaert C., Cornet J., Cardinal M., wood extracts against spoilage and pathogenic micro- Monteau F., Le Bizec B. and Prost C. 2007c. Organolep- organisms associated with foods. Letters in Applied tic characterization and PAH content of salmon (Salmo Microbiology 27: 45–48. salar) fillets smoked according to four industrial smok- Suñén E., Fernández-Galian B. and Aristimuno C. 2001. ing techniques. Journal of the Science of Food and Ag- Antibacterial activity of smoke wood condensates riculture 87: 847–854. against Aeromonas hydrophila, Yersinia enterocolitica Varlet V., Serot T., Cardinal M., Knockaert C. and Prost C. and Listeria monocytogenes at low temperature. Food 2007d. Olfactometric determination of the most potent Microbiology 18: 387–393. odor-active compounds in salmon muscle (Salmo sal- Suñén E., Aristimuño C. and Fernández-Galian B. 2003. ar) smoked by using four smoke generation techniques. Activity of smoke wood condensates against Aerom- Journal of Agricultural and Food Chemistry 55: 4518–25. onas hydrophila and Listeria monocytogenes in vacu- Ward D.R. 2001. Description of the situation. Journal of um-packaged, cold-smoked rainbow trout stored at 4ºC. Food Science 66: 1067–1071. Food Research International 36: 111–116. Webb J., Verspoor E., Aubin-Horth N., Romakkaniemi A. Thiansilakul Y., Benjakul S. and Richards M.P. 2010. and Amiro, P. 2007. The Atlantic Salmon: genetics, con- Changes in heme proteins and lipids associated with servation and management. In: The Atlantic salmon. Ver- off-odour or seabass (Lates calcarifer) and red tilapia spoor, Stradmeyer & Nielsen (Eds). Blackwell Publish- (Oreochromis mossambicus x o. niloticus) during iced ing. London. pp. 17–56. storage. Food Chemistry 121: 1109 –1119. Yanar Y., Celik M. and Akamca E. 2006. Effects of brine Thorarinsdottir K.A., Arason S., Bogason S. and Krist- concentration on shelf-life of hot-smoked tilapia (Ore- ochromis niloticus) stored at 4°C. Food Chemistry 97: 244–247.

226 227 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Rajala, A. Vol.et al. 20(2011): Seed quality 228–234. effects in barley

Seed quality effects on seedling emergence, plant stand establishment and grain yield in two-row barley

Ari Rajala*1, Markku Niskanen1, Mika Isolahti2 and Pirjo Peltonen-Sainio1 1MTT Agrifood Research Finland, Plant Production Research, FI-31600 Jokioinen, Finland 2Boreal Plant Breeding Ltd. FI-31600 Jokioinen, Finland *e-mail: [email protected]

Seed viability and vigour play important roles in seedling emergence, plant stand establishment and yield potential. The majority of cereal fields in Finland are typically sown with farm saved seed (FSS). If the quality of the seed is not known, there can be insidious yield reduction. This research was conducted to study the effects of seed quality on seedling emergence rate, seedling number and yielding capacity. The study comprised three-year field experiments conducted during 2007–2009, established at three sites: Jokioinen, Nousiainen and Ylistaro. Spring barley cultivars Saana (2007) and Annabell (2008-2009) were sown at rate of 500 germinating seeds m-2. Five seed lots were included as treatments: farm saved seed (FSS); down-

graded seed <2.5 mm (FSS<2.5 mm); upgraded seed >2.7 mm (FSS>2.7 mm); upgraded seed >2.7 mm with disinfection

(FSS>2.7 mm + dis); and commercial certified seed with disinfection (CCS). Up- and down-graded seed lots

(FSS<2.5 mm, FSS>2.7 mm, and FSS>2.7 mm + dis) all originated from the FSS. Seedling emergence rate was measured from the time when coleoptiles started to break through the soil surface. The number of seedlings (3 × 1 m row per plot) was recorded at five-day intervals four times from the same rows. Plots were harvested at physiological maturity and grain yield (kg ha-1), hectolitre weight (HLW, kg) single grain weight (SGW,

mg) and grain protein content (%) were recorded. Seed lots of CCS and FSS>2.7 mm + dis enhanced seedling emergence rate and increased the number of plants compared with other treatments. These two seed lots also produced the highest grain yield and had the lowest grain protein. Seed quality had an apparent effect on plant stand establishment and grain yield. A seed lot effect was evident despite identical targeted sowing rates that took into account germination rate and seed weight. Therefore, differences in seedling emergence and yielding capacity were likely outcomes of variation in seed vigour among the five treatments.

Key-words: barley, emergence, plant stand, seed quality, yield

Rajala, A. Vol.et al. 20(2011): Seed quality 228–234. effects in barley Vol. 20(2011): 228–234.

Introduction The shoot apex initially produces the leaf and tiller primordia and after a transitional phase, the spikelet and floral organs (Bonnett 1961; Wadding- Seed viability and vigour determine how success- ton et al. 1983). Thus, yield formation actually be- fully seedlings emerge, plant stand is established gins already at seedling emergence. During a short and yield potential is built up. Seed viability gener- intensive growing season, as prevails in northern ally refers to germination rate whereas seed vigour Europe, cereal growth and development are rapid represents the capacity of the seed to develop and due to the long days. This secures successful crop produce a normal seedling under various growing production in just 80 to 110 days (Peltonen-Sainio conditions. These two traits are affected by several and Rajala 2007; Peltonen-Sainio et al. 2009a). Ac- factors, starting from grain-filling period and con- celerated development rate reduces a plant’s poten- tinuing throughout harvest and storage conditions. tial to compensate for potential setbacks occurring Seed age, moisture content, infection by seed-borne during emergence and early plant stand formation pathogens and seed weight are among the charac- (Peltonen-Sainio et al. 2009a). This emphasizes the teristics associated with seed viability and vigour central role of seed quality and successful sowing (Doling 1968, Ching et al. 1977, Khah et al. 1989, to facilitate even and adequately dense seedling Gan et al. 1992, Chastain et al. 1995, Grilli et al. emergence. Successful seedling emergence sets the 1995, Reddy et al. 1999, Nonogaki et al. 2010). Dif- foundation for the formation of full (= adequately ferences between germination rates in the laboratory dense) plant stands that utilise available water and and seedling emergence rates in the field may vary radiation efficiently from the start of the intensive considerably: seed lots with a germination rate of growing period (Peltonen-Sainio et al. 2003; Pel- 85–100% produced seedlings at rates of 47 to 93% tonen-Sainio et al. 2009a). according to Steiner et al. (1989). This difference Depending on a farmer’s preferences, either between germination rate and emerged seedling FSS or CCS are used to sow a crop. Unlike in other number was attributed to differences in seed vigour European countries, in Finland most cereal fields under varying conditions (Steiner et al. 1989). Ac- are typically sown with FSS (Peltonen-Sainio et cordingly, seed performance is a consequence of al. 2003). The proportion of FSS varies among ce- an interaction between seed vigour and prevailing real species and years, ranging from 63% to 80% conditions such as sowing depth, soil characteristics, (Fig. 1). The quality control for FSS depends on the soil moisture and temperature. farmer running the required tests. For the certified

% 40

0 Barley Oats Wheat Fig. 1. The certified seed percent- age (%) in spring cereal sowings. 2007 2008 2009 Source: Nordman (2010). 228 229 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Rajala, A. et al. Seed quality effects in barley seed sector, the seed quality is regulated by em- ypry –compounds (trademark Ariane S). Five seed ploying a threshold value for seed germination and lots (treatments) were compared: FSS; FSS<2.5 mm; other seed quality requirements, which are set by FSS>2.7 mm; FSS>2.7 mm + dis (disinfection by using legislative authorities (http://www.evira.fi). Com- triconazole and imazalil -compounds, trademark pared with viability (=germination rate), seed vig- Robust); and CCS. FSS<2.5 mm, FSS>2.7 mm, and our is more laborious to measure and define and it FSS>2.7 mm + dis originated from the same FSS mate- is not included in regular seed testing and approval rial, which was produced by applying good farm programmes. However, seed vigour may strongly management practices. Seed lot FSS<2.5 mm consist- affect seedling emergence rate, plant stand estab- ed of seeds passed through a 2.5 mm sieve plate, lishment and yield performance, especially under whereas for the FSS>2.7 mm, and FSS>2.7 mm + dis lots less favourable growing conditions (Steiner et al. the seeds remaining on the 2.7 mm sieve plate were 1989, López-Castaňeda et al. 1996). This research used. Germination rate and single grain weight of was conducted to study the potential differences seed lots are shown in Table 1. in seedling emergence rate, seedling number and Seedling emergence rate was assessed at the yielding capacity of seed lots with similar germina- Jokioinen site (2008–2009) when coleoptiles began tion rates but varying in other quality traits. Special to emerge through the soil surface. At first count- interest was to reveal potential differences in yielding, three rows per plot were marked (length 1 ing capacity and ranking of seed lots when grown m). Number of seedlings was counted at five-day in varying growing conditions. intervals four times at the same marked seeding rows. At physiological maturity, plots were harvested and grain yield (kg ha-1 at 15% moisture content) hectolitre weight (HLW, kg) and single grain weight (SGW, mg) were measured. Grain N Material and methods concentration was measured with a NIR analyzer (Foss InfraXact). This study comprised three-year field experiments Trial set-up was a randomised block design conducted during 2007–2009 established at three with four replicates. All statistical analyses were sites: Jokioinen (60° 48′ 3 N, 23° 28′ 5 E), Nou- carried out with PASW Statistics 18.0 software us- siainen (60° 35′ 6 N, 22° 50′ 0 E) and Ylistaro ing the MIXED procedure (Littell et al., 1996). In (62° 56′ 3 N, 22° 31′ 0 E). Spring barley cultivars the model, seed lot, trial site and year were consid- Saana (2007) and Annabell (2008–2009) were ered to be fixed effects, while blocks nested within sown at 500 germinating seeds m-2. Sowing rate trial and year were considered to be random when was calculated using following equation computing means for grain yield and yield quality parameters. By considering seed lot, trial site aimed density (plants m-2) × thousand grain weight (g) -1 ______and year as a fixed effect, the model was more ap- kg ha = germination rate (%) propriate to reveal potential seed lot by growing Thus all plots were sown with an identical condition interaction, i.e. seed lot differences in number of viable seeds (500 seeds m-2) and thereby yielding capacity when grown in varying growing the potential seed lot effect on growth was not a conditions (site and year). Comparisons of treat- consequence of different germination rates. ments were done using the Bonferroni test. Plot size was 6 m2 in Jokioinen and Nousiainen Seed lot was considered to be a fixed effect, and 10 m2 in Ylistaro. Plots were fertilized with while year and block were considered to be random NPK- (23-3-6) fertilizer at 80 kg N ha-1. Weeds when computing means for seedling numbers of cv were controlled using MCPA, clopyralid, fluorox- Annabell for trials carried out at Jokioinen.

230 231 AGRICULTURAL AND FOOD SCIENCE

Rajala, A. et al. Seed quality effects in barley Vol. 20(2011): 228–234.

Table 1. Single grain weight (SGW) and germination rate (%) of seed categories. +, disinfectant treatment; - no disinfectant treatment. Year Cultivar Seed lot Seed lot SGW Germination Disinfectant abbreviation mg % 2007 Saana farm saved seed (FSS) FSS 38.8 96 -

2007 Saana downgraded FFS FSS<2.5 mm 30.7 96 -

2007 Saana upgraded FSS FSS>2.7 mm 48.0 96 -

2007 Saana upgraded + disinfected FFS FSS>2.7 mm + dis 48.0 96 + 2007 Saana commercial certified seed CCS 51.2 94 + 2008 Annabell farm saved seed (FSS) FSS 41.0 89 -

2008 Annabell downgraded FFS FSS<2.5 mm 32.7 89 -

2008 Annabell upgraded FSS FSS>2.7 mm 45.3 89 -

2008 Annabell upgraded + disinfected FFS FSS>2.7 mm + dis 45.3 89 + 2008 Annabell commercial certified seed CCS 50.7 92 + 2009 Annabell farm saved seed (FSS) FSS 41.7 91 -

2009 Annabell downgraded FFS FSS<2.5 mm 37.2 91 -

2009 Annabell upgraded FSS FSS>2.7 mm 43.6 91 -

2009 Annabell upgraded + disinfected FFS FSS>2.7 mm + dis 43.6 91 + 2009 Annabell commercial certified seed CCS 50.7 92 + CCS = commercial certified seed; FSS = farm saved seed.

our (Peltonen-Sainio et al. 2009a, Peltonen-Sainio et Results and discussion al. 2010). Hence, actual seedling number may differ markedly from potential seedling emergence based

CCS and FSS>2.7 mm + dis seed lots enhanced seedling solely on germination rate (Fig. 2). This was evident emergence rate and increased the number of plants in this study as the number of emerged seedlings compared with the other seed lots at trials carried did not reach the target 500 seedlings m-2 for any of out in Jokioinen (Fig. 2). Improved emergence the seed lots. The range of emerged seedlings was was evident despite similar numbers of germinat- from 68% to 86%, which is comparable with the ing seeds per m2. Apparently, seed treatment with values presented by Steiner et al. (1989). Barley has fungicide improved seed’s potential to produce live the highest tillering capacity of the cereal species and vigour seedlings to emerge through the soil. grown in northern growing conditions (Peltonen- Similar response has been reported in an extensive Sainio et al. 2009b). Despite of this, sparse and/ Swedish study carried out on several locations and or uneven seedling emergence may result in yield years (Johnsson et al. 1998). Under field condi- reductions in a short and intensive growing period tions, also seed vigour is known to affect seedling (Peltonen-Sainio et al. 2009a,b). emergence (Khah et al. 1989, Steiner et al. 1989, Seed lots varied in yielding capacity (Table 2). López-Castaňeda et al. 1996). Post-sowing drought The grain yield varied within trial sites and be- periods occur regularly in the main cereal produc- tween years. However, the ranking of seed lots was tion areas of Finland (Pajula and Triipponen 2003, similar over the years and sites and no significant Peltonen-Sainio et al. 2010). Sparse precipitation interaction between seed lots, trial sites and years in conjunction with heavy soils (typically clay in was recorded. CCS produced the highest grain

southern Finland) set high requirements to seed vig- yield, followed by seed lot FSS>2.7 mm + dis. These 230 231 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Rajala, A. et al. Seed quality effects in barley

2 seedlings/m 500

a a 0.86 a a a 0.81 400 a b b b 0.72 b 0.69 300 b b Fig. 2. Number of seedlings per square meter after different seed categories of cv Annabell. Counted 200 at Jokioinen trial site 2008–2009. Countings were carried out at 5 days intervals. Numbers in the box indicate the ratio of realised and 100 0 1 2 3 4 5 aimed seedling number. The same counting lowercase letter indicates no statistical difference at p < 0.05. CCS = CCS FSS<2.5 mm commercial certified seed; FSS = FSS>2.7 mm + dis FSS farm saved seed. FSS>2.7 mm

Table 2. Seed lot effect on grain yield (kg ha-1), protein concentration (%), hecto liter weight (HLW, kg) and single grain weight (SGW, mg). Protein is shown at 0% moisture content, other results at 15%. Results are shown as means over year, trial site and genotype. The same lowercase letter indicates no statistical difference at p < 0.05 in the test of Bonferroni. Yield Protein HLW SGW kg ha-1 % kg mg FSS 5692 bc 11.6 a 67.3 a 44.8 a

FSS<2.5 mm 5506 c 11.6 a 67.3 a 44.9 a

FSS>2.7 mm 5820 b 11.6 a 67.2 a 45.2 a

FSS>2.7 mm + dis 6126 a 11.3 b 67.3 a 44.9 a CCS 6346 a 11.3 b 67.4 a 45.3 a p-value < 0.001 < 0.001 0.948 0.612 sem 80.1 0.06 0.18 0.29 CCS = commercial certified seed; FSS = farm saved seed; sem = standard error of means.

two seed lots expressed also higher seedling emer- yield improvement is comparable with the results gence rate when counted at Jokioinen trial site. Im- presented previously in other Nordic studies (Tah- proved seedling emergence has likely a positive vonen et al. 1995, Johnsson et al. 1998). CCS and effect on yield performance (Steiner et al. 1989, FSS>2.7 mm + dis attained a slightly lower grain pro- López-Castaňeda et al. 1996). Seed restoration tein concentration (%) as compared with other actions (up-grading and seed disinfection) of FSS seed lots (Table 2). In contrast with the general improved yield performance. When comparing protein requirements for feed and food, low pro- seed lots FSS>2.7 mm and FSS>2.7 mm + dis, seed disin- tein content is required in malting barley (Pelto- fection increased grain yield by 300 kg ha-1. This nen-Sainio et al. 2003). Yield and grain protein 232 233 AGRICULTURAL AND FOOD SCIENCE

Rajala, A. et al. Seed quality effects in barley Vol. 20(2011): 228–234.

are typically negatively correlated (Boonchoo et ment needs to exceed the seed cost difference al. 1998), which was the case also in this study. with unrestored FSS. The relatively strong fluc- Hence, any increase in grain yield using the same tuations in cereal market price during recent years available growth resources (namely nitrogen) will somewhat complicate the economic comparison likely result in reduction in grain protein concen- between seed lots, and consequently a farmer’s tration. Even moderate reduction in grain protein decision making. In our estimations approxi- may determine whether yield ends up for feed or mately 150 to 250 kg ha-1 more grain is needed malting. This translates into a substantial difference to cover the difference in cost between CCS and -1 in the value of the yield as malting barley attracts FSS>2.7 mm + dis, while 300 to 500 kg ha covers the a higher price than feed barley. Despite of lower difference between CCS and unrestored FSS, and -1 grain protein concentration in CCS and FSS>2.7 mm + 100 to 200 kg ha covers the difference between dis seed lots, these two produced the highest protein FSS>2.7 mm + dis and FSS. These rough threshold val- yield (kg protein per hectare). This means more ues were met in this study (Table 2). The economic efficient use of nitrogen to build up grain yield. assessment of the situation changes even more to- Other quality traits, like single grain weight (mg) wards a high quality seed advantage when other and hectolitre weight (kg), were not affected by issues are taken into account. If other inputs (cul- seed lot (Table 2). tivation, fertilisation, plant protection) are used at a The seed weight effect was also emphasised in similar level, then the low quality seed with lower this experiment as both up- and down-grading of yielding potential inevitably reduces the input use-

yielding capacity. FSS<2.5 mm was associated with efficiency. This represents a drawback in economic a grain yield reduction of ca. 200 kg ha-1 when as well as environmental terms. compared with FSS, ca. 300 kg ha-1 when com- -1 pared with FSS>2.7 mm and of ca. 600 kg ha when compared with FSS>2.7 mm + dis (Table 2). The seed weight effect on yielding potential is ambiguous, but some literature reports a positive correlation Conclusions between them (Mikkelsen 1963, Hampton 1981, Chastain et al. 1995). Accordingly, grading of the In this study, seed quality had an apparent effect seed lot seems to improve yield formation. FSS is on plant stand establishment and yield. Differences often heterogeneous regarding seed weight (Pelto- were obtained despite similar germination rates nen-Sainio et al. 2011) and therefore grading is an for the various seed lots and identical seed rates important procedure for improving and stabilising of germinable seeds (500 germinating seeds m-2). the seed quality of FSS. Thus, differences in seedling emergence and yield FSS is used extensively by Finnish farmers performance were the outcome of variation in seed (Fig. 1). A common reason for this is the low price quality. Higher yield improves input use-efficiency, paid for grain and farmers are apt to reduce pro- which is an important aspect of economical results duction costs by using FSS. However, according and environmental impact of cereal production. to our results, CCS produced the highest grain yield. Also, good quality FSS, when it is graded Acknowledgements. The authors want to acknowl- and treated with disinfectant, can be comparable edge the technical staff of Boreal Plant Breeding to the quality of CCS, but these restoration actions Ltd and MTT Plant Production Research, who also increase the costs of FSS. To justify the use participated in organizing the trials. The authors

of more expensive seed (CCS or FSS>2.7 mm + dis) are grateful to Kylvösiemensäätiö (foundation) for merely for economic reasons, the yield improve- financing the research.

