Qliest#s - V. 9, 022 Guny 1985) pp. 95-104 VARIABILITY OF THE RADIUS GROWTH AT 25 YEAR OLD PINUS UNCINATA MILLER EX MIRBEL SPECIES IN THE SPANISH NorTH-EAsT PYRENEES EMILIA GUTIERREZ UNIVERSIDAD DE BARCELONA It haa been studied the variabitity of tem radial growth in four radiuacrtentations at the Age of 25 yaare old trees. the species was Pinus wninata. Thedeampling outes, Located i te North East Spanteh Pévencee belong to different plane gomnitiee. an Ai07h ~ analysie as performed in order to find out union factore, the radio orientation, plant comunity of"the oite and tree ttechf are aignifoative of have some Vind of influsnce on the tree redial gravth. Keywords: ANOVA, NESTEL MODELS, MIXED MODELS, PSEUDO-F, SECONDARY GROW, RADIO ORIENTATION, COMMUNITY EXPOSURE. LINTRODUCT LON. ‘The variability in the wooden species radius growth can be due to different factors such as the slope, the exposure to solar radia~ tion of the site, the soil, the tree habitat and the tree itself ,among others. The tree response matanisms to all of these factors througn both geotropic and phototropic proce- s8es produce a mechanical stress in branches and stem (Sinmermann et. a./3/). May be tne most notorious example is tne reaction wocd formation produced in gym Rosperms as well as in angiosperms, when -- trees grow on a very strong slope. However , tne eccentricity in the radial growth is not 80 Spectacular most of the times. It is a -- consequence of many factors and maybe of them has a major effect. ‘Tree physiology mediates the responses - by varying and making the distribution of assini~ lates and hormones asymenetrical. So we know that tension wood formation (in angiosperns) is induced by a low auxin concentration while compression wood formation is promoted by — high auxin levels (Morey, R.C. /2/). The stem radial growth irregularity diminish- es from the bottom to the dominant apical - meristem. This is logical because it is the ~ Emilia Cutiéeraz - Universi¢ar de Barcelona - Departament de Heologia = Article rebut ol juliol de 1985. lower part of the stem where the root vessels and where reaction wood is formed in a high slope. This is also the part where ilematic stem vessels reinforce the whole tree. 2. OBJECTIVE. The objective of this work is to study the variability of the radiye growth at 25 years of 0c of the tree species Pinus uncinata Miller ex Mizbel. This species, as the majority of those of the temperature zones, formes growth rings as a consequence of its secondary growth. They are easy to measure with a binocular micros~ cope with a micronetric. We had two aims in taking the radiis until 25 years of age variable: on one hand to study the radius growth variability itself caused by the factors of interest and on the other hand to exclude other factors that affect radial growth rings negatively, that is the produe- tion of fruits. This species growth is main- ly vegetative during the first 20-25 years. ‘Trees expecting a long future invest in their structure the major part of their production during the first period of their life, which has adifferent length in years, for each species. = Av. Diagonal, 637 ~ Barcelona 95,Qliest¥6 -V. 9, n= 2 (juny 1985) S, ANALYSIS DESIGN, FACTORS OF INTEREST. ‘The external factors we have mpreciated in this analysis as the most important ones in the radial growtn of Pinus wncinata have been: tne plant community, the radius growtn orientation, that is the orientation corresponding to tne direction of growth Fadiis of the stem and the last factor is tree itsels. ed. THE PLANT COMMUNITY. The natural habitat of this species is 1o- cated in the Pyrenees where it forms fo- rests belonging to different plant conmu- nities petween 1600 and 2000 meters 2.8.1. ‘The characteristics of these communities re~ ferring to the soil, its exposure and the composition of the accompanying species are specific of each one of them (Table 1). P. uncinata is tne dominant species of the com munities we have chosen. These are: -- Pulsatillo-Pinetum uncinatae, Arctostapiy1o- Pinetum uncinatae and Rhododentro-Pinetum uncinatae. We have included another Pulsa- tillo-Pinetum uncinata D because it has different cnaracteristics from the typical community. TABLED coMMUNITZES 3.2, ORIENTATION FACTOR OF THE SAMPLED RADIUS IN THE TREE. This factor refers to the direction of the growth radiw in the stem of the tree. Your radii with different orientation in the tree were taken North (N), South (S), ast (f) and West (Wi). he purpose of tnis was to pick up the possible variability producea by factors such as: the soil slope, sranch orientation and distribution along the tree, and competition among trees. 