232 233 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Rajala, A. et al. Seed quality effects in barley

ings Tolvte Kongress, Helsingfors 1963. p. 154–155 References (in Norwegian). Nonogaki, H., Bassel, G.W. & Bewley, J.D. 2010. Germina- tion – Still a mystery. Plant Science 179: 574–581. Bonnett, O.T. 1961. The oat plant: Its histology and devel- Nordman, R. 2010. Kevätviljoista kolmannes sertifioidulla. opment. Bulletin 672. University of Illinois. Agricultural Kylvösiemen 49:29. (in Finnish) Experiment Station. 112 p. Pajula, H. & Triipponen, J.P. 2003. Selvitys Suomen Boonchoo, S., Fukai, S. & Hetherington, S. E. 1998. Bar- kastelutilanteesta. Esimerkkialueena Varsinais-Suomi. ley yield and grain protein concentration as affected by Suomen Ympäristö 629. 86 p. (In Finnish). assimilate and nitrogen availability. Australian Journal Peltonen-Sainio, P., Rajala, A. & , Kontturi, M. 2003. Devel- of Agricultural Research 49: 695–704. opment and growth of barley stands determining malting Chastain, T.G., Ward, K.J. & Wysocki, D.J. 1995. Stand quality: copying growing conditions with crop manage- establishment responses of soft white winter wheat ment. In: Proceedings of the 29th EBC Congress Dub- to seedbed residue and seed size. Crop Science 35: lin 2003. p. 116–124. 213–218. Peltonen-Sainio, P. & Rajala, A. 2007. Duration of vege- Ching, T.M., Hedtke, S., Boulger, M.C. & Kronstad, W.E. tative and generative development phases in oat culti- 1977. Correlation of field emergence rate and seed vigor vars released since 1921. Field Crops Research 101: criteria in barley cultivars. Crop Science 17: 312–314. 72–79. Doling, D.A. 1968. Effects of infection with Ustilago nuda Peltonen-Sainio, P., Rajala, A., Känkänen, H., Hakala, K. and of seed size on the vigour of barley plants. Transac- 2009a. Improving farming systems in northern European tions British Mycological Society 51: 179–183. conditions. In: Editors Victor O. Sadras, Daniel Calderini. Gan, Y., Stobbe, E.H. & Moes, J. 1992. Relative date of Crop Physiology : Applications for Genetic Improvement wheat seedling emergence and its impact on grain yield. and Agronomy. Amsterdam: Elsevier. p. 71–97. Crop Science 32: 1275-1281. Peltonen-Sainio, P., Jauhiainen, L., Rajala, A., Muurinen, Grilli, I., Bacci, E., Lombardi, T., Spano, C. & Floris, C. S. 2009b. Tiller traits of spring cereals under tiller-de- 1995. Natural ageing: Poly(A) polymerase in germinat- pressing long day conditions. Field Crops Research ing embryos of Triticum durum wheat. Annals of Bota- 113: 82-89. ny 76: 15–21. Peltonen-Sainio, P., Jauhiainen, L. & Hakala, K. 2010. Crop Hampton, J.G. 1981. The extent and significance of seed responses to temperature and precipitation according to size variation in New Zealand wheats. New Zealand long-term multi-location trials at high latitude conditions. Journal of Experimental Agriculture 9: 179–183. Journal of Agricultural Science 149: 49-62 Johnsson, L., Hökeberg, M. & Gerhardson, B. 1998. Per- Peltonen-Sainio, P., Rajala, A. & Jauhiainen, L. 2011. Hid- formance of the Pseudomonas chlororaphis biocontrol den viability risks in use of farm saved barley and wheat agent MA 342 against cereal seed-borne diseases in seed? Journal of Agricultural Science, DOI:10.1017/ field experiments. European Journal of Plant Patholo- S0021859611000335 gy 104: 701–711. Reddy, M.V.B., Arul, J., Angers, P. & Couture, L. 1999. Chi- Khah, E.M., Roberts, E.H. & Ellis, R.H. 1989. Effects of tosan treatment of wheat seed induces resistance to seed ageing on growth and yield of spring wheat at dif- Fusarium graminearum and improves seed quality. Jour- ferent plant-population densities. Field Crops Research nal of Agriculture and Food Chemistry 47: 1208–1216. 20: 175–190. Steiner, J.J., Grabe, D.F. & Tulo, M. 1989. Single and mul- Littell, R.C., Milliken, G.A., Stroup, W.W., Wolfinger, R.D., tiple vigor tests for predicting seedling emergence. Crop 1996. SAS System for Mixed Models. SAS Institute Science 29: 782–786. Inc, Cary, NC. Tahvonen, R., Hannukkala, A. & Avikainen, H. 1995. Ef- López-Castaňeda, C., Richards, R.A., Fraquhar, G.D. & fect of seed dressing treatment of Streptomyces griseo- Williamson, R.E. 1996. Seed and seedling characteris- virides on barley and spring wheat in field experiments. tics contributing to variation in early vigor among tem- Agricultural Science in Finland 4: 419–427. perate cereals. Crop Science 36: 1257–1266. Waddington, S.R., Cartwright, P.M. & Wall, P.C. 1983. A Mikkelsen, K. 1963. Resultater av forsøk med ulik sterk quantitative scale of spike initiation and pistil develop- sortering såkornet i havre. Nordisk Jordbrugsforsking. ment in barley and wheat. Annals of Botany 51: 119 – Supplement 8-9. Nordiske Jordbrugsforskeres Foren- 130.

234 PB AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat Vol. 20(2011): 235–244.

Meat bone meal as fertiliser for barley and oat

Lin Chen1, Jukka Kivelä1*, Juha Helenius1 and Arjo Kangas2 1Department of Agricultural Sciences, Fin-00014 University of Helsinki 2MTT Agrifood Research Finland, Alapääntie 104, 61400 Ylistaro, Finland *e-mail: [email protected]

The traditional production of mineral N and P fertilisers is unsustainable due its reliance on fossil fuels in the case of N, and on limited mineral resource stocks in the case of P. The use of alternative or complementary fertilisers that originate from organic waste materials is gaining interest. Organic farms, especially arable organic farms without livestock, need usable sources of plant nutrients. Meat bone meal (MBM), a potential organic fertiliser for agricultural crops, contains considerable amounts of nutrients (on average 8% N, 5% P, 1% K and 10% Ca). In EU countries, Commission regulation (EC) No 181/2006 authorised the use of MBM as an organic fertiliser. In this study, MBM was compared to conventional mineral NPK fertiliser. Two randomised complete block split-plot field experiments were conducted: one with spring barley (Hordeum vulgare) in two years; and another with oat (Avena sativa) for three years, including a fourth year of testing for residual effect. Compared to mineral fertiliser (20% N, 3% P and 9% K), MBM was applied at three N levels: 60, 90 and 120 kg N ha-1. The grain yield of both cereal species supported by MBM, did not differ from the yield obtained with the mineral fertiliser at any N level. At 120 kg N ha-1, the grain yield level with either type was ca. 4500 kg ha-1 of barley and 5000 kg ha-1 of oat, representing fair averages for Finnish conditions. Moreover, MBM and mineral fertilisation showed no differences in quality in terms of 1000-grain weight, test-weight, protein content and protein yield. Since MBM has a low N/P ratio, P was applied in surplus to attain comparable N levels. Therefore MBM fertilisation should be fitted for crop rotation and for meeting environmental requirements.

Key-words: meat bone meal, nitrogen, phosphorus, organic fertiliser

PB 235 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat

Introduction cally grown wheat (Fredriksson et al. 1997). The concentration of heavy metals, the endogenous infection of moulds, and mycotoxins showed no con- As a product of the rendering industry, MBM sistent differences in wheat fertilised with MBM or contains about 8% N, 5% P, 1% K and 10% Ca, with urea (Salomonsson et al. 1995). Using MBM varying according to the rendering process and to as a fertiliser could also reduce the incidence of the origin of the offal. Before MBM was suspected potato scab (Verticillium dahliae) and curb popu- as the cause of bovine spongiform encephalopathy lations of parasitic nematodes (Lazarovits et al. (BSE) when fed to ruminants, it was commonly 1999). used to feed animals (Brewer 1999). The use of The availability of P in MBM has also been animal meal in the food and feed industry was studied. P is mostly present in MBM as apatite in banned in 2000, and its use as fertiliser material is the bone fraction whereas it exists in organic form currently strictly regulated in the European Union in the meat fraction (Jeng et al. 2006). pH is an (EU) (Werner 2003). Since 2006, however, Com- important factor influencing P release from bone mission regulation (EC) No 181/2006 has permitted meal. Ylivainio et al. (2007) found that 90% of P the use MBM as fertiliser for arable crops in the EU. in MBM was soluble only in 1 M HCl. P in MBM Following the overall idea of industrial ecol- has a residual effect: P applied in soil will remain ogy (Graedel 1996, Frosch and Gallopoulos 1989), for at least three to five years in acid soils and even the recycling of nutrients in MBM should improve longer if the pH is over 6.5. In barley and canola the eco-efficiency of natural resource use in the seed production, bone meal proved to be a more food system. The annual production of MBM in the effective P fertiliser than rock phosphate (Bekele EU, for example, was about three million tonnes and Höfner 1993). in 2001 (Werner 2003). The total amount of N and The aim of this study was to analyse the effect P from animal meals in 15 EU countries could be of MBM as an organic fertiliser on crop yield and estimated at 120 million kg N and 155 million kg quality, and to compare this with conventional min-

P, or about 360 million kg P2O5 (Werner 2003). In eral fertiliser. In particular, the aim was to compare Finland, at least 200 million kg (fresh weight, in the effects over a range of N fertilisation levels, in ca. 35 % moisture content) of animal by-products order to obtain applicable results for recommenda- are generated annually (Salminen 2002), of which tions to farmers. Spring cereals barley and oat were about 15 million kg are classified as high-risk ma- chosen as test crops because of their importance terial and must be incinerated. The sales for ferti- in Finland. liser nutrients were about 154 million kg of N and 12 million kg of P in Finland in 2009. If all but the high-risk fraction of Finnish animal by-products were recycled as fertiliser, the amount of nutrients would equal at least 6 % of N and 50 % of P in Materials and methods the industrial chemical fertilisers used on Finnish agricultural lands. Study area and soil analysis As a fertiliser, MBM has recently been tested for cereals. Jeng et al. (2004, 2006) found MBM to The experiments were carried out during 2000–2003 be an effective organic fertiliser for spring wheat at the Ylistaro research station (62° 56′ 25″ N, and barley. Salomonsson et al. (1994, 1995) studied 22° 31′ 0″ E) of MTT Agrifood Research Finland. the impact of MBM fertiliser on the grain protein Ylistaro is located in the boreal zone in the prov- content in spring and winter wheat and found N ince of Western Finland. Agriculture in the region availability in MBM to be similar to that in urea. is dominated by production of spring cereals such Additionally, MBM has been found to provide suf- as barley, especially for feed, oat, wheat and spring ficient N for good baking performance of organi- canola. 236 237 AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat Vol. 20(2011): 235–244.

The weather conditions were typical of the re- Material used in the experiment gion. May 2001 stands out as cooler and rainier than the month of May in the other years of experi- The spring barley (Hordeum vulgare L.) cultivar mentation (Table 1). The soil had a texture of silty used in the experiment was Thule, a six-row barley. clay loam. According to the WRB system (FAO The spring sown oat (Avena sativa L.) cultivar in 2006), the experimental soil was tentatively clas- 2000 and 2001 was Veli, and in 2002 and 2003 sified as Endogleyic Cambisols (Humid, Dystric, Belinda. Seed dressing fungisides were not applied. Siltic) (see also Yli-Halla et al. 2000). The soil was Pellon Y3® 20-3-9 (PY3) from Kemira GrowHow, classified in the category of low organic matter a granular mineral fertiliser with 20.0% N (about content, with a class range of 3–6% organic mat- 11.0% ammonium, 9.0% nitrate), 3.0% P and 9.0% ter (Table 2). Soil P, K, Ca, and Mg were extracted K, served as the mineral comparison. Other nutrient by acid ammonium acetate, pH 4.65, ratio 1:10 v/v, content in PY3 includes 0.5% magnesium, 3.0% 1 h (Vuorinen and Mäkitie 1955). sulfur, 0.02% boron and 0.0015% selenium. Honka- joki Oy provided MBM 7-5-1 for the experiment as well as the data on its content (Table 3).

Table 1. Temperature and rainfall in the growing seasons from 2000 to 2002, and in the 1960–91 reference period at Ylistaro. Mean temperature (°C) Total Rainfall (mm) Month 2000 2001 2002 2003 1960–91 2000 2001 2002 2003 1960–91 May 10.3 7.6 11.3 9.7 8.8 36.9 89.2 44.4 69.5 38 June 13.9 14.2 15.8 13.1 14.0 38.5 63.0 64.2 68.2 42 July 16.1 17.6 18.1 19.5 15.5 75.2 61.4 82.3 72.3 68 Aug 13.9 14.5 17.7 14.3 13.6 77.6 79.4 56.6 31.2 70 Mean 13.6 13.5 13.5 14.2 13.0 228.2 293.0 247.5 241.2 218

Table 2. Soil pH and nutrient contents in the experimental fields of barley and oat in 2000, before the experiments. Barley Conditions1 Oat Conditions pH 5.8 tolerable 5.4 rather low Ca, mg l-1 726 rather low 728 rather low P, mg l-1 7.0 tolerable 5.5 tolerable K, mg l-1 131 acceptable 113 tolerable Mg, mg l-1 173 acceptable 144 acceptable

1 The growth condition interpreted according to the Finnish standard system of soil fertility assessment (for details, see for example Peltovuori 1999).

236 237 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat

Table 3. The macronutrient (%), micronutrient (mg kg-1) and heavy metal (mg kg-1) content of MBM. The data are from an analysis of the rendering plant (Honkajoki Oy in Finland) from which the MBM used in the experiment originated (ws = water soluble). Macronutrients Micronutrients Heavy metals N 7.0 B 25 Pb 0.50 N ws 2.5 Co 0.15 Cd 0.1 P 5 Cu 3.9 Hg 0.01 P ws 0.15 Fe 58 Ni 0.55 K 1 Mn 4 Ca 12.0 Zn 55 Mg 0.8 Se 0.19 S 0.5 Na 0.5

Experimental design rate x fertiliser type, an additional control plot (no fertilisation) was randomised for each block. The two experiments were both factorial randomised complete block split-plot designs with four replicates. The sowing and harvesting dates Crop management of barley were 19 May, 17–19 August 2000; and 22 May, 16–20 August 2001. The dates for oat were 19 May, 27–28 August 2000, 22 May, 20–24 The crops were rain-fed. Based on monitoring for August 2001, and 15 May, 15 August 2002. These need of pest control, control measures were not dates were typical to farming in the area. To test needed. Weeds were controlled by spraying with for a residual effect on the following year’s crop, standard tribenuron-methyl herbicide Express Clas- the experiment with oat was continued into 2003. sic® and fluroxypyr herbicide Starane® 0.41 ha-1. In this final year, all the plots were sown 22 May The previous crop in the barley experiment was with cv. Belinda without fertiliser. The experimental rye; in the oat experiment, the crops were winter plots were physically maintained in the same places rye and wheat. over the years. The main plot size was 2.2 m × 10 m. The blocks were placed 3 m apart. The main plot factor was fertilisation rate at Measurements three levels, adjusted for N at steps 60, 90, and 120 kg ha-1 (Table 4). At each N step, levels at which P, K, and the other nutrients were applied also in- Barley and oat grain yields (kg ha-1) were harvested creased at the same rate as N. The sub-plot factor and reported at 15% moisture. Four parameters of was fertiliser type. In addition to treatments derived grain quality, namely 1000-grain weight (TGW, g), from factorial combinations of fertiliser application test weight (kg hl-1), protein content (%) and pro-

Table 4. Amount of fertiliser and N, P, and K applied in the experiments (kg ha-1). N 60 N 90 N 120 Fertiliser type Fertiliser P K Fertiliser P K Fertiliser P K PY3 300 9 27 450 14 41 600 18 54 MBM 857 43 9 1285 64 13 1714 86 17

238 239 AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat Vol. 20(2011): 235–244.

tein yield (kg ha-1), were analysed. Protein content highest levels (p < 0.05). In the second year, both was measured with the Kjeldahl method at MTT’s PY3 and MBM significantly increased barley yield laboratory in Jokioinen. by 91% and 84% respectively (p < 0.000) even at the lowest fertilisation level. As with barley, oat grain yield (Fig. 2) did not Data analysis differ between the fertiliser types. In the fertilised

Barley yield (kg ha-1) Analysis of variance (ANOVA) was performed with 5000 fertiliser SPSS ® according to the experimental design. The 0 GLM repeated measures procedure was used to test PY3 4000 MBM linear response effects in the full factorial split-split plot design. The year was included as the split-split plot level factor (Gomez and Gomez 1984, pp. 153 3000 and 265), as the treatments defined by fertilisation rate × fertiliser type, and the respective measurements at 2000 split-plot level were repeated over two years (in the

case of barley) or three years (in the case of oat). When 1000 including the control, the MBM and PY3 fertiliser effects were analysed separately, the model being a 0 fully randomised complete blocks ANOVA at the 0 60 90 120 main plots level. Tukey’s test was used to compare N kg ha-1 means at a 5% risk. Nitrogen use efficiency (NUE %) was calculat- Fig. 1. Barley grain yield in 2000 and 2001. Comparison ed as the degree of recovery of fertiliser. NUE was at three N-levels of the effect of meat bone meal (MBM) measured by the following equation: NUE (%) = (N with the effect of mineral fertiliser (PY3), and with a control without fertilisers. uptake (kg ha-1) in fertilised plots – N uptake (kg ha- 1) in unfertilised plots) / N application rate (kg ha-1). Oat yield (kg ha-1) 6000 fertiliser 0 5000 PY3 Results MBM 4000 Grain yield 3000 The same barley grain yield (Fig.1) was obtained with MBM as with PY3 (no significant difference). 2000 Barley grain yield increased from 3561 kg ha-1 (SD 1000 = 307 kg ha-1, n = 16) at the lowest fertilisation level -1 -1 to 4309 kg ha (SD = 476 kg ha , n = 32) at the two 0 higher fertilisation levels (p < 0.05). Compared to 0 60 90 120 -1 unfertilised control, less yield increase was obtained N kg ha with fertilisation in the first year than in the second Fig. 2. Oat grain yield in 2000, 2001 and 2002. year (p < 0.01). In the first year, PY3 significantly Comparison at three N-levels of the effect of meat bone increased barley yield by 41% only at the highest meal (MBM)with the effect of mineral fertiliser (PY3), fertilisation level and MBM by 40% at the two and with a control without fertilisers. 238 239 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat treatments, the average oat yield was 4913 kg ha-1 received the lowest fertilisation level in the three (SD = 708 kg ha-1, n = 69), and differences among previous years, to 2606 kg ha-1 (SD = 372 kg ha- the fertilisation levels were not significant. Com- 1, n = 7) in the plots with a history of the highest pared to unfertilised control, both fertilisers signifi- fertilisation level (Tukey’s test, p < 0.05) (Fig. 3). cantly increased oat yield at the two highest ferti- In comparison to the unfertilised control (Fig. 3), lisation levels. In the first year, PY3 increased oat oat grain yield significantly increased only when yield even at the lowest application rate (p < 0.01). the fertilisation rate in the previous years had been the highest (Tukey’s test, p < 0.05). Residual effect on oat grain yield 1000-grain weight (TGW) The result showed that MBM had a higher residual fertilisation effect on oat grain yield than PY3 (p Average TGW of barley (Table 5) was 39.8 g (SD < 0.05). The plots fertilised by MBM in the three = 2.3 g, n = 54). No consistent effect of either ferti- previous years produced, on an average, 381 kg liser type or fertilisation level on barley TGW over ha-1 (20 %) more than the plots fertilised by PY3 the two years was found (three-way interaction, (Fig. 3). The advantage of MBM over PY3 was p < 0.05). In comparison to unfertilised control, consistent over all the three fertilisation levels (no in the first year, both PY3 and MBM significantly significant interaction). increased barley TGW only at the highest fertilisa- The residual effect on yield increased with in- tion level (p < 0.05). In the second year, PY3 at creasing fertilisation level (p < 0.05). There was the two highest fertilisation levels and MBM at a significant rise in oat grain yield from 1780 kg the lowest fertilisation level significantly increased ha-1 (SD = 165 kg ha-1, n = 7) in the plots that had barley TGW (p < 0.001). Average TGW of oat was 40.2 g (SD = 1.9 Oat yield (kg ha-1) g, n = 70). Oat TGW did not differ between the 3000 fertiliser two fertiliser types nor between the fertilisation

0 levels and no interaction between the factors was PY3 found. In comparison to unfertilised control, ap- MBM plying PY3 increased oat TGW by 4% at the two 2000 highest application levels (p < 0.05), but applying MBM did not have a significant effect. No significant residual effect on oat TGW was found.

1000 Test weight

0 Barley test weight (Table 5) did not depend on 0 60 90 120 fertiliser type. A significant (p < 0.05) increase -1 N kg ha in barley test weight from 64.4 kg hl-1 (SD = 2.8 kg hl-1, n = 16) at the lowest fertilisation level to Fig. 3. Oat grain yield in the fourth year (2003), in which an average 65.5 kg hl-1 (SD = 2.7 kg hl-1, n = 32) fertilisers were not applied, after three previous seasons at the two higher fertilisation levels was observed of fertilisation. Comparison at three N levels of the residual effect of meat bone meal (MBM) with the residual ef- (p < 0.01). In comparison to unfertilised control, fect of mineral fertilizer (PY3), and with a control of no PY3 and MBM significantly increased barley test fertilization (0) in the three previous years. 240 241 AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat Vol. 20(2011): 235–244.

Table 5. Average barley and oat 1000 grain weight (TGW, g), test weight (kg hl-1), protein content (%), Protein yield (kg ha-1) and nitrogen use efficiency (NUE, %) during two years. N TGW Test weight Protein content Protein yield NUE Fertiliser (kg ha-1) Barley Oat Barley Oat Barley Oat Barley Oat Barley Oat

0 37.2 39.7 63.4 55.4 10.1 10.9 215.8 293.6 60 PY3 39.0 40.2 64.6 56.5 10.2 11.6 306.0 456.5 24.2 43.5 MBM 38.4 40.3 64.3 55.7 10.6 11.0 328.5 414.2 30.2 31.0 90 PY3 39.6 40.5 65.2 56.2 11.2 12.3 388.0 517.2 30.6 39.8 MBM 40.8 40.0 65.4 55.7 11.0 12.0 386.7 498.1 30.4 35.6 120 PY3 41.4 40.6 65.9 54.9 12.0 13.0 452.6 592.7 31.5 39.9 MBM 42.0 40.0 65.6 55.0 11.8 13.1 459.4 594.9 32.6 40.7

weight at the two highest fertilisation levels by Oat protein content did not differ between 3.3% (p < 0.05). MBM and PY3. At each step of increase in ferti- Oat test weight was on average 55.7 kg hl-1 lisation rate, a significant increase in oat protein (SD = 2.5 kg hl-1, n = 70). No consistent effect content was obtained from 11.4% (SD = 1.3%, of either fertiliser types or levels on oat gain test n = 23) through 12.2% (SD = 1.3%, n = 23) to weight were observed, over the three years (three- 13.1% (SD = 1.1%, n = 23) respectively (p < way interaction, p < 0.05). In comparison to un- 0.01). Compared to unfertilised control, with an fertilised control, neither fertilisers´ effects were average protein content of 10.8% (SD = 1.0%, n = consistent during the three years. In the first two 12), both fertilisers at the two highest fertilisation years, both fertilisers gave the same level oat test levels significantly increased oat protein content weight as the control. In the third year, PY3 at by 16% (p < 0.01). the highest fertilisation level decreased oat test weight by 4% (p < 0.001) and MBM did not differ from the control at any of the three levels. Protein yield

Protein content The protein yield of barley (Table 5) did not differ between MBM and PY3. At each step of increase in fertilisation rate, a significant 22% and 44% Barley protein content (Table 5) did not differ increase in barley protein yield was achieved (p between PY3 and MBM. At each step of increase < 0.01). Compared to the control, in which the in fertilisation rate, a significant increase in barley average was 216 kg ha-1 (SD = 79 kg ha-1, n = 8), protein content was obtained from 10.5% (SD = irrespective of the type of fertiliser, fertilisation 1.1%, n = 16) through 11.1% (SD = 1.1%, n = 16) even at the lowest level significantly increased to 11.9% (SD = 0.9%, n = 16) respectively (p < barley protein yield by 47% (p < 0.001). 0.001). Compared to the control, with an average As in the case of barley, oat protein yield did protein content of 10.1% (SD = 1.1%, n = 8), PY3 not differ between the two fertiliser types. At each at the two highest fertilisation levels and MBM step of increase in fertilisation rate, oat protein at the highest fertilisation level significantly in- yield increased significantly by 16% and 36% creased barley protein content by 15% and 18% from 435 kg ha-1 (SD = 73 kg ha-1, n = 23) (p < respectively (p < 0.01). 0.01). Compared to the control, in which the aver- 240 241 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat age was 294 kg ha-1 (SD = 86 kg ha-1, n = 12), PY3 yields, with equal nitrogen use efficiency. The re- even at the lowest level significantly increased oat sult was consistent with those of experiments with protein yield by 56% (p < 0.001). The fertilisation winter and spring wheat in Sweden, which showed effect of MBM was not consistent during the three that MBM and urea produced equally good yield years (p < 0.05). In the first year, MBM signifi- increases (Salomonsson et al. 1994, 1995). The cantly increased oat protein yield by 54% only doubling of the application rate from 857 kg to at the two highest fertilisation rates (p < 0.001), 1714 kg MBM ha-1 did not significantly increase while in the last two years, MBM increased oat barley and oat grain yield in the first year; it did, protein yield by 61% even at the lowest fertilisa- however, in the second year. A possible explanation tion level (p < 0.001). could be better initial nutrient status in the soil in the first year than in the second year. Lundström and Lindén (2001) found that in soils with high Nitrogen use efficiency plant-available N that was partly released in spring and partly released by mineralisation during the growing season, an N application of as little as In the barley experiment NUE (Table 5) ranged 40 kg N ha-1 in MBM was sufficient. Jeng et al. from 9.2% to 40.2%. No difference in NUE was (2006), experimenting with spring barley, obtained observed between the two fertiliser types or ferti- a yield increase with 500 kg MBM ha-1, but no lisation rates, but NUE in the second year (average further yield increase with increased application 20.8%, SD = 9.1%, n = 24) was significantly higher rates. Obviously, the fertilisation effect of MBM (p < 0.01) than in the first year (average 38.9%, SD is determined by soil nutrient status. = 6.1%, n = 24). The average C/N ratio in MBM is around 3–4 In the oat experiment NUE ranged from 14.6% (Jeng et al. 2004), favourable to microbial miner- to 58.8%. No significant difference appeared be- alisation. Rapid mineralisation of N from MBM tween the two fertilisers. NUE differed among the in soil was found by Chaves et al. (2005) and by three fertilisation rates, but the difference was not Mondini et al. (2008). Jeng et al. (2004), experi- consistent over the years (fertilisation rate × N menting with spring cereals, however, found that level, two way interaction, p < 0.05). the application of small amounts of MBM (60 kg N ha-1) resulted in a high proportion of immobilisation shortly after the seedling stage. When the application rate was higher, the immobilisation had a lesser effect on N uptake. The MBM they Discussion used contained 7 % N, of which 2.5 % was soluble ammonium-N and the remaining 4.5% organic This study aimed at screening MBM as a fertiliser substances. for spring cereals at variable application rates rather In addition to N, the P effect of MBM is also than testing any potential MBM-derived fertilisers considerable. Ylivainio et al. (2007) found that al- of adjusted nutrient contents. Hence, the levels of though P in MBM was mainly acid soluble early all nutrients, not just N, increased with increased on, it was eventually converted to a plant-available application rates. A single nutrient’s contribution to form. In the first year, only 19 % of the P became the results remained confounded with the N-level. available to the plants, but in the third year of the The N rate itself was adjusted according to the total experiment, MBM was equal to dairy manure and content, as plant availability and rate and seasonal super phosphate in producing ryegrass yield (Yli- pattern of release of N is not sufficiently known. We vainio et al. 2007). Moreover, MBM serves not only discuss the results with these reservations in mind. as a fertiliser, but by increasing activity of micro- In the present study, MBM was as effective as organisms, it serves as soil conditioner (Mondini et mineral fertiliser in increasing barley and oat grain al. 2008). Kahiluoto and Vestberg (1998) reported 242 243 AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat Vol. 20(2011): 235–244.

that MBM as well as manure enhanced arbuscular ratio would be about eight, which is close to crops’ mycorrhiza, while mineral fertiliser suppressed it. needs. Therefore MBM can be a source of P for In our experiment, the residual effect on oat organic farming and especially for perennial plants grain yield of MBM was better than the effect of and pastures, which can absorb large quantities of the mineral fertiliser. With both, the residual yield P. In this way, the runoff of P can be reduced. increase was significant only at the highest rates Another essential macronutrient for crops (is of application over the previous three years. The potassium. To balance the nutrient contents in slow release of P, and even some residual effect MBM, K from different by-products, such as vi- from the mineralisation of organic N in the years nasse from the sugar beet industry, could be added that followed, may explain the advantage of MBM to MBM to achieve NPK contents more optimal over mineral fertiliser. Further process-oriented for crops. Formulation, such as granulation or slur- studies covering more than one season’s residual ring, could be developed. Such development and effect are needed to understand the long-term ef- testing would be needed for both agricultural and fects of MBM application. horticultural crops. The market value of a grain crop is largely de- Jeng and Vagstad (2009) found that autumn and termined by its test weight and TGW. Like the min- early spring MBM application may cause nutrient eral fertiliser, MBM performed well in building up loss in the environment, especially leaching or run barley and oat grain TGW and test weight. Protein off of N and P. Thus, the timing of MBM appli- content is an important factor of grain quality, and cation should be close to sowing or planting, or also contributes to the market price. Protein content MBM could be applied to an established crop. In is often restricted by N supply. MBM showed a Finnish conditions, granulated fertiliser for arable similar good effect on barley and oat, which was crops is usually applied simultaneously with the in line with results from spring and winter wheat seed by a combined application machine. This is an published by Salomonsson et al. (1994, 1995) and additional argument in support of granulated MBM by Fredriksson et al. (1997, 1998). Protein yield products. The study did not assess environmental is determined by grain yield and protein content. risks or benefits. The further development of MBM As the N level increased, both the grain yield and fertilisation should include environmental impact protein content increased, leading to an increase in assessment. protein yield. As with the mineral fertiliser, MBM Considering the high N and P content of MBM, showed the same effect in increasing grain protein disposal in landfills or incineration in specific fa- yield. cilities for garbage incineration, power stations and Raising N results in taller plants and may lead the cement industry seems wasteful. Because 70% to lodging, which increases the risk of quality of the MBM produced originates from “no risk” or loss. Optimal N management must balance yield “low risk” material, meals of animal origin should improvement against reductions in grain quality. be recycled within the agrifood system. In the present study, lodging was negligible in all We conclude that MBM is an efficient fertiliser treatments, and did not contribute to the results. for cereals. We recommend product development Mineral fertilisation and MBM produced equally for improved nutrient ratios by mixing MBM with good quality at all N application levels. other suitable by-products of the food industry, The nutrient content of MBM (7-5-1) differs and formulation into granulated products. More from crop requirements. The N/P uptake ratios for research in P dynamics in MBM in soil is needed, most crops range from 4.5 to 9, so applying MBM both for crop management and for evaluating en- according to N needs would lead to a surplus of P. vironmental risks. Recycling nutrients within the According to Ylivainio et al. (2007), only about agrifood system is an important path towards in- 20 % of the P in MBM is immediately available to dustrial ecologisation within the food sector. plants. Thus, the immediately plant-available N/P

242 243 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Chen, L. et al. Meat bone meal as fertiliser for barley and oat