3.3, THE TREE AS A FACTOR IN THE ANALYSIS. As biological systems trees present an in- trinsic variability per se . Their sexual reproduction gives them an individual gene- tie information and thus different from one another. Therefore, althougn they can grow in similar conditions their responses have need not be the same necessarily. on the other nana, the totality or only some of tne trees of an area may suffer danages caused by insects, fire, snow, ete. that - can make the radial growth smaller across the whole section or partially. Pulsatillo Arctostaphylo Rhododendro Pulsatillo-D sor. Basic acta Acid EXPOSURE 30 (W-NE) 190 (S) 1s (w) ALTITUDE 1780 m 1900 1800 m suoPs, 26 35 30 SITE NAME -MASHLLA COLL, DEL PAL MOTXERO Basic 350 (wy 1700 35 PEDRAFORCA Characteristics and sites of the communities where samples have been taken. ‘The Pulsatillo-D conmunity has been included and analysed as one of the levels because it presents a species com position that belongs more to acid soils than to basic ones. Tne sites wnere these communities have been sampled are in table 1, all of these belong to the Cadi mountains in tne Pyrenees (Fig. 1) We have taken the samples almost at the same slope and altitude in order to enable comparison. 96Qllest#'6 -V. 9, n° 2 (juny 1985) 3.4. NUNBER OF SAMPLES, 13 is the plant community factor also represented by four levels j=1,2,3,4 that ‘The numer of samples was 48. Three trees are respectively the four sites’ conmunities. from each of the four conmunities and four stem radius growth orientations from each Ak (5) 48 the tree factor with k 1,2,3 le- tree. So, it is 3 trees x 4 communities vels, the 3 trees sampled in each community. % 4 radio orientations. In this way we Factcr Ak(}) is nested in the community as had a complete balanced design for each one of the factors (Table 2). each tree is different from the other. Obi} and OAix(3) represent the interac- tions radiusorientation-conmunity and radius 4, ACQUISITION AND PRCPARATION OF DAT. orientation-tree respectively. We carrot take into account the LAjk interaction be~ So as to measure the 25 year old tree cause they are nested. radius we took the samples from the tree stem, at a height of 1 metre from e4jk is the aleatory error in the usual the base with an increment borer. hypothesis iia N(0;02). The cores are mounted and the ring widths are counted fron the pit to the 25th tree 5.1. ANALYSIS OF THE VARIANCE. growth ring. The summatory of these is the length of the tree radiss at 25 expressed The ANOVA models are among the most frequen- in om (Table 2) tly Used in statistical studies of this kind. ——E——EE TABLE 2 UA MASSELLA COLL DE PALL = MOTKERO PEDRAFORCA, AL AZ A390 Ad AS AB OAT ABO. ALL Az 7,88 8.61 11.56 4.95 5.30 8.01 5.76 4.95 4.49 6.71 4.68 5.13 8.27 8.81 12.66 6.95 6.73 8.11 4.70 5.00 5.62 6.09 4.87 5.35 6.88 8.69 11.35 6.93 7.34 8.80 4.98 $.20 5.20 5.40 4.12 4.25 8.41 7.77 12.42 6.95 7.60 9.76 $.01 5.11 5.17 6.94 4.59 7.43, Length of the radii at 25 years in cm., for the sample sites of the 4 communities, 12 trees (Al, A2,...,A12) and for the 4 stem radius orientations. EE 5+ STATISTICAL MODEL OF THE DESIGN What ve have expressed in (1) is the design model of @ three way analysis of the variance mixed (2 factors are of random efects and 1 is fixed), factorial in 2 fac- tors [radio orientation-vegetal community and radio orientation-tree) and nested in the factors tree-plant community. Repli- According to the objective and the design followed in the gathering of samples plan, the implicit statistic model is: YAK —p+OL+L5+OLA FAK (4) HORLK(Z) Fete UL) cations are not retrievable due te the Mhere Yijk 18 the logarithm of the radio construction of the design (Table 3). at 25 years of age (radio 25). 0k represents the radio orientation factor with i-1,2,3,4 levels, that are the samples Of every tree growth radius orientation (5,8 and Ww) 97Qtests6 -V. 9,22 (juny 1985) TABLE 3 Source of variation Yadiusorientation: Oi RANDOM 4 Ni-1 = 4-1 Plant community: Lj FIXED 4 = Nj-2 = 4-1 interaction: OLij RANDOM (i) (j-1) tree: Ak(3) RANDOM 3 (Nke1) (Nj = interaction: OAik(}) RANDOM (i=1) (k= 1) TOTAL we wNeL= 48-1 Characteristics of the factors and their interactions. To analyze the data we took their logarithns because as can be semin figure 2, variance is not independent from the average: as the level of the observations increasesso does the variance. Furthermore, we point out that the radius orien tation factor is treated as although the samples were taken in a fixed radius orientation, because it is a conti- uous variable and we took a high enough nun ber of radius orientation samples to cover yandom lange range of variation (A. Prat, verbal communication) « The model hypotheses are: on Aid Nt 15 1,2,3/4 y Lugs 0 3 = 1,2,3,4 ous 48 NCOF8FOL) Ff onts40 jou. i Ak(3) 4d N(O784a) k= 1,2,3 OAGK(3) ila N(0,6%0A) 6 RESULTS. The calculations corresponding to the Sun Te de im nested of Squares (SS) are given below portant to be careful with the factor. Sum of squares for the radi orientation factor. Effect levels degrees of freedom =3 3 a3 3 Gna = 8 Nj = 34204 = 24 247 443 Loh vise ouna = S22 371 KD ~ N(F)#=0.016278 (ay 43 Sum of squares for the plant factor. community J. vagey? wed =n(@) #=0.512554 (a) Sum of squares for the tree factor. i (Eo wager? akg) = 5 Et gh nn 4 a xs5K)? - Sa et = 0.147905 «> oe Sum of squares for the O11} interaction. ouij-1 (orn) (ue1) 14g - y= Ob + a 443 Lod od waaay? SEL EL BEL | pjsoser =0.043005 (0) 2 Total sum of squares (TSS). 443 mss = | ff wisn? - wth) te0.75¢514(6) ah geal ke Sum of squares for the OAik(}) interaction that is the same as the residual squares (RSS), (see below). sum of ESS = (5) = (A) ~ (B) - (©) ~ (D) = 0.034772 98Qliestsi6 -V. 9, n2 2 (juny 1985) 6.1, EXPECTED MEAN SQUARES (EMS), ‘The expected moan squares turn out to be extremely important as an aid in deciding how to set up an F test for significance - (ilicks /1/). We do EMS using the Yates al- gorithm (Table 4). interactions. The F's for the fixed effect factors are calculated with the interaction expression that contains this factor. In our case all tne F's can be calculated without difficulty except for that of the — TABLE Expressions of the expected mean squares. souRCE 4430 or RF RR VARIATION i jk om EMS's 08 14301 otE te%0R + 120%0 uy 40 3 1 GB +o40A + deta + 3otoL + 2p b ous3 10 3 1 GE +a%0a + 3u%on ak (3) 40101 1 088 +20a + ag2a onik (3) 1 1 1 2 GE + 3% Bijk 1 1 1 2 a@® (not retrievable). As there are not replications, the degrees Of freedom are quite low for the tests in~ dicated in table 4, and since there is no Separated estimate of error variance, we do 0A 0 to serve as the error variance, and from a biological point of view, OAik has not much interest. ee community factor (Lj). Tais F camot be obtained directly since no expression con tains its EMS. In these circumstances it ss of common practice to do a new F called F' or poeudo-F, and new degrees of fredom by a Lineal combination taking the right factors and interactions. In this way, we obtain a Se TABLE 5 ANOVA table obtained from table 4 with the new expected mean squares expressions. SOURCE OF VARIATION D. F,BNS's SS's MS's on 3 7B + 120% 0.016278 0.c054260 1 3 o4Br4g*Ar30%OL+128 L 0.512554 01708515 o1ij 9 0B + 30%0L 0.043005 0.c047783 RG) 8 OE + 407A 0.147908 0.c1sager Eijk’ 24 oe 0.034772 0.corssee TOTAL 47 0.754514 ee To obtain the F's from table 5, we calculate them with the error variance for all the main factors of random effects and random Rew mean squares from the interaction ob, the factor A and the error (0A). 99Qliesti6 -V. 9, n° 2 (iuny 1985) MS MS (OL) + MS(A) ~ MS(OA) MS OE + of0L + o?E + 407A - oftz OPE + 30%OL + 402A, Where MS(OL) 0.0047783 with 9 d. of £. ms(n) 0.018081 * a ms(oR) 9.001488 "240 he F* calculus is pr = MSLLH) | 0.1708515 _ gg) Ms 0.0247152 and the new degrees of freedom, v, are: o.o247152 a (2)? (00)7/9 6 (2)74098 (4)7/03¢ -) tes (0n)?/24) = 135 Jn this way, P's value for Lj factor is 6,91 with 13,5 d, of £. (Table 6). ‘TABLE 6 F values and signification level: s+ at 18 level and tes at 5 1 te a pseudo-F. souRce OF VARIATION F's D. of Py ob 332 3 624 + By) 691 3 613 ae 3 eld ae ous 3.92 9 6 24 we ak (3) 22.21 @ 6 24 ane Hott: oFA1 = g%a2 = o7A3 All the trees grow in the same manner and this factor does not add new variability to the radio growth. Obviously we have to reject all the null hypothesis (Table 6) and to assume that, there are significative diferences rads 25 growth between stem radius orienta- tions, plant in the communities and trees. ZDISCUSSION AND INTERPRETATION OF THE Statistically, the tree factor is the most, Its interpretation, based facts, following way: trees from the same area pro~ significative one. on biological can be done in the cess a similar external information (climate * significative at 2.5% level level. The Lj factor F (.) is Starting from these results ve will