Kahiluoto, H. & Vestberg, M. 1998. The effect of arbuscu- References lar mycorrhiza on biomass production and phosphorus uptake from sparingly soluble sources by leek (Allium porrum L.) In Finnish field soils. Biological Agriculture Bekele, T. & Höfner, W. 1993. Effects of different phosphate and Horticulture 16: 65–85. fertilisers on yield of barley and rape seed on reddish Lazarovits, G., Conn, K. & Potter, J. 1999. Reduction of po- brown soils of the Ethiopian highlands. Nutrient Cycling tato scab, Verticillium wilt, and nematodes by soymeal in Agroecosystems 34: 243–250. and meat and bone meal in two Ontario potato fields. Brewer, M.S. 1999. Current status of bovine spongiform Canadian Journal of Plant Pathology 21: 345–353 encephalopathy - A review. Journal of Muscle Food Lundström, C. & Lindén, B. 2001. Nitrogen effects of hu- 10: 97–117. man urine, meat bone meal (Biofer) and chicken manure Chaves, C., Canet, R., Albiach, R., Marin, J., & Pomar- (Binadan) as fertilisers applied to winter wheat, spring es, F., 2005. Meat and bone meal: fertilizing values and wheat, and spring barley in organic farming. Swedish rates of nitrogen mineralization. In: Bernal, M.P., Moral, University of Agricultural Sciences, Department of Ag- R., Clemente, R., Paredes, C. (Eds.). Proceedings of the ricultural, Skara, Series B crops and soils, report 8:51. 11th International Conference RAMIRAN, Murcia, Spain, Mondini, C., Cayuela, M.L., Sinicco, T., Sanchez-Monede- vol. 1, 6–9 October 2004, p. 177–180. ro, M., Bertolone, E. & Bardi, L. 2008. Soil application of FAO 2006. World reference base for soil resources 2006. meat and bone meal. Short-term effects on mineraliza- A framework for international classification, correlation tion dynamics and soil biochemical and microbiological and communication. Food and Agriculture Organization properties. Soil Biology & Biochemistry 40: 462– 474. of the United Nations, Rome. 103 p. Peltovuori, T. 1999. Precision of commercial soil testing Fredriksson, H., Salomonsson, L. & Salomonsson, A. 1997. practice for phosphorus fertilizer recommendations in Wheat cultivated with organic fertilisers and urea: bak- Finland. Agricultural and Food Science in Finland 8: ing performance and dough properties. Acta Agricul- 299–308. turae Scandinavica, Section B - Soil & Plant Science Salomonsson, L., Jonsson, A., Salomonsson, A. & Nils- 47: 35–42. son, G. 1994. Effects of organic fertilisers and urea when Fredriksson, H., Salomonsson, L., Andersson, R., Salo- applied to spring wheat. Acta agriculturæ Scandinavica. monsson, A.-C. 1998. Effects of Protein and Starch Section B, Soil and Plant Science 44: 170. Characteristics on the Baking Properties of Wheat Cul- Salomonsson, L., Salomonsson, A., Olofsson, S. & Jons- tivated by Different Strategies with Organic Fertilisers son, A. 1995. Effects of organic fertilisers and urea when and Urea. Acta agriculturæ Scandinavica. Section B - applied to winter wheat. Acta agriculturæ Scandinavica. Soil & Plant Science 48: 49. Section B, Soil and Plant Science 45: 171. Frosch, R.A. & Gallopoulos, N.E. 1989. Strategies for man- Salminen, E. 2002. Finnish expert report on best available ufacturing. Scientific American 261: 144–152. techniques in slaughterhouses and installations for the Gomez, K. A. & Gomez, A. A. 1984. Statistical Proce- disposal or recycling of animal carcasses and animal dures in Agricultural Research, 2nd edition. 680 p. waste. The Finnish Environment 539: 42. New York, Wiley Vuorinen, J. & Mäkitie, O. 1955. The method of soil testing Graedel, T.E. 1996. On the concept of industrial ecology. used in Finland. Agroecological publications 63:1–14. Annual Review of Energy and Environment 21: 69–98. Werner, W. 2003. Complementary nutrient sources. In: Jeng, A., Haraldsen, T.K., Vagstad, N., Grønlund, A & IFA-FAO Agriculture Conference Global Food Securi- Tveitnes, S. 2004. Meat and bone meal as nitrogen fer- ty and the Role of Sustainable Fertilisation, Rome, 26– tiliser to cereals in Norway. Agricultural and Food Sci- 28 March 2003. ence 13: 268–275. Yli-Halla, M., Mokma, D., Peltovuori, T. & Sippola, J. 2000. Jeng, A.S., Haraldsen, T.K., Grønlund, A. & Pedersen, P.A. Suomalaisia maaprofiileja (Abstract: Agricultural soil 2006. Meat and bone meal as nitrogen and phospho- profiles in Finland and their classification). Maatalou- rus fertiliser to cereals and rye grass. Nutrient Cycling den Tutkimuskeskuksen Julkaisuja, Sarja A 78. 104 p. in Agroecosystems 76: 183–191. (In Finnish). Jeng, A.S. & Vagstad, N. 2008. Potential nitrogen and Ylivainio, K., Uusitalo, R. & Turtola, E. 2007. Meat bone phosphorus leaching from soils fertilized with meat and meal and fox manure as P sources for ryegrass (Lolium bone meal. Acta Agriculturae Scandinavica, Section B multiflorum) grown on a limed soil. Nutrient Cycling in - Plant & Soil Science 59: 1–8. Agroecosystems, 81:267–278.

244 PB AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland Vol. 20(2011): 245–261.

Composition of weed flora in spring cereals in Finland – a fourth survey

Jukka Salonen*, Terho Hyvönen and Heikki Jalli MTT Agrifood Research Finland, Plant Production Research, FI-31600 Jokioinen, Finland, *e-mail: [email protected]

The weed flora in conventionally and organically grown spring cereals was investigated in southern and central Finland during 2007–2009. The survey was conducted in 16 regions, 283 farms and 595 fields (72 organically cropped and 523 conventionally cropped fields, of which 503 were treated with herbicides). The occurrence of weeds was assessed in late July–early August. Altogether 148 weed species were identified, of which 128 were broad-leaved and 20 grass species. In organically cropped fields, the average species number per field was 21 and the most frequent species were Chenopodium album 96%, Stellaria media 94%, Viola arvensis 94% and Elymus repens 89%. In conventionally cropped fields, the average species number was 12 and the most frequent weed species were Viola arvensis 83%, Stellaria media 65%, Galeopsis spp. 59% and Galium spurium 59%. The average density of weeds was 160 m-2 (median = 112) in sprayed conventional fields and 519 -2m (468) in organic fields. The average air-dry biomass of weeds was 167 kg ha-1 (median = 82) and 775 kg ha-1 (563), respectively. Elymus repens, the most frequent and abundant grass species, produced the highest proportion (about 30%) of the total weed biomass in both cropping systems. The frequency of Galium spurium in conventional cropping and Fumaria officinalis in organic cropping had increased substantially since the previous survey in 1997–1999. The average size of the weed seedbank in the 5 cm surface layer was about 1 700 seeds m-2, the most predominant seeds being of C. album. Although the weed flora in Finnish spring cereal fields consists of numerous species, only a fraction of them severely threaten crop production in terms of their frequency and abundance. Weeds in conventional cropping were effectively controlled with available herbicides whereas weed management in organic cropping calls for urgent measures such as direct mechanical weed control in crop stands, which was not practised at all in survey fields.

Key-words: weeds, spring cereals, biodiversity, conventional farming, organic farming, crop protection, seedbank, Elymus repens, Viola arvensis, Galium spurium

PB 245 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland

Introduction continuous changes in cropping practices affect the species composition of the weed flora and the level of weed infestation, as discussed for instance by Three extensive surveys of weeds in spring cereal Andreasen and Streibig (2010). This report focuses fields have been carried out in Finland since the on the composition of weed flora in spring cereals 1960s, the first in 1961–1964 (Mukula et al. 1969), in 2007–2009. The changes in weed occurrence the second in 1982–1984 (Erviö and Salonen 1987) over the decades will be reported in detail sepa- and the third in 1997–1999 (Salonen et al. 2001a). rately. In addition to conventionally cropped fields, organi- As a supplement to previous weed surveys, cally cropped fields were included in the survey sampling of soil seed banks was included in the protocol in the 1990s. present survey. The seed density data provide in- Similar comprehensive weed surveys have formation on the potential of germinating weed been conducted earlier in many countries, but new seedlings in the soil. Reference mapping of soil surveys and follow-up monitoring have been real- seed banks in cultivated soils in Finland dates back ized to a lesser extent. Nonetheless, recent informa- to the 1960s (Paatela and Erviö 1971). Since then tion on composition of weed floras and factors af- factors like the use of herbicides (Hyvönen and fecting weed floras in cereal fields is available, e.g. Salonen 2003), crop rotations (Sjursen 2001) and from Bulgaria (Milanova et al. 2009), Denmark changes in tillage practices (Yenish et al. 1992) (Andreasen and Stryhn 2008, Andreasen and Sko- have been of central importance in determining vgaard 2009), France (Fried et al. 2008), Hungary the amount of seeds in the soil seed bank. (Novak et al. 2010), the UK (Potts et al. 2010) and Weed surveys in spring cereal fields in Fin- the US (Conn et al. 2011). In Finland, a small-scale land are important in terms of crop dominance as survey of weeds in organically cropped spring ce- spring cereals account for more than half of the reals in coastal regions was carried out in the early total cropped area of around 2 million hectares 2000s (Riesinger and Hyvönen 2006a). Recently, (Niemi and Ahlstedt 2009). In addition to biodiver- the EWRS (European Weed Research Society) sity aspects (species richness, high-value species, established a working group on weed mapping to findings of new alien species etc.), the updated in- facilitate communication among research groups formation on weed infestation is aimed at farmers, monitoring weeds in arable habitats in Europe advisory services and the chemical industry, with a (http://www.ewrs.org/weedmapping/default.asp). view to promoting specific weed control measures. Since the previous extensive weed survey in Weed shift is of agronomic interest primarily as Finland (Salonen et al. 2001a), which was con- regards the most abundant weeds species. ducted in 1997–1999, agri-environment policy has influenced the preconditions for cereal production. Some measures originating from EU regulations, such as restricted use of fertilizers, demand for over-wintering plant cover, reduced tillage and Material and methods uncultivated buffer zones along watercourses have been extensively implemented in Finnish farms. Study regions, farms and fields Furthermore, several studies have demonstrated that organic farming clearly promotes weed floras, The weed survey was carried out in southern and both in terms of diversity and abundance (e.g. van central Finland in 2007–2009 (Fig. 1). The highest Elsen 2000, Salonen et al. 2001b, Hyvönen et al. number of fields (N = 369) was studied in southern 2003a, Hyvönen 2007, Romero et al. 2008). Finland (Table 1), where spring cereals are grown In the present study, we report the results of on more than 50% of the arable area and annual the fourth extensive weed survey of Finnish spring spring-sown crops are predominant. The number of cereals, conducted in 2007–2009. We expected that study fields decreased towards the east and north, 246 247 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland Vol. 20(2011): 245–261.

Fig.1. Weed survey regions (1-16) in three zones in 2007- 2009. Key to region numbers: 1 = Jokioinen, 2 = Lammi, 3 = Nauvo/Korppoo, 4 = Tammela, 5 = Iitti, 6 = Kitee, 7 = Laukaa/ Toivakka, 8 = Vieremä, 9 = Laihia, 10 = Laitila, 11 = Lieto/ Paimio, 12 = Nurmijärvi, 13 = Ruokolahti, 14 = Mikkeli, 15 = Kihniö/Parkano, 16 = Nivala.

where spring cereals account for less than 50% of the 3 378 ha. In the largest fields the investigated area cultivated area, and many farms include grassland was restricted to a maximum of 25 ha and in some in their crop rotation. fields the surveyed field area was adjusted accord- The number of farms visited was 283, of which ing to previous surveys. The same 16 regions were 243 were engaged in conventional and 40 in organ- surveyed ten years ago and altogether 443 out of ic farming. Three survey regions, Nauvo/Korppoo the 595 surveyed fields were the same as in 1997– (region number 3), Tammela (4) and Laihia (9), 1999 (Salonen et al. 2001a). were study regions without organically cultivated A great majority of study fields had been sown survey fields. The proportion of organic survey by mid-May (mainly during weeks no. 18–20) and fields was highest, 30–35%, in eastern Finland sprayed with herbicides by the end of June (weeks (regions 6, 7 and 13) and in Kihniö/Parkano (15). no. 24–26). The difference in cropping systems between survey The permission to investigate weeds in survey regions is reflected to some extent in the results on fields was asked from farmers only in early July overall weed occurrence. with the aim that all cropping practices influenc- One to five spring cereal fields were examined ing the weed infestation, including the level of on each farm, giving a total of 595 fields: 267 un- fertilization or choice of herbicide product, dose der barley (Hordeum vulgare L.), 175 under oats and time of application, were realized according (Avena sativa L.), 148 under wheat (Triticum aes- to “house-style”. At the sampling time, information tivum L.) and 5 under oat/pea or spring-sown rye on cropping measures was recorded by interview- cultivation. Altogether 523 study fields were culti- ing the farmers. vated conventionally and 72 organically. Herbicides had been applied in 503 out of 523, i.e. The average area of the study fields was 5.7 ha 96%, conventionally cultivated fields. A common (range 0.3 ha– 35.6 ha) and the total area surveyed, reasoning for not using chemical weed control 246 247 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland

Table 1. Number of fields surveyed by region, production type and cereal species. Region Number of fields Production type Cereal species7 Year No Zone5 Total Conventional 6 Organic Barley Oat Wheat Municipality 2007 Jokioinen1 1 S 43 37 (36) 6 23 14 5 Lammi2 2 S 50 45 (42) 5 18 15 17 2008 Nauvo/Korppoo 3 S 28 28 (26) 0 9 10 9 Tammela 4 S 40 40 (40) 0 20 14 6 Iitti 5 S 35 31 (27) 4 13 12 8 Kitee 6 E 20 13 (10) 7 8 10 2 Laukaa/Toivakka 7 E 35 32 (30) 3 12 19 3 Vieremä 8 W 26 24 (20) 2 18 6 2 Laihia 9 W 41 41 (41) 0 24 1 16 2009 Laitila 10 S 48 43 (42) 5 27 14 7 Lieto/Paimio3 11 S 66 60 (60) 6 18 7 40 Nurmijärvi 12 S 59 52 (52) 7 24 11 24 Imatra/Ruokolahti 13 E 23 15 (15) 8 9 10 4 Mikkeli4 14 E 25 17 (17) 8 9 13 3 Kihniö/Parkano 15 W 26 18 (18) 8 10 16 Nivala 16 W 30 27 (27) 3 25 3 2 Total 595 523 (503) 72 267 175 148 1 including Humppila, Jokioinen, Koski Tl, Loimaa municipality, Somero, Ypäjä 2 including Hämeenkoski, Kärkölä, Lammi, Mäntsälä, Pukkila 3 including Lieto, Marttila, Paimio, Tarvasjoki 4 including Joroinen, Juva, Jäppilä, Mikkeli, Pertunmaa 5 S = South, E = East, W = West 6 number of sprayed fields in parentheses 7 added to this 5 fields with either cereal/pea or spring-sown rye

was either rainy weather conditions, particularly no. 28–29), by which time the spring cereals had in 2008, or undersown clover/grass mixture. Four reached their heading stage and in most cases around major types of active ingredients were used: pure one month had elapsed since herbicide treatment. MCPA (4-chloro-2-methylphenoxyacetic acid) in The weed sampling protocol was exactly the 7% of treated fields, phenoxy acid mixtures (22%), same as in the previous survey in 1997–1999 (Sa- sulphonylureas (50%) and phenoxy acids + sulpho- lonen et al. 2001a). The occurrence of weeds was nylurea tank mixtures (20%). assessed from 10 sample quadrats randomly placed in each field. For this purpose, each field was split into a 10 x 10 cell grid in which the positions of Weed sampling sample quadrats were set with a random number calculator. The size of grid cells varied among fields according to the area of each field. The weed survey was carried out in 2007–2009 Weed density was determined by counting the during a 4-week period starting in mid-July (weeks number of plants or shoots of grass weeds by spe- 248 249 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland Vol. 20(2011): 245–261.

cies in a rectangular frame measuring 0.1 m2 (25 with attribution are given to the 40 most frequent cm × 40 cm), which was a corner area within a species (Table 2). larger quadrat measuring 1.0 m2 (1.0 m × 1.0 m). The term frequency indicates the proportion The larger quadrat was used for observations on of fields where the species was found. -Differ the presence/absence of each species. The results ences between frequency values were tested with presented in the tables and figures derive from data Fisher’s Exact Test. For each field, the total weed collected from the 0.1 m2 quadrats and pooled over density and biomass were summed, and the aver- the 10 samples in each field. A complete list of ages, standard deviations and median values are the additional weed species found in the presence/ presented by survey regions. Differences between absence observation (1 m2) is given in Appendix the regions were tested with log-transformed val- 1 as a supplement to the 40 most frequent species ues using the MIXED procedure of SAS 9.2 (SAS presented in the tables. Institute Inc, Cary, NC, USA). If not mentioned In four out of ten small sample quadrats, weeds otherwise, the specific results concerning the weed and cereals were cut at the soil surface and their occurrence in conventional cropping derive from biomass was weighed by species after the samples the 503 fields treated with herbicides, i.e. the 20 had been dried in an air-flow dryer at 40°C for sev- non-sprayed fields were excluded. eral days. The air-dry biomass results are presented The similarity of species composition between in kg ha-1 which is the equivalent of 10 × g m-2. zones was compared using Jaccard’s similarity co-

efficientS j (Jaccard 1912) (Sj=c/(A+B-c), where c = number of species common to both samples A and Seed bank sampling B, A = number of species in sample A, B = number of species in sample B). The data were pooled over all fields of each zone before the analysis. The seed sampling was conducted by taking one soil The diversity of weed species was described sample (10 × 10 cm in size, around 5 cm depth) from by species richness. The number of species was every weed sampling quadrat (0.1 m2, see above). used as a measure of species richness. Since the The ten samples were mixed in the bucket and one 4 number of species depends on the sample size deciliter soil sample was taken for the final sample. and since the number of sampled fields varied Soil samples were taken to the laboratory and dried among zones, regions and production types total at room temperature. Seeds were separated from soil species numbers could not be compared directly.

with tweezers and identified by species. Therefore, the expected number of species E(Sn) was calculated for each zone, region and production type using rarefaction: Nomenclature and data analysis   N − N     i   S  n  =  −  All weed species found in sampling areas were E(Sn ) ∑1  =  N assessed. Nevertheless, some genera or taxa, e.g. i 1     Galeopsis spp. and Lamium spp., were pooled since  n   they could not be unequivocally identified to species

level at the small seedling stage. where E(Sn) = expected number of species in a The plant species nomenclature follows Hämet- random sample of n individuals, S = total number

Ahti et al. (1998), and the abbreviations of species of species in the entire collection, Ni = number of names are according to the EPPO code system individuals in species i, N = total number of indi- (formerly BAYER codes) available at the EPPO viduals in the collection, n = sample size (number web site (www.eppo.org). The full scientific names of individuals) chosen for standardization (see Heck et al. 1975). 248 249 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland

Table 2. Frequencies (%) of weed species by regions Year / Region Species / Taxon 2007 2008 2009 Average 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 2007–09 Achillea millefolium L 0 8 0 0 9 20 9 0 2 2 2 8 17 4 8 0 5 Alopecurus geniculatus L. 5 0 0 0 3 5 6 12 2 0 8 7 4 4 8 10 4 Artemisia vulgaris L. 0 10 0 3 0 20 0 0 0 21 0 14 17 4 8 0 6 Brassica rapa L. ssp. oleifera (DC.) Metzg. 9 4 4 3 3 10 3 0 10 4 0 3 4 24 4 3 5 Capsella bursa-pastoris (L.) Medik. 23 14 25 23 34 60 37 35 15 42 9 22 35 56 19 30 27

Cerastium fontanum Baumg. 7 6 0 5 11 60 0 27 0 0 0 12 9 12 12 0 8 Chenopodium album L. 65 56 57 60 63 90 66 62 78 44 64 31 70 48 35 83 59 Cirsium arvense (L.) Scop. 26 52 4 20 37 40 9 19 24 19 44 32 35 24 12 23 28 Elymus repens (L.) Gould 30 50 46 63 51 80 63 85 54 46 45 44 96 72 69 63 56 Epilobium angustifolium L. 0 6 0 8 9 0 14 12 0 2 2 0 4 4 12 13 5

Equisetum arvense L. 16 18 32 5 9 5 11 19 5 13 30 20 9 8 8 7 15 Erysimum cheiranthoides L. 19 38 39 20 40 80 69 96 29 19 18 15 39 44 46 27 35 Fallopia convolvulus (L.) À. Löve 51 62 39 68 54 45 26 8 78 38 70 68 96 28 12 63 53 Fumaria officinalisL. 58 64 39 43 63 60 66 42 29 35 53 69 65 36 27 17 49 Galeopsis L. spp.a 72 72 29 70 74 80 60 58 51 50 76 54 70 48 85 70 64

Galium spurium L.b 77 48 71 78 77 20 31 19 76 52 71 64 74 16 23 13 55 Gnaphalium uliginom L. 5 36 0 30 37 75 66 88 2 25 0 8 43 64 58 17 29 Juncus bufonius L. 0 0 0 3 14 20 0 23 0 2 0 0 0 16 12 0 4 Lamium L. spp.c 42 28 93 35 54 10 14 0 20 60 68 56 48 24 8 7 39 Lapsana communis L. 51 78 71 65 91 70 86 8 20 79 41 68 91 80 19 3 58

Matricaria matricarioides (Less.) Porter 16 20 7 25 29 55 57 58 17 21 12 19 52 52 58 47 29 Myosotis arvensis (L.) Hill 42 58 32 38 66 70 69 54 39 40 44 39 57 76 38 33 48 Persicaria lapathifolia (L.) Gray 21 36 14 43 26 50 46 31 17 15 44 34 57 48 62 60 36 Phleum pratense L. 2 6 0 3 3 25 3 12 12 4 2 5 0 12 15 7 6 Plantago major L. 5 18 7 5 11 55 31 42 5 8 5 14 13 24 4 10 14

Poa annua L. 12 14 14 25 14 45 71 65 29 33 21 20 9 44 62 60 31 Poa pratensis L. 2 14 0 5 3 35 11 19 7 13 8 2 0 12 4 3 8 Polygonum aviculare L. 35 40 36 53 57 65 57 58 44 60 61 59 26 40 62 73 52 Ranunculus repens L. 7 26 7 8 29 65 40 69 7 4 9 19 26 24 50 20 22 Sonchus arvensis L. 40 66 54 53 49 30 46 31 27 27 21 37 48 36 35 33 39

Sonchus asper (L.) Hill 7 10 0 5 0 0 0 0 5 0 20 5 0 0 0 0 5 Spergula arvensis L. 16 46 32 53 34 85 71 62 22 44 24 17 78 56 88 43 43 Stellaria media (L.) Vill. 86 54 50 83 66 80 57 58 78 73 82 47 78 60 73 80 69 Taraxacum Weber spp. 53 64 14 40 60 70 17 81 32 6 24 63 57 24 8 17 39 Thlaspi arvense L. 14 16 11 5 26 15 0 0 10 0 2 3 22 4 12 10 8

Trifolium L. spp.d 35 54 7 40 31 65 57 77 12 15 27 12 13 36 23 17 31 Tripleurospermum inodorum (L.) Sch. Bip. 63 62 57 40 60 30 43 19 17 23 53 42 43 12 31 30 41 Veronica L. spp.e 2 2 14 5 9 25 6 38 10 15 2 15 22 28 15 0 11 Vicia L. spp.f 12 32 11 18 17 15 9 4 7 10 14 10 26 32 12 3 14 Viola arvensis Murrayg 79 98 89 85 89 100 91 92 83 98 70 69 83 96 88 83 85 amainly G. bifida and G. speciosa, bincl. G. aparine, cmainly L. purpureum, dmainly T. repens, emainly V. serpyllifolia, fmainly V. cracca, gincl. V. tricolor 250 251 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland Vol. 20(2011): 245–261.

In rarefaction, the number of species of larger in conventionally farmed fields, 12 (SD = 4.7., min samples is scaled down to the given number of in- 3, max 26). dividuals, which permits the comparison of species numbers among samples differing in size. Since the lowest numbers of individuals observed were Frequency of weed species 23081, 5316 and 37403 for zone, region and production type, respectively, we scaled sample sizes down to 23000, 5000 and 37000, respectively. Data The occurrence of the 40 most frequent weed from the ten 0.1 m2 sample quadrats were pooled species in the small 0.1 m2 sample quadrats is before the calculation. presented by region (Table 2) and the remaining species observed from the larger 1.0 m2 quadrats are listed by region (Appendix 1). The most common weeds were broad-leaved species, including Results

Table 3. Observed (SOBS) and rarefied (E(Sn) and SD) Species richness number of species by zone, region and production type.

a In total 175 weed species were found in the large SOBS E(Sn) SD 2 (1.0 m ) sampling quadrats and 148 in the small Zone 2 (0.1 m ) quadrats (see Appendix 1 for the species South 131 113 2.91 recorded in large quadrats but not included in the East 105 101 1.81 list of 40 most frequent species). The majority (i.e. West 84 84 0.16 110 species) of the observed species occurred in Region less than 5% of the fields studied. Altogether 104 species were observed in organically cropped fields Jokioinen 49 48.4 0.69 (N = 72) and 133 species in conventionally cropped, Lammi 64 60.1 1.55 herbicide-treated fields (N = 503). Nauvo/Korppoo 48 47.5 0.71 Tammela 63 59.1 1.60 The total number of observed species, SOBS, in regions ranged from 48 to 76 (Table 3). In four Iitti 64 58.5 1.82 regions (Jokioinen, Nauvo/Korppoo, Laihia and Kitee 63 61.3 1.18 Laitila), the number of observed species was be- Laukaa/Toivakka 63 55.5 2.01 low 50 and in one region (Nurmijärvi) it exceeded Vieremä 57 57.0 0.17 70. The same regions had the lowest and the high- Laihia 49 48.3 0.78

est expected number of species, E(Sn), calculated Laitila 49 45.7 1.44 by rarefaction analysis (Table 3). The total number Paimio/Tarvasjoki 62 56.2 1.86 of species was greatest in the southern and lowest Nurmijärvi 76 71.5 1.75 in the western zone. The species composition of Imatra/Ruokolahti 64 60.6 1.42 the southern zone was also most dissimilar among Mikkeli mlk 65 62.9 1.28 zones (Jaccard’s similarity: east vs. south = 0.411, Kihniö/Parkano 56 54.1 1.22 south vs. west 0.311 and east vs. west 0.59). Nivala 51 50.6 0.64 Surprisingly, the total species number for conventionally farmed fields exceeded that for organi- Production type cally farmed fields (Table 3). However, the average Organic 105 105 0.33 observed species number was higher in organically Conventional 135 123 2.63 farmed fields, 21 (SD = 4.5, min 11, max 34), than aThe number of individuals used in the analysis: zone 23000, region 5000 and production type 37000

250 251 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland

Viola arvensis, Stellaria media and Galeopsis spp., more frequent (p < 0.001) in organic than in con- which were frequently found both in convention- ventional farming. ally and organically cropped fields. Nevertheless, Some weed species were characteristic of cer- there were evident differences in the ranking list tain regions or zones in Finland; e.g. Lamium spp. of predominant weed species and their frequency and Galium spurium were most frequently found levels between conventionally and organically in southern regions, whereas the frequency of Ely- cropped fields (Table 4). mus repens, Gnaphalium uliginosum, Matricaria With a frequency of 56%, Elymus repens was matricarioides, Poa annua and Spergula arvensis by far the most common grass weed (Table 4). The increased towards the north. One of the most com- next most common grass species were Poa annua mon species, Lapsana communis, was relatively (31%), Poa pratensis (8%), Phleum pratense (6%) rare in the regions of western Finland. and Alopecurus geniculatus (4%). Altogether 23 grass species were identified in the sample quadrats (1.0 m2). Weed density Highly productive perennial broad-leaved weed species occurred more frequently (p < 0.001) in organic than in conventional farming; Sonchus The average density of weeds in all fields surveyed arvensis had a frequency of 61% in organic and was 209 plants m-2 (SD = 211, median = 138, N = 35% in conventional fields and Cirsium arvense 595) (Table 5). However, the difference between 46% and 25%, respectively. Typical grassland the two cropping systems was considerable; the weed species such as Achillea millefolium 26% average density of weeds in sprayed conventional vs. 1% and Ranunculus repens 47% vs. 17% were fields was 160 plants m-2 (SD = 155, median = 112,

Table 4. Frequency of ten most common weed species in two cropping systems and the change in frequency-% from 1997–1999 to 2007–2009. For comparison, additional four species from the other cropping system are included. Sprayed conventional (N = 503) Organic (N = 72) Rank Weed species % Change Weed species % Change 1 Viola arvensis 83 +2 Chenopodium album 96* 0 2 Stellaria media 65 0 Stellaria media 94* -1 3 Galeopsis spp. 59 -1 Viola arvensis 94* +1 4 Galium spurium 59 +16 Elymus repens 89* +8 5 Lapsana communis 57 +5 Spergula arvensis 89* +6 6 Fallopia convolvulus 53 +5 Galeopsis spp. 88* -5 7 Chenopodium album 52 -1 Erysimum cheiranthoides 86* +4 8 Elymus repens 50 -9 Myosotis arvensis 72* +12 9 Fumaria officinalis 48 +9 Polygonum aviculare 72* +2 10 Polygonum aviculare 48 -2 Persicaria lapathifolia 68* +15 Spergula arvensis 34 -1 Galium spurium 39* +2 Erysimum cheiranthoides 25 -4 Lapsana communis 60 +3 Myosotis arvensis 44 +4 Fallopia convolvulus 58 -5 Persicaria lapathifolia 30 -3 Fumaria officinalis 58 +17 * indicates significant difference in species frequencies between cropping systems (Fisher's Exact Test, p < 0.05) 252 253 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland Vol. 20(2011): 245–261.

Table 5. Weed density (plants m-2) by regions. + indicates < 1 plant m-2 Year / Region Species / Taxon 2007 2008 2009 Average 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 2007–09 Achillea millefolium 0 + 0 0 + 2 + 0 + + + 1 + + + 0 + Alopecurus geniculatus + 0 0 0 + 1 2 2 + 0 2 + + + + + + Artemisia vulgaris 0 + 0 + 0 + 0 0 0 + 0 + + + + 0 + Brassica rapa ssp. oleifera. + + + + + + + 0 + + 0 + + 1 + + + Capsella bursa-pastoris + + 2 1 4 13 7 2 + 2 + + 7 7 + 4 3

Cerastium fontanum + + 0 + + 3 0 + 0 0 0 + 1 + + 0 + Chenopodium album 15 30 26 27 12 39 11 8 36 32 19 10 39 55 7 27 23 Cirsium arvense + 2 + 1 1 + + + + + 3 + 1 + + + + Elymus repens 8 17 16 15 27 21 18 27 16 10 11 13 73 32 51 17 20 Epilobium angustifolium 0 + 0 + + 0 + + 0 + + 0 + + + + +

Equisetum arvense + 1 2 + + + + + + + + 1 1 + + + + Erysimum cheiranthoides + 2 4 + 5 19 7 9 1 2 2 + 6 8 15 3 4 Fallopia convolvulus 6 4 2 7 4 2 1 + 5 2 5 6 6 2 + 4 4 Fumaria officinalis 4 3 2 3 7 3 9 2 1 2 6 7 17 + 1 1 4 Galeopsis spp. 7 9 11 8 13 8 7 9 9 8 13 21 5 13 8 12 11

Galium spurium 20 12 14 13 12 1 2 2 12 18 8 9 5 1 + + 9 Gnaphalium uliginosum + 3 0 8 21 40 16 14 + + 0 2 3 5 8 + 6 Juncus bufonius 0 0 0 + 3 3 0 + 0 + 0 0 0 3 + 0 + Lamium spp. 7 3 28 6 6 + 8 0 2 6 24 5 4 + + + 7 Lapsana communis 13 16 9 11 34 16 22 + 1 26 2 11 50 10 3 + 13

Matricaria matricarioides + 2 + 1 7 7 18 7 + + + 3 12 9 5 4 4 Myosotis arvensis 2 4 + 1 22 21 7 2 1 1 3 2 9 6 2 4 5 Persicaria lapathifolia 2 8 12 4 6 3 16 1 + + 6 2 5 2 8 7 5 Phleum pratense + 1 0 + + 2 + 2 + + + + 0 4 + 2 + Plantago major + + + + + 7 5 2 + + + 3 + 3 + + 1

Poa annua 5 2 + 18 6 4 56 16 7 27 3 2 + 12 8 17 11 Poa pratensis + 1 0 + + 3 6 + + 2 + + 0 + + + + Polygonum aviculare + + + 2 4 3 2 1 2 1 4 3 1 1 2 3 2 Ranunculus repens + + + + + 3 2 4 + + + + + + 21 + 2 Sonchus arvensis 2 5 4 6 5 + 6 4 3 5 1 1 12 4 + + 4

Sonchus asper + + 0 + 0 0 0 0 + 0 + + 0 0 0 0 + Spergula arvensis 3 8 15 7 16 12 28 27 1 2 3 4 42 29 60 13 13 Stellaria media 13 4 14 22 7 12 5 7 13 19 22 4 14 11 10 14 12 Taraxacum spp. 6 3 2 1 11 3 + 4 + + + 5 4 2 + 1 3 Thlaspi arvense 2 + + + 5 2 0 0 + 0 + + 8 + + 2 1

Trifolium spp. 4 4 + 2 1 2 2 7 + + + + + 2 5 1 2 Tripleurospermum inodorum 3 3 3 2 11 4 4 + + + 2 1 9 + + + 3 Veronica spp. + + 2 + + 1 + 2 + 4 + 1 1 2 + 0 + Vicia spp. + 2 1 1 + + + + + 2 + + 4 3 + + + Viola arvensis 17 18 20 24 34 41 36 22 23 50 10 13 28 21 30 33 25

Mean total, plants m-2 150 175 199 204 293 311 319 193 143 228 158 140 387 266 257 177 209 Median total, plants m-2 114 118 104 159 199 301 200 130 110 153 109 90 324 185 182 86 138 252 253 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland

Table 6. Weed biomass (kg ha-1) by regions. + indicates < 1 kg ha-1 Year / Region Species / Taxon 2007 2008 2009 Average 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 2007–09 Achillea millefolium 0 3 + 0 + 2 0 0 0 0 0 6 + 0 0 0 + Alopecurus geniculatus 0 0 0 0 + 2 3 1 2 0 2 3 + 0 + + 1 Artemisia vulgaris 0 + 0 0 0 + 0 0 0 + 0 1 0 0 0 0 + Brassica rapa ssp. oleifera 0 0 0 0 + 2 0 0 1 + 0 0 1 10 0 3 + Capsella bursa-pastoris + 0 + + 3 2 2 6 + 3 0 + 2 4 + 4 1

Cerastium fontanum 0 0 0 0 + + 0 0 0 0 0 + 0 0 0 0 + Chenopodium album 11 25 17 46 37 60 8 5 51 29 12 8 31 104 4 40 27 Cirsium arvense 8 6 0 7 29 1 2 0 7 3 18 5 8 23 + 2 8 Elymus repens 19 63 34 95 72 50 65 102 79 45 38 49 217 127 196 61 72 Epilobium angustifolium 0 0 0 0 0 0 + 0 0 0 0 0 0 0 0 + +

Equisetum arvense + 6 7 + 1 0 + + + 1 3 11 9 + 0 + 3 Erysimum cheiranthoides + + + 1 1 12 4 2 1 3 + + + 4 4 2 2 Fallopia convolvulus 3 3 2 7 7 4 1 0 5 3 2 5 9 3 + 4 4 Fumaria officinalis 2 2 1 2 7 3 9 + 3 1 3 7 19 1 0 3 4 Galeopsis spp. 8 11 2 32 21 20 12 11 30 32 9 20 9 20 13 13 17

Galium spurium 14 17 16 10 10 + 3 + 13 17 4 4 1 0 10 + 8 Gnaphalium uliginosum + + 0 + + + + 0 0 + 0 0 + + 3 + + Juncus bufonius 0 0 0 + 2 5 0 + 0 0 0 0 0 + 0 0 + Lamium spp. 3 + 10 2 7 0 10 0 1 3 6 3 2 + 0 + 3 Lapsana communis 4 12 4 11 12 12 10 0 1 16 + 6 53 8 1 0 8

Matricaria matricarioides + + 0 + 5 2 7 1 + + + + 2 6 + 4 2 Myosotis arvensis 1 + + + 4 8 3 + + + + + 2 1 + 1 1 Persicaria lapathifolia + 14 3 11 14 4 11 1 + + 9 2 6 5 11 22 7 Phleum pratense 0 12 0 0 0 2 1 0 8 0 13 1 0 3 4 0 4 Plantago major + + 0 0 + 5 + 0 0 + 0 + + 4 0 1 +

Poa annua + 0 0 6 3 + 11 9 12 21 + 2 + 5 2 24 6 Poa pratensis 0 + 0 + 0 3 + 4 1 + + + 0 + 0 + + Polygonum aviculare + + + 5 7 6 3 + 1 + 8 4 + 2 + 8 3 Ranunculus repens 0 + + + 0 2 + 2 0 0 + + + 0 + + + Sonchus arvensis 3 9 0 29 27 + 40 13 7 15 2 2 32 14 4 1 12

Sonchus asper + + 0 + 0 0 0 0 0 0 + + 0 0 0 0 + Spergula arvensis 1 3 31 8 16 15 33 65 3 1 2 7 44 50 89 13 18 Stellaria media 4 + 7 25 8 12 2 4 8 34 9 5 7 23 19 14 11 Taraxacum spp. + 3 + + 3 4 + + + 0 + 6 4 + 0 + 1 Thlaspi arvense + 0 0 0 3 5 0 0 + 0 0 + 6 + + + +

Trifolium spp. + + 0 + 3 + + 0 + + + 2 0 0 0 0 + Tripleurospermum inodorum + 3 0 3 4 1 1 0 6 + + 4 3 0 0 + 2 Veronica spp. + 0 1 0 0 0 + 1 + 5 + + 0 + 0 0 + Vicia spp. + 2 2 3 2 0 4 0 0 2 + 1 4 7 0 + 2 Viola arvensis 4 7 + 11 11 17 10 7 11 26 1 3 7 8 4 12 8

Mean total, kg ha-1 93 212 140 338 333 272 263 247 255 288 148 189 494 449 377 241 251 Median total, kg ha-1 28 84 36 141 195 220 155 163 153 144 61 88 335 195 248 109 110

254 255 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland Vol. 20(2011): 245–261.

Table 6. Weed biomass (kg ha-1) by regions. + indicates < 1 kg ha-1 N = 503) and in organic production 519 plants m-2 Sonchus arvensis (52) and Galeopsis spp. (39). Year / Region (SD = 244, median = 468, N = 72). The total weed Correspondingly, the five most productive weed Species / Taxon 2007 2008 2009 Average densities were significantly lower (p < 0.05) in species in sprayed conventional fields were Ely- 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 2007–09 the regions with intensive cereal production (e.g. mus repens (46), Galeopsis spp. (13), Chenopo- Achillea millefolium 0 3 + 0 + 2 0 0 0 0 0 6 + 0 0 0 + regions 1, 4, 9, 12) compared with regions (e.g. 6, dium album (13), Viola arvensis (9) and Lapsana Alopecurus geniculatus 0 0 0 0 + 2 3 1 2 0 2 3 + 0 + + 1 7, 13) where cattle farms and organic production communis (8). Artemisia vulgaris 0 + 0 0 0 + 0 0 0 + 0 1 0 0 0 0 + predominated. Although the number of recorded weed spe- Brassica rapa ssp. oleifera 0 0 0 0 + 2 0 0 1 + 0 0 1 10 0 3 + Capsella bursa-pastoris + 0 + + 3 2 2 6 + 3 0 + 2 4 + 4 1 The five most abundant weed species in organ- cies was relatively high, the number of species ic fields were Chenopodium album (on average mainly contributing to the biomass production Cerastium fontanum 0 0 0 0 + + 0 0 0 0 0 + 0 0 0 0 + 107 plants m-2), Elymus repens (68), Spergula ar- was relatively low. The ten most abundant weed Chenopodium album 11 25 17 46 37 60 8 5 51 29 12 8 31 104 4 40 27 vensis (50), Viola arvensis (31) and Stellaria me- species accounted for 72% of the weed biomass Cirsium arvense 8 6 0 7 29 1 2 0 7 3 18 5 8 23 + 2 8 dia (26). Correspondingly, the five most abundant production in the fields of the southern survey Elymus repens 19 63 34 95 72 50 65 102 79 45 38 49 217 127 196 61 72 weed species in sprayed conventional fields were zone, 87% in the western zone and 80% in the Epilobium angustifolium 0 0 0 0 0 0 + 0 0 0 0 0 0 0 0 + + Viola arvensis (24), Lapsana communis (13), Poa eastern zone. Equisetum arvense + 6 7 + 1 0 + + + 1 3 11 9 + 0 + 3 annua (13), Elymus repens (12) and Chenopodium Elymus repens was the most efficient biomass Erysimum cheiranthoides + + + 1 1 12 4 2 1 3 + + + 4 4 2 2 album (11). producer, accounting for 31% of the total weed Fallopia convolvulus 3 3 2 7 7 4 1 0 5 3 2 5 9 3 + 4 4 Although not revealed in observations from biomass production pooled over all organic fields Fumaria officinalis 2 2 1 2 7 3 9 + 3 1 3 7 19 1 0 3 4 sample quadrats, the substantially increased and for 28% in sprayed conventional fields (Fig. Galeopsis spp. 8 11 2 32 21 20 12 11 30 32 9 20 9 20 13 13 17 number of fields infested withAvena fatua was an 2). Galium spurium 14 17 16 10 10 + 3 + 13 17 4 4 1 0 10 + 8 evident discovery of the survey. Typically, some The proportion of weed biomass relative to Gnaphalium uliginosum + + 0 + + + + 0 0 + 0 0 + + 3 + + individuals of A. fatua grew in a few patches in total vegetative biomass (crop + weeds) was Juncus bufonius 0 0 0 + 2 5 0 + 0 0 0 0 0 + 0 0 + infested fields. The information from farmers’ in- relatively low (mean = 2.9%, median=1.2%) in Lamium spp. 3 + 10 2 7 0 10 0 1 3 6 3 2 + 0 + 3 terviews provided support to this conclusion as sprayed conventional fields, but considerably Lapsana communis 4 12 4 11 12 12 10 0 1 16 + 6 53 8 1 0 8 they classified almost 30% (168 fields) of fields higher (mean = 21.3%, median=16.0%) in organic Matricaria matricarioides + + 0 + 5 2 7 1 + + + + 2 6 + 4 2 as infested and we detected even more fields. In fields. For conventional cropping the proportion Myosotis arvensis 1 + + + 4 8 3 + + + + + 2 1 + 1 1 comparison, the proportion of surveyed fields in- of weed biomass was less than 5% in 84% of the Persicaria lapathifolia + 14 3 11 14 4 11 1 + + 9 2 6 5 11 22 7 fested with A. fatua was only around 10% (74 survey fields whereas the majority (89%) of or- Phleum pratense 0 12 0 0 0 2 1 0 8 0 13 1 0 3 4 0 4 fields out of 690 fields) ten years ago. ganically cropped fields fell into classes above Plantago major + + 0 0 + 5 + 0 0 + 0 + + 4 0 1 + 5% infestation (Fig. 3). In most cases the high relative proportion of weeds was due to abundant Poa annua + 0 0 6 3 + 11 9 12 21 + 2 + 5 2 24 6 infestation of Elymus repens. Poa pratensis 0 + 0 + 0 3 + 4 1 + + + 0 + 0 + + Weed biomass Polygonum aviculare + + + 5 7 6 3 + 1 + 8 4 + 2 + 8 3 Ranunculus repens 0 + + + 0 2 + 2 0 0 + + + 0 + + + Sonchus arvensis 3 9 0 29 27 + 40 13 7 15 2 2 32 14 4 1 12 The average biomass production of weeds was 251 Seed bank kg ha-1 (SD = 382, median = 110, N = 595) (Table Sonchus asper + + 0 + 0 0 0 0 0 0 + + 0 0 0 0 + 6). The difference between cropping practices was Spergula arvensis 1 3 31 8 16 15 33 65 3 1 2 7 44 50 89 13 18 as clear as in the case of weed densities, in that Altogether, 27 species were encountered in seed Stellaria media 4 + 7 25 8 12 2 4 8 34 9 5 7 23 19 14 11 for sprayed conventional fields the average weed bank samples (Table 7). Only seven species ex- Taraxacum spp. + 3 + + 3 4 + + + 0 + 6 4 + 0 + 1 biomass was 167 kg ha-1 (SD = 239, median = 83, ceeded the frequency level of 10%, Chenopodium Thlaspi arvense + 0 0 0 3 5 0 0 + 0 0 + 6 + + + + N = 503) and for organic production it was more album and Fallopia convolvulus being the most Trifolium spp. + + 0 + 3 + + 0 + + + 2 0 0 0 0 + than four times higher, namely 775 kg ha-1 (SD = common species. The average size of the seed bank Tripleurospermum inodorum + 3 0 3 4 1 1 0 6 + + 4 3 0 0 + 2 590, median = 563, N = 72). in the 5 cm surface layer was 1684 seed m-2. C. Veronica spp. + 0 1 0 0 0 + 1 + 5 + + 0 + 0 0 + The five weed species producing the highest album was the most abundant species, comprising Vicia spp. + 2 2 3 2 0 4 0 0 2 + 1 4 7 0 + 2 amounts of biomass in organic fields wereElymus 66.2% of the total seed sample and having 1115 Viola arvensis 4 7 + 11 11 17 10 7 11 26 1 3 7 8 4 12 8 repens (average air-dry weight 242 kg ha-1), Che- seeds per m-2, followed by Spergula arvensis and Mean total, kg ha-1 93 212 140 338 333 272 263 247 255 288 148 189 494 449 377 241 251 nopodium album (115), Spergula arvensis (70), F. convolvulus. Median total, kg ha-1 28 84 36 141 195 220 155 163 153 144 61 88 335 195 248 109 110

254 255 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland

Species rank Sprayed conventional (N=503) SONAR POAAN SPRAR GALSP STEME LAPCO VIOAR CHEAL GAESS AGRRE 0 5 10 15 20 25 30 35 Fig. 2. Proportion of the total Proportion, % weed biomass production by the most abundant species in conventional and organic fields. Species rank Organic (N=72) Key of EPPO code abbrevia- POLAV tions: AGRRE Elymus repens, AVEFA AVEFA Avena fatua, CHEAL CIRAR Chenopodium album, CIRAR POLLA Cirsium arvense, GAESS, Galeopsis spp., GALSP Galium STEME spurium, LAPCO Lapsana com- GAESS munis, POAAN Poa annua, SONAR POLAV Polygonum aviculare, SPRAR POLLA Persicaria lapathifo- CHEAL lia, SONAR Sonchus arven- AGRRE sis, SPRAR Spergula arvensis, STEME Stellaria media, 0 5 10 15 20 25 30 35 VIOAR Viola arvensis. Proportion, %

Fields (%) 50

10 Fig. 3. Relative number of fields (%) in different weed biomass 0 0-1 1-5 5-10 10-25 >25 classes in two cropping systems. Weed biomass is proportioned to Proportion of weed biomass, % total vegetative biomass (crop Conventional Organic + weeds).

256 257 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland Vol. 20(2011): 245–261.

Table 7. Occurrence of weed seeds in the top-soil (0–5 cm) of the survey fields. Frequency of occurrence Proportion of total sample Seeds Weed species (%) (%) m-2 Agrostis capillaris 0.5 0.1 1 Alopecurus geniculatus 6.1 1.6 26 Brassica rapa 0.5 0.1 1 Capsella bursa-pastoris 0.3 0.1 2 Chenopodium album 79.8 66.2 1115 Cirsium arvense 0.5 <0.1 1 Elymus repens 3.9 0.6 10 Fallopia convolvulus 41.5 8.0 136 Fumaria officinalis 1.5 0.1 2 Galeopsis spp. 23.4 3.8 64 Galium album 0.2 <0.1 <1 Galium spurium 17.0 2.6 44 Lapsana communis 1.2 0.1 2 Myosotis arvensis 6.2 1.2 20 Persicaria lapathifolia 13.8 1.8 30 Phleum pratense 2.0 0.2 4 Poa annua 0.3 0.1 1 Polygonum aviculare 12.1 1.9 32 Ranunculus repens 0.3 <0.1 1 Rumex longifolius 0.2 <0.1 <1 Sonchus asper 0.2 <0.1 <1 Spergula arvensis 10.1 10.6 179 Stellaria media 3.9 0.6 9 Taraxacum officinale 0.2 0 <1 Thlaspi arvense 1.0 0.1 1 Vicia cracca 1.0 0.1 1 Viola arvensis 1.3 0.3 4

Discussion

The weed survey in 2007–2009 can be regarded 100 000 ha ten years ago (TIKE 2009). The areas as a follow-up study for the previous survey car- of spring barley (52% of spring cereal area) and ried out in 1997–1999 in the same 16 regions and oats (30%) remain much higher than that of spring mainly at the same farms and fields (Salonen et wheat. A marked shift towards winter cereal crop- al. 2001a). Moreover, the monitoring protocol was ping, as in Denmark (Andreasen and Stryhn 2008), exactly the same as in the 1990s. Spring cereals, has not occurred in Finland. covering about 50–55% of the arable land, still The number of weed species, 148, found in dominate crop production in Finland and the only small sample quadrats (0.1 m2) was slightly lower major shift in cereal cropping during the last ten than ten years ago when 160 species were record- years has been the doubled area of spring wheat, ed (Salonen et al. 2001a). In the larger (1.0 m2) reaching almost 200 000 ha, having been around quadrats the number of observed species was now 256 257 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland

175 versus 188 species ten years ago. An obvious ping. The impact of intensity is an obvious reason reason for reduction in floral diversity is that the also for the tremendous decline in the number of total number of surveyed fields was lower (595 seeds in the soil between the present survey (1168 vs. 690). Particularly the number of organically seeds per m-2) and the 1960s survey (43850 seeds grown fields was less than half compared with the per m-2) (Paatela and Erviö 1971). Such a differ- 1997–1999 survey. In comparison, the total num- ence is partly due to the difference in the soil sam- ber of weed species found in spring cereal fields ple, which was taken from a 5 cm top layer in our in the Czech Republic was 101, with an average study and from a 20 cm layer in the 1960s. How- of 13 species per field (Tyšer & Kolárová 2010). ever, Chenopodium album was the most common Much lower total numbers of weeds, 64 species, and abundant species in both decades. were found in arable fields in Alaska, at a similar The proportion of organically cropped survey latitude as in Finland (Conn et al. 2011), whereas fields was relatively high in certain regions such as the species number recorded in Danish arable fields Kitee, Kihniö/Parkano and Imatra/Ruokolahti and in 2001–2004 was as high as 225 (Andreasen and this is reflected to some extent in the results, for Stryhn 2008). instance as a higher frequency of Spergula arvensis Only 38 weed species or taxa exceeded the in those regions. On the other hand, some common overall frequency level of 5%. Among the most species were relatively rare in certain zones, like common weed species, Galium spurium, (+16% Lapsana communis in western Finland, as discov- units increase since the 1990s), Fumaria officinalis ered already in the 1960s (Mukula et al. 1969). The (+9), Lapsana communis (+5) and Fallopia convol- regional specialization is a complex of many fac- vulus (+5) have become more frequent, whereas tors, including prevailing crop rotations, soil types, Elymus repens (-9) has become less frequent since tillage practices, proportion of organic farming etc., 1997–1999. The newcomers in the list of 40 most still existing and affecting the specific weed occur- frequent species were Alopecurus geniculatus, rence in different regions. Artemisia vulgaris, Phleum pratense, Poa praten- Galium spurium (probably including some pop- sis and Sonchus asper. Appearance of new grass ulations of Galium aparine) is a typical example of species in the list is evidently a consequence of this ‘regionalization’ as it has become very frequent increased proportion of survey fields with reduced (66% in southern vs. 35–37% in eastern and west- tillage, which favours grasses, as demonstrated in ern zones) in the southern survey zone, which is other studies (e.g. Gruber et al. 2000, Jalli et al. characterized by clay soils, monoculture of cereals 2006, Tørresen et al. 2006, Thomas et al. 2011). and reduced tillage. For instance, reduced tillage Moreover, it has been demonstrated that reduced practices seem to favour Galium species (Tørresen tillage has a major, although complex, role in pro- & Skuterud 2002, Thomas et al. 2011). In addition, moting weed species diversity (Murphy et al. 2006) a long-lasting use of first generation sulphonylurea and in determining the weed community compo- herbicides with weak efficacy againstGalium spe- sition (Légère et al. 2005). In our survey fields, cies has been an evident promoting factor, although the proportion of non-inversion tillage (i.e. only Hyvönen et al. (2003b) found only weak statistical tine-cultivated or direct-drilled) had increased from support for the selection pressure on weed commu- 14% in 1997–1999 to 44 % (27% tine-cultivated nities due to the application of sulphonylureas. This and 17% direct-drilled) in 2007–2009. aspect is worth re-analyzing by further dissecting Overall, the frequencies of occurrence for the the survey data series and making comparisons be- most common weed species were significantly tween the decades now that sulphonylureas have higher in organic cropping than in conventional largely replaced phenoxy acid herbicides. cropping. Evidently this is partly due to the success Elymus repens is the most harmful weed spe- of chemical weed control, but is also a consequence cies for spring cereals in Finland, producing about of more intensive crop management resulting in 30% of weed biomass in both cropping systems. more competitive crop stands in conventional crop- In conventional cropping, farmers seemingly re- 258 259 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland Vol. 20(2011): 245–261.

spond to the situation as the sales of glyphosate enhancement of biodiversity with rich and ad- have more than doubled in ten years (Londesbor- vantageous flora should be primarily realized by ough et al. 2000, Evira 2010). Approximately 60% establishing specific flower-strips (e.g. van Elsen of the direct-drilled survey fields had been treated and Hotze 2008), “ecological compensation areas” with glyphosate in the spring before sowing and (Aviron et al. 2009) or by introducing completely the annual application of glyphosate before sowing new incentives in which farmers commit them- time was a common practice in these fields. Like- selves to conservation of weed species by reduc- wise, the decline of E. repens has been observed ing the chemical inputs, primarily fertilizers and in Denmark and associated with increased use of herbicides (Ulber et al. 2010). glyphosate (Andreasen and Stryhn 2008). New grass weed species were introduced on to The abundance of weeds in organic cropping is the list of most frequent species, possibly reflect- slowly approaching the infestation level of 1 000 ing the changes in tillage practices towards non- kg ha-1 recorded in unsprayed conventionally inversion soil cultivation. The frequent occurrence cropped fields in the 1960s (Mukula 1974). There- of Avena fatua is alarming and requires prompt fore, measures promoting crop competition, as well action to manage the problem. As suggested by as reliance on physical weed control should be em- Conn et al. (2011), research and implementation phasized in organic cropping (Kolb et al. 2010). of prevention programmes should be emphasized For instance, the abundance of E. repens in organic as the costs of controlling an outbreak are much cropping is a cause for some concern and calls for higher than for preventing one. Likewise, the suc- improved control strategies. Contradictory findings cess of Poa annua is worth taking note of as its on the success of E. repens in organic cropping frequency (31%) had significantly increased in ten over time have been reported by Becker and Hurle years, having been 12% in 1997–1999 (Salonen et (1998), who demonstrated decreased frequency of al. 2001a). The success of P. annua is in line with E. repens, whereas Riesinger and Hyvönen (2006b) observations in the Nordic countries (Tørresen et found increased abundances as a function of the al. 2006, Andreasen and Stryhn 2008). Among the duration of organic farming. In general, E. repens broad-leaved weed species, the success of Galium is probably going to benefit from climate change, spurium in conventional cropping, as well as Fu- which could extend its period of autumn growth maria officinalis, Myosotis arvensis and Persicaria (Tørresen et al. 2010). lapathifolia in organic cropping, should be regard- The area of organic farming in Finland has ed as a weed shift. levelled out at around 6–7% of arable land (TIKE In conclusion, although the weed flora in Finn- 2009) and subsidized by agri-environment policy ish spring cereal fields consists of numerous weed this has successfully promoted diverse and abun- species, only a fraction of them severely threaten dant weed flora to the benefit of biological diver- crop production at national level in terms of fre- sity. However, the success of E. repens in organic quency and abundance. A wider range of grass cropping should not be regarded as a positive trend weed species may interfere with the production of because it is a very competitive species that greatly cereals in the future, particularly if there will be a reduces crop yield and has only little benefit for shift towards winter cereal cropping with reduced biodiversity, and actually mainly benefits harmful tillage. Thus, new grass weed management strate- phytophagous insects (Hyvönen and Huusela-Veis- gies, both preventive and curative, should be em- tola 2008). As regards biological diversity, other phasized. In general, the range of herbicides with frequently observed perennial weed species such different modes of action (mainly sulphonylureas, as Cirsium arvense and Sonchus arvensis are more phenoxy acids and glyphosate) is limited on the favourable (Hyvönen and Huusela-Veistola 2008) Finnish market, and the risks of weeds evolving In conventional cropping, the relatively high herbicide resistance should be taken into account. efficacy of weed control and good competitive At present, weeds in conventional cropping can ability of spring cereals leads to a conclusion that be effectively managed with the current, relatively 258 259 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland

Fried, G., Norton, L.R. & Reboud, X. 2008. Environmen- intensive cropping measures and with available tal and management factors determining weed species herbicides whereas weed management in organic composition and diversity in France. Agriculture, Eco- cropping calls for urgent measures such as direct systems and Environment 128: 68–76. Gruber, H. Händel, K. & Broschewitz, B. 2000. Einfluß der mechanical weed control in crop stands, which was Wirtschaftsweise auf die Unkrautflora in Mähdrusch- not practiced at all in survey fields. This would früchten einer sechsfeldrigen Fruchtfolge. Zeitschrift likely benefit also species richness by diminishing für Pflanzenkrankheiten und Pflanzenschutz, Sonder- heft XVII, 33–40. the dominance of competitive perennial weeds. Hämet-Ahti, L., Suominen, J., Ulvinen, T. & Uotila, P. 1998. Retkeilykasvio (Field Flora of Finland), 4th ed. Finnish Acknowledgements. The weed survey was part of the Museum of Natural History, Botanical Museum. Hel- follow-up study on the impacts of agri-environment policy sinki. 656 p. measures (MYTVAS) and part-financed by the Ministry of Heck, K.L.J., van Belle, G. & Simberloff, D. 1975. Explic- Agriculture and Forestry. We thank all farmers who allowed it calculation of the rarefaction diversity measurement and the determination of sufficient sample size. Ecolo- us to visit their fields and provided us with information gy 56: 1459–1461. about their farming practices. Our permanent technical Hyvönen, T. 2007. Can conversion to organic farming re- staff, Eira-Maija Tanni, Anne Muotila and Marjo Segerstedt store the species composition of arable weed commu- and a team of field assistants Miko Häsä, Osmo Jalli, Mar- nities? Biological Conservation 137: 382–390. jaana Järveläinen, Saara Kinnunen, Sini Stenbacka and Hyvönen, T. & Huusela-Veistola, E. 2008. Arable weeds Tuija Vihervirta are acknowledged for their enthusiastic as indicators of agricultural intensity – A case study from Finland. Biological Conservation 141: 2857–2864. toil. Timo Hurme advised us in statistics. This paper is Hyvönen, T. & Salonen, J. 2003. Weed seedbank develop- dedicated to Professor Leila-Riitta Erviö who passed away ment under low-input and conventional cropping practic- in May 2011. es. Aspects of Applied Biology 69: 119–123. Hyvönen, T., Ketoja, E. & Salonen, J. 2003b. Changes in the abundance of weeds in spring cereals in Finland. Weed Research 43: 348–356. Hyvönen, T., Ketoja, E., Salonen, J., Jalli, H. & Tiainen, J. 2003a. Weed species diversity and community com- References position in organic and conventional cropping of spring cereals. Agriculture, Ecosystems and Environment 97: 131–149. Andreasen, C. & Skovgaard, I.M. 2009. Crop and soil fac- Jaccard, P. 1912. The distribution of the flora of the alpine tors of importance for the distribution of plant species zone. New Phytologist 11: 37– 50. on arable fields in Denmark. Agriculture, Ecosystems Jalli, H., Laine, A. & Känkänen, H. 2006. No-till cultivation and Environment 133: 61–67. suppresses broad-leaved weeds but favours grasses. Andreasen, C. & Streibig, J. 2011. Evaluation of changes in ‘Extended Abstracts’’ of the NJF Seminar 378, p. 72–77. weed flora in arable fields of Nordic countries – based on (available at: http://www.njf.nu/seminars/378/). Danish long-term surveys. Weed Research 51:214–226. Kolb, L.N., Gallandt, E.R. & Molloy, T. 2010. Improving Andreasen, C. & Stryhn, H. 2008. Increasing weed flora weed management in organic spring barley: physical in Danish arable fields and its importance for biodiver- weed control vs. interspecific competition. Weed Re- sity. Weed Research 48: 1–9. search 50: 597–605. Aviron, S., Nitsch, H., Jeanneret, P., Buholzer, S., Luka, H., Légère, A., Stevenson, F.C. & Benoit, D.L:.2005. Diver- Pfiffner, L, Pozzi, S., Schüpbach, B., Walter, T. & Her- sity and assembly of weed communities: contrasting zog, F. 2009. Ecological cross compliance promotes responses across cropping systems. Weed Research farmland biodiversity in Switzerland. Frontiers in Ecol- 45: 303–315. ogy and the Environment 7: 247–252. Londesborough, S., Hynninen, E.-L. & Blomqvist, H. Becker, B. & Hurle, K. 1998. Unkrautflora auf Feldern mit 2000. Pesticide sales in Finland in 1999. Kemia-Kemi unterschiedlich langer ökologischer Bewirtschaftung. 27: 492–494. Zeitschrift für Pflanzenkrankheiten und Pflanzenschutz, Milanova, S., Dimitrova, T., Valkova, M., Tachkov, J., Atan- Sonderheft XVI: 155–161. asova, L., Ilieva, L. & Christov, C. 2009. Weed infesta- Conn, J.S., Werdin-Pfisterer, N.R. & Beattie, K.L. 2011. De- tion of winter wheat in Pleven region, Bulgaria. Herbo- velopment of the Alaska agricultural weed flora 1981– logia 10: 1–11. 2004: a case for prevention. Weed Research 51: 63–70. Mukula, J. 1974. Weed competition in spring cereals in Fin- Erviö, L.-R. & Salonen, J. 1987. Changes in the weed pop- land. Forskning och forsøk i landbruket 25: 585–592. ulation of spring cereals in Finland. Annales Agriculturae Mukula, J. , Raatikainen, M., Lallukka, R. & Raatikainen, T. Fenniae 26: 210–226. 1969. Composition of weed flora in spring cereals in Fin- Evira 2010. Sales of agricultural herbicides. Available at: land. Annales Agriculturae Fenniae 8: 59–109. http://www.evira.fi/portal/en/plant_production_and_ Murphy, S.D., Clements, D.R:, Belaoussoff, S., Kevan, P.G. feeds/plant_protection_products/statistics/. Accessed & Swanton, C.J. 2006. Promotion of weed species di- 15 Jun 2010. versity and reduction of weed seedbanks with conserva-

260 261 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weed flora in spring cereals in Finland Vol. 20(2011): 245–261.

tion tillage and crop rotation. Weed Science 54: 69–77. Holm, F.A. & Gradin, B. 2011. Weed community re- Niemi, J. & Ahlstedt, J. (eds.) 2009. Finnish Agriculture and sponses to contrasting integrated weed management Rural Industries 2009. MTT Economic Research Publi- systems for cool dryland annual crops. Weed Rese- cations, No 109a. 96. p. arch 51: 41–50. Novak, R., Dancza, I., Szentey, L., Karamán, J., Béres, TIKE 2009. Maatilatilastollinen vuosikirja 2009 (Yearbook I., Kazinczi, G. & Gólya, G. 2010. ARABLE weeds of of farm statistics). Maa- ja metsätalousministeriön tie- Hungary. The fifth national weed survey (2007–2008). topalvelukeskus TIKE. 268 p. Proceedings of the 15th EWRS Symposium, Kaposvar, Tørresen, K.S., Fykse, H. & Rafoss, T. 2010. Autumn growth Hungary. p. 8–9. of Elymus repens, Cirsium arvense and Sonchus arven- Paatela, J. & Erviö, L-R. 1971. Weed seeds in cultivat- sis at high latitudes in an outdoor pot experiment. Weed ed soils in Finland. Annales Agriculturae Fenniae 10: Research 50: 353–363. 144–152. Tørresen, K.S., Salonen, J., Fogelfors, H., Håkansson, S. Potts, G.R., Ewald, J.A. & Aebischer, N.J. 2010. Long-term & Melander, B. 2006. Weed problems in various tillage changes in the flora of the cereal ecosystem on the Sus- systems in the Nordic countries. ‘Extended Abstracts’’ sex Downs, England, focusing on the years 1968–2005. of the NJF Seminar 378, p. 54–60. (available at: http:// Journal of Applied Ecology 47: 215–226. www.njf.nu/seminars/378/). Riesinger, P. & Hyvönen, T. 2006a. Weed occurrence in Tørresen, K.S. & Skuterud, R. 2002. Plant protection in Finnish coastal regions: a survey of organically cropped spring cereal production with reduced tillage. IV. Chang- spring cereals. Agricultural and Food Science 15: 166– es in the weed flora and weed seedbank. Crop Protec- 182. tion 21: 179–193. Riesinger, P. & Hyvönen, T. 2006b. Impact of management Tyšer, L. & Kolárová, M. 2010. Current weed spectrum in on weed species composition in organically cropped selected areas of the Czech Republic. Proceedings of spring cereals. Biological Agriculture and Horticulture the 15th EWRS Symposium, Kaposvar, Hungary. p. 133. 24: 257–274. Ulber, L., Klimek, S., Steinmann, H. & Isselstein, J. 2010. A Romero, A., Chamorro, L. & Xavier Sans, F. 2008. Weed market-based payment scheme for conservation of weed diversity in crop edges and inner fields of organic and species diversity in arable systems. Proceedings of the conventional dryland winter cereal crops in NE Spain. 15th EWRS Symposium, Kaposvar, Hungary. p. 122–123. Agriculture, Ecosystem and Environment 124: 97–104. van Elsen, T. 2000. Species diversity as a task for organic Salonen, J., Hyvönen, T. & Jalli, H. 2001a. Weeds in spring agriculture in Europe. Agriculture, Ecosystems and En- cereal fields in Finland – a third survey.Agricultural and vironment 77: 101–109. Food Science in Finland 10: 347–364. van Elsen, T. & Hotze, C. 2008. The integration of autoch- Salonen, J, Hyvönen, T. & Jalli, H. 2001b. Weed flora in thon arable field plants and the corncockle into flower organically grown spring cereals in Finland. Agricultural strips for organic farming. Journal of Plant Diseases and and Food Science in Finland 10: 231–242. Protection 21: 373–377. Sjursen, H. 2001. Change of the weed seed bank during Yenish, J.P., Doll, J.D. & Buhler, D.D. 1992. Effects of till- the first complete six-course crop rotation after conver- age on vertical distribution and viability of weed seed in sion from conventional to organic farming. Biological soil. Weed Science 40: 429–433. Agriculture and Horticulture 19: 71–90. Thomas, A.G., Légere, A., Leeson, J.Y., Stevenson, F.C.,

260 261 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. etVol. al. Weeds20(2011): in oilseed 262–268. crops in Finland

The main weed species and their control in oilseed crops in Finland

Jukka Salonen1*, Pauliina Laitinen1, Marketta Saastamoinen2 and Juha Salopelto3 1MTT Agrifood Research Finland, Plant Production Research, FI-31600 Jokioinen, Finland 2Satafood Development Association, FI-32700 Huittinen, Finland 3Hankkija Maatalous Ltd., FI-05801 Hyvinkää, Finland * e-mail: [email protected]

A survey of weeds in spring-sown oilseed crops (Brassica rapa ssp. oleifera and Brassica napus ssp. oleifera) was conducted in southern and central Finland during 2007–2009, representing the first such extensive investigation in the country. The occurrence of the most abundant weed species in oilseeds was surveyed in 429 fields. In the fields with moderate or high weed infestation, 1–6 harmful weed species were recorded by visual observation according to their biomass production. About 40 weed species were recorded, the most predominant being Chenopodium album, Galeopsis spp., Galium spurium, Sonchus arvensis and Tripleurospermum inodorum. Elymus repens was the only major grass weed. Chemical weed control of broad-leaved weeds had been practised in 53% of the fields, resulting in relatively good control. In addition, both selective graminicides and glyphosate were used to control E. repens. Mechanical weed control was not practised in any field. The crop yield level was about 300 kg ha-1 higher in the fields with low weed infestation compared with in the highly infested fields. New promising options to replace the banned herbicide trifluralin are available. Thus, the most harmful weeds, such as C. album, which interferes with the production of high-quality oil for human consumption, can still be effectively controlled.

Key-words: Brassica rapa ssp. oleifera L., Brassica napus ssp. oleifera L., herbicides, spring turnip rape, spring oilseed rape, trifluralin, weed flora, weed control, yield

262 263 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weeds in oilseed crops in Finland Vol. 20(2011): 262–268.

Introduction um sativum L.) (Salonen et al. 2005). Reference surveys in oilseeds are available, for example, from Spring-sown cultivars of oilseed crops, predomi- Denmark (Andreasen & Stryhn 2008), Germany nantly turnip rape Brassica rapa ssp. oleifera, and (Goerke et al. 2008) and Slovak Republic (Týr & to minor extent oilseed rape, Brassica napus ssp. Vereš 2010). oleifera, are commonly grown in Finland as far north The objective of the study was to rank the most as the central areas from 60˚N to 65˚N. Harvested harmful weed species that currently interfere with seed is mainly processed for food oils and the oilseed oilseed cropping and to outline the current weed meal is used as a feed component. In addition, the control options applied in practice. The survey in- use of oilseeds for biofuel is foreseen. Cultivation formation is needed to complement national weed of oilseed crops has increased markedly during re- mapping data, for advisory purposes and for dem- cent years, and they are considered to be excellent onstrating the need for updated control strategies. break-crops for cereal-dominated cropping systems (Peltonen-Sainio et al. 2009). The harvested area of oilseed crops in Finland ranged from 63 800 ha Material and methods to 89 500 ha in 2007–2009 (FAOSTAT 2011). The interest in growing oilseed crops instead of spring Study regions, farms and fields cereals has been partly related to fluctuating grain prices over time, and the cultivated area of spring- sown oilseeds reached its record level, 158 000 ha, The weed survey was carried out in southern and in 2010 (TIKE 2010). central Finland from 60˚N to 63˚N during 2007– The long-term declining trend in the yields of 2009. The survey regions and farms were selected oilseed crops in Finland has not been particularly randomly using national statistics on oilseed crop encouraging (Peltonen-Sainio et al. 2007). In- cultivation from previous years, which were pro- creased pressure by plant pathogens is assumed to vided by the Information Centre of the Ministry be one of the major constraints to achieving yield of Agriculture and Forestry. Information about the potentials with modern high-yielding cultivars. visited regions and oilseed cropping in general is Evidently a proportion of potential oilseed yield is reported in detail elsewhere (Peltonen-Sainio et al. also lost due to competition with weeds. The selec- 2011). The information on cropping measures was tion of herbicides for broad-leaved weed control in recorded by interviewing farmers. The questionnaire oilseed crops in Finland has been very limited and included, for example, information on crop variety, relatively expensive. Moreover, the most applied crop rotation, soil properties, tillage practices, her- active ingredient, trifluralin, is not registered for bicide use and estimated yield. use beyond 2009 as a result of the EU Commis- The number of fields examined in the weed sion decision 2010/355/EU to not include trifluralin survey was 144 in 2007, 155 in 2008 and 130 in on the list of marketable plant protection products. 2009. In most cases just one field per farm was Trifluralin has been used because of its efficacy studied annually. All farms, except one, practiced against the major weeds in spring-sown crops, such conventional cropping. The previous crop in sur- as Chenopodium album, Galeopsis spp. and Stel- veyed fields was predominantly (75% of fields) a laria media. spring cereal (barley, oats or wheat). This study was the first comprehensive survey of weeds in oilseed crops carried out in Finnish farmers’ fields. The survey provides new informa- Weed samples and yield estimates tion on the composition of weed floras in arable fields in Finland. Earlier sources of such information are the surveys in spring cereal fields (Salonen The occurrence of weeds was visually assessed et al. 2001, Salonen et al. 2011) and field peas (Pis- in August by walking through the entire field (or 262 263 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weeds in oilseed crops in Finland a representative part of a very large field) along a Table 1. Incidence of 38 weed species recorded in the W-shaped route with 15–20 stops and recording fields rated as either “Moderate” or “High” weed infestation in a survey conducted in 2007-2009 in 429 spring- the 1–6 most abundant weed species that appeared sown oilseed fields. either evenly distributed or in patches. The main No. of purpose of field tours was to carry out a crop disease No Scientific name infested survey and therefore the assessment method was fields1 selected according to sampling techniques relevant 1 Galium spurium L. 112 to estimation of disease incidence (Delp et al. 1986). 2 Sonchus arvensis L. 99 The level of weed infestation in each field, as- 3 Chenopodium album L. 81 sociated with visual estimation of weed biomass 4 Galeopsis L. spp. 78 production, was categorized into three classes; 1) 5 Tripleurospermum inodorum (L.) Sch. Bip. 73 negligible_low, 2) moderate (including patchy oc- 6 Cirsium arvense (L.) Scop. 58 currence) or 3) high. In the case of the field being 7 Stellaria media (L.) Vill. 29 put into the first class, no individual weed species 8 Elymus repens (L.) Gould 25 were recorded. Thus, the results on particular weed 9 Fallopia convolvulus (L.) À. Löve 13 species presented in this report derive either from 10 Persicaria lapathifolia (L.) Gray 10 the fields with moderate or high weed infestation. 11 Taraxacum officinale Weber in Wigg. 7 To ensure congruence in visual assessment of weed 12 Thlaspi arvense L. 6 infestation, three survey teams exercised the sam- 13 Artemisia vulgaris L. 5 pling procedure and scaling before the annual field 14 Matricaria matricarioides (Less.) Porter 5 tour in late July, and the team members remained 15 Polygonum aviculare L. 5 the same during the survey years. 16 Stachys palustris L. 4 Farmers estimated the harvested yields ac- 17 Stellaria graminea L. 3 cording to official procedures used e.g. by TIKE 18 Viola arvensis Murray 3 Information Centre of the Ministry of Agriculture 19 Centaurea cyanus L. 2 and Forestry. Data on estimated crop yields were 20 Equisetum arvense L. 2 analysed using analysis of covariance, taking into 21 Fumaria officinalis L. 2 account the following factors: year, crop, soil type, 22 Lamium purpureum L. 2 special soil properties, tillage, year since last culti- 23 Lapsana communis L. 2 vation of Brassica crops, incidence of clubroot and 24 Arabidopsis thaliana (L.) Heynh. 1 weed abundance (3 classes). The model was fitted 25 Capsella bursa-pastoris (L.) Medik. 1 using SAS/MIXED software. 26 Cardamine L. spp. 1 27 Euphorbia helioscopia L. 1 28 Myosotis arvensis (L.) Hill. 1 29 Plantago major L. 1 Results 30 Polygonum L. spp. 1 31 Raphanus raphanistrum L. 1 Some 40 species were recorded in the list of abundant 32 Rumex longifolius DC. 1 species, many of them, however, occasionally in 33 Solanum tuberosum L. 1 1–5 fields only (Table 1).Galium spurium, Sonchus 34 Spergula arvensis L. 1 arvensis, Chenopodium album, Galeopsis spp. and 35 Trifolium hybridum L. 1 Tripleurospermum inodorum were most often rated 36 Trifolium (L.) spp. 1 as abundant species in survey fields (Fig. 1). Elymus 37 Tussilago farfara L. 1 repens was the only grass species recorded. 38 Urtica dioica L. 1 The weed infestation was highest in 2007 when 1) Out of the 259 fields rated as either moderate or high the oilseed fields suffered from a severe flea beetle weed infestation 264 265 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weeds in oilseed crops in Finland Vol. 20(2011): 262–268.

Number of fields infested 120

Fig. 1. Incidence of weed spe- 100 cies in oilseed crops in 2007- 2009. The number of fields 80 62 where the weed species were recorded in the 259 fields rated as 62 either “Moderate” or “High” in- 60 42 festation. Key to EPPO codes of 46 weed species: GALSP=Galium 55 40 spurium, SONAR=Sonchus ar- 42 vensis, CHEAL=Chenopodium 50 album, GAESS=Galeopsis spp., 20 39 15 MATIN=Tripleurospermum 37 32 16 inodorum, CIRAR=Cirsium 18 16 14 7 6 9 6 arvense, STEME=Stellaria 0 4 media, AGRRE=Elymus repens, POLCO=Fallopia con- MATIN CIRAR POLLA GALSP CHEAL GAESS STEME POLCO SONAR AGRRE volvulus, POLLA=Persicaria Moderate High lapathifolia.

(Phyllotreta spp.) attack. Only 20% of surveyed either in the previous autumn or at sowing time, fields were ranked as low-infestation in 2007 while mainly in conjunction with direct drilling. Con- the proportion of these ‘weed-free’ fields was about trol had apparently been successful since almost 55% and 45% in 2008 and 2009, respectively. In total, there were 170 fields with negligible or low weed pressure. Their proportion was significantly (χ2 = 31.77, df = 2, p < 0.001) higher in the fields Proportion of fields, % treated with herbicides compared with in the un- 100

treated fields (Fig 2). 90 Chemical weed control against broad-leaved weeds was practised on 53% of fields. None of the 80 survey farms practised mechanical weed control. 70

Trifluralin was the most common active ingredi- 60 High ent applied to total of 190 fields with an annual range of 41–53% of treated fields. Napropamide, 50 Moderate metazachlor, clopyralid/picloram and imazamox 40 Low were the other active ingredients used for this pur- 30 pose. In addition to the control of broad-leaved 20 weeds, grass weeds, Elymus repens in particular, 10 were controlled separately with selective gramini- 0 cides, including propaquizafop, quizalofop-P- No Yes ethyl, fluazifop-P-butyl and cycloxydim. However, Weed control selective grass weed control was a less common Fig. 2. Proportion of survey fields in three weed infesta- practice and was carried out only on 123 survey tion classes as a result of either refraining from or rely- fields. Glyphosate had been applied to 47 fields ing on chemical control of broad-leaved weeds. 264 265 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weeds in oilseed crops in Finland all fields with moderate or high infestation of E. crop rotations. The most abundant species, like repens had not been treated with graminicides. Chenopodium album and Sonchus arvensis, have Based on the yield results estimated by farm- been among the most harmful species in spring ce- ers, the average yield in all three years was higher reals and in field peas (Salonen et al. 2001, Salonen in the fields with low weed infestation compared et al. 2005). In contrast, some rare weed species in with in the highly infested fields. Estimated yields arable fields, likeArabidopsis thaliana, Centaurea were analyzed statistically and significant differ- cyanys and Euphorbia helioscopia, were ranked ences (p<0.001) between weed infestation levels abundant in 1–2 fields, probably as a result of slow were detected; the average yield in the fields with and uneven establishment of crop stand and/or re- low level infestation was 1853 kg ha-1 (N = 113, fraining from chemical weed control. s.e. 59), in moderately-infested fields 1703 kg ha-1 There were no clear regional differences in the (N = 109, s.e. 57) and in highly-infested fields 1551 occurrence of weed species. The success of some kg ha-1 (N = 60, s.e. 63), respectively. The aver- species, like Galium spurium, is well in line with age crop yields were highest in the fields where the latest observations from spring cereal fields both the broad-leaved and grass weeds had been (Salonen et al. 2011). G. spurium is a character- controlled. istic species of Finnish arable fields, whereas G. aparine prevails in oilseed crops in other countries (Andreasen & Stryhn 2008, Goerke et al. 2008). In general, comparison of results from Finland with those of other countries should be done with Discussion caution since spring-sown oilseed cultivars pre- dominate in Finland whereas winter oilseeds, with The weed survey was part of a research project aimed an over-wintering weed flora, are predominant in at improving oilseed crop cultivation in Finland. other countries. The main purpose for farm visits in 2007–2009 was The efficacy of chemical control was relatively to collect crop samples for plant disease and root good in most of the survey fields. The most com- growth monitoring (Peltonen-Sainio et al. 2011). monly used herbicide, trifluralin, did not seem to Therefore, the weed occurrence was observed in control the late-emerging and climbing G. spurium a less comprehensive manner compared with the effectively enough in all survey fields. As always, earlier weed surveys carried out by MTT. However, the final outcome of weed control is a combination the results from oilseed crop fields are applicable of herbicide efficacy and the competitive ability for comparison of the weed species producing the of the crop, which was not optimal in all fields highest biomass, for example in spring cereals. As and particularly not in 2007 when insect pests af- an example, tall perennial species such as Cirsium fected the crop. Farmers obviously recognize the arvense, Elymus repens and Sonchus arvensis were need for controlling E. repens because both selec- naturally easy to observe in crop stands in August tive graminicides and glyphosate were commonly and therefore were often rated as abundant species used. In reference to the results of our weed survey, when present in survey fields. In most cases, how- henceforth advisory efforts should place greater ever, they occurred only in patches, as observed also emphasis on the need for new strategies and op- in other surveys (e.g. Salonen et al. 2005, Goerke tions to control certain harmful broad-leaved spe- et al. 2007). All weed species were still relatively cies. In this respect, launching of new selective fresh and easy to observe at the time of the field herbicide products on the Finnish market would tour in August. be desirable. The observed weed species, particularly the Chenopodium album markedly reduces both first ten species in the rank, are characteristic of the quantity and quality of oilseed yield and should arable fields in Finland where annual spring-sown consequently be regarded as one of the main targets crops, particularly spring cereals, dominate the of chemical weed control. In this respect, introduc- 266 267 AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weeds in oilseed crops in Finland Vol. 20(2011): 262–268.

tion of the clopyralid/picloram product on the Finn- are highly competitive under favourable conditions ish market in 2010 and promising experiences with and new options for chemical weed control have the CLEARFIELD® production system (Ruuttunen been introduced. et al. 2010) represent feasible options to replace trifluralin as both effectively controlC. album, which Acknowledgements. The weed survey was part of was one of the predominant weed species. In addi- the research project “Improving competitive ability tion, both new options improve the control of the of Finnish oil crop production: enhancing yields perennial broad-leaved species Cirsium arvense and increasing cultivation area (RYPSINOSTE)“, and Sonchus arvensis, which were not effectively coordinated by Professor Pirjo Peltonen-Sainio. controlled with trifluralin. It is likely that farmers She, Terho Hyvönen and Pentti Ruuttunen provided either place reliance on chemical weed control or valuable comments on the manuscript. Lauri Jauhi- leave the oilseed fields untreated instead of choos- ainen assisted with the statistics. We thank farmers ing mechanical weed control, which has become who allowed us to visit their fields and provided us apparent also in the weed surveys in spring cereal with information about farming practices. Sampling fields (Salonen et al. 2001, Salonen et al. 2011). In procedure was a joint effort by MTT, Satafood De- general, more sustainable use of herbicides is pos- velopment Association and Suomen Rehu Ltd. The sible with the new compounds because they can survey was financed by MTT Agrifood Research be applied according to observed need after the Finland, the Ministry of Agriculture and Forestry, emergence of weed seedlings, which was not the Mildola Ltd. and Suomen Rehu Ltd. case with trifluralin, which had to be incorporated into the soil at sowing time. Plant breeding programmes in Finland have been successful in terms of increasing yield potential and improving chemical composition of References oilseed yield. Unfortunately, it seems that the Andreasen, C. & Stryhn, H. 2008. Increasing weed flora newly released cultivars are targeted for cultivation in Danish arable fields and its importance for biodiver- in highly productive environments but are sensi- sity. Weed Research 48: 1–9. tive to the stressful growing conditions that often Beckie, H.J., Johnson, E.N., Blackshaw, R.E. & Gan, Y. 2008. Weed suppression by canola and mustard culti- occur in Finland. Therefore, the current realistic vars. Weed Technology 22: 182–185. national potential for spring turnip rapeseed yield is Delp, B.R., Stowell, L.J. & Marois J.J. 1986. Evaluation of estimated to be around 2 100 kg ha-1, which has ac- field sampling techniques for estimation of disease incidence. Techniques 76: 1299–1305. tually not been realized during recent years (Pelto- FAOSTAT 2011. Agricultural data. Available online at http:// nen-Sainio et al. 2007). In principle, oilseed crops faostat.fao.org/site/567/default.aspx. Accessed 31 Jan- suppress weed growth efficiently, particularly if uary 2011. Goerke, K., Richter, U., Schulte, M. & Gerowitt, B. 2008. Re- the conditions for establishment and early growth gionale Unterschiede in der Rapsunkrautflora Deutsch- are favourable (Beckie et al. 2008). In our survey, lands. Gesunde Pflanzen 60: 151–158. one third of the fields rated ‘low weed infestation’ Goerke, K., Schönhammer, A., Schulte, M. & Gerowitt B. 2007. Weeds in oilseed rape in Germany. Status and as- had not been treated with herbicides. However, the sessment of changes. Proceedings 14th EWRS Sympo- yield levels in weed survey fields, 1 550 – 1 850 kg sium, Hamar, Norway. p. 198. ha-1, were clearly lower than the above-mentioned Peltonen-Sainio, P., Jauhiainen, L. & Hannukkala, A. 2007. Declining rapeseed yields in Finland: how, why and what potential average yield. next? Journal of Agricultural Science 145: 587–598. In conclusion, the main weeds of oilseed crop Peltonen-Sainio, P., Jauhiainen, L., Laitinen, P., Salopel- fields were the same as those recorded in earlier to, J., Saastamoinen, M. & Hannukkala, A. 2011. Iden- tifying difficulties in rapeseed root penetration in farm- surveys of spring cereal and pea fields in Finland. ers’ fields in northern European conditions.Soil Use and Weeds should be considered as a one of the main Management 27: 229–237. constraints to oilseed production, both in terms of Peltonen-Sainio, P., Jauhiainen, L. & Venäläinen, A. 2009. yield quantity and quality. Fortunately, the crops Comparing regional risks in producing turnip rape and 266 267 AGRICULTURAL AND FOOD SCIENCE AGRICULTURAL AND FOOD SCIENCE

Salonen, J. et al. Weeds in oilseed crops in Finland

oilseed rape – Today in light of long-term datasets. Acta Salonen, J., Hyvönen, T. & Jalli, H. 2005. Weed flora and Agriculturae Scandinavica Section B – Soil and Plant weed management of field peas in Finland.Agricultural Science 59: 118-128. and Food Science 14: 189–201. Ruuttunen, P., Lassi, K. & Pfenning, M. 2010. Effective Salonen, J., Hyvönen, T. & Jalli, H. 2011. Composition of weed control in spring turnip rape (Brassica rapa sub- weed flora in spring cereals in Finland – a fourth survey. sp. oleifera) using the CLEARFIELD ® production system. Agricultural and Food Science 20:p. 245–261. Proceedings of the 15th EWRS Symposium, Kaposvár TIKE 2010. Matilda Agricultural Statistics service. Availa- Hungary. p. 297–298. ble online at: http://www.maataloustilastot.fi/en/utilised- Salonen, J., Hyvönen, T. & Jalli, H. 2001. Weeds in spring agricultural-area. Accessed 10 Febuary 2011. cereal fields in Finland – a third survey.Agricultural and Týr, S. & Vereš, T. 2010. Temporal dynamics of weed in- Food Science in Finland 10: 347–364. festation in the winter oilseed rape canopies. Herbolo- gia 11, No. 2: 1–8.

268 PB 2011

cience www.mtt.fi/afs 3,

No.

food

20,

and

Vol. Agricultural Economics Agricultural Engineering Animal Science Environmental Science Food Science Horticulture Plant and Soil Science

gricultural A Agricultural Society The Scientific of Finland Agrifood Research Finland MTT

AGRICULTURAL AND FOOD SCIENCE V o l . 20, 3 (2011) 191–268 235 245 262 191 206 127 228

cience

AND FOOD S FOOD AND ISSN electronic edition 1795-1895