test the Ho,Ho' and Ho! hypothesis: Ho: a%0n = o70s = o%00 = ow =o? Which means that the growth variance in the 4 stem radinsorientations is the same; chat is to say,the radiis orientation factor does not add variability to the radim growth. Hot : L(ceso) L(cHSQ): L(CCP)= L(CHX)= L(CPF)= 0 LICCP) =u LAO) p L(CP)= ye The width growth or secondary one of P. un~ cinata species is the same in any of the plant communities. soil, accompanying species, ...) and they give a different response in their growth A set of exogemus factors , those we know and those we do not notice, have an radio influence on the level of their physiologi- cal processes ina particular way in tree each as a result of the variability in the genetic background. So, while the physiolo~ gical processes are the ones that mediate the exterml and internal interactions, the genetic information potentiates and leads these responses. Bvory tree is adapted and prepared to give response to the great amount of changing factors that modify tree space, and habitat, 100Qlests'6-V, 9, n° 2 (juny 1985) wich 1s of a fantastic hetereogenity. There is not 1 square metre equal to another nei- ther in time (climate, other species...) nor (so41 composition, rockness in space th RADIUS ORIENTATION AND THE PLANT commu nity also present significative differences in the trees'secondary growth as well as in their interaction. Im table 1 you can observe the growth of the different communities in the sampling sites: La Masella, represented by CESQ, Coll del Pal by CCP, Moixero, CMX, and Pedraforca that we will call CPF. The community found in CHSQ is the one that presents larger dif. ferences as we compare them. The differences, among the other communities are smaller, -- specially between CMX and CPF. The reasons for CBSQ to have a bigger radius growth (25) reside in the particular conditions of this site which are nore favorable than the others. The topography allows the trees to receive solar radiation almost during the whole day because it is an opener zone. The highest mountain tops to the South, are far away and do not obstruct the direct penetra- tion of solar radiation. It also has the most developed soil of ali the sampled communi~ ties and is the one that has less slope. The slope has to be interpreted as the soil's capacity for retaining nutrients and water, 4£ it is large the losses are supposedly ‘The community from which we took samples in site CCP, Arctotaphylo-Pinetun uneinatae, For this rea~ grows in a Southern expos: son they are exposed to photosyntetically active radiation (PAR) for nore hours during the day inctding favorably on production on the contrary the communities with Northern orientation have generally less direct sun radiation and fewer hours of exposure. ‘The differences among communities are ins- tead a consequence of other intrinscal fac~ tors such as insolation, tics, topography and the The first soil characteris-- kind of accompa—— 2 of these have,in our opinion, the largest influence. The — nying species. tree configuration also supports this idea. Wnile in CH and CPF the fort occupy more space vertically, than in CCP and CESQ. the branches and the shape of the trees is more similar between the last two communities which also have a greater radial growth even though they are not so high. ‘The strategy for radiation is different in CCP and CESQ. The branches grow all along the stem starting almost at the base in most of the trees. They grow horizontally and form a large canopy slightly turned to the SW in cee. branches is in some cases almost apical, spe- In CHX and CPF the distribution of the cially so in CPF. The lower branches lose their functionality as support elements of assimilatory surface and they finally die. The wood quality related to the height and Aianeter of the tree growth, is an important item of forestry wood production studies. The quality increases when the trees are hi- gher. Among the studied communities, the —- best quality is that of the Pedraforca pine woods (CPF). ‘The radius orientation factor has been com mented implicitly when we discussed the plant community since it is closely related to this and to the strategy that the tree has to follow in order to adapt to the given habitat and sun exposure conditions. It 4s important to point out that there is RO compression wood in none of the studied sites at 1m height from the base. ‘The variability among radius orientations within communities is consequence of the incident solar radi. fon, the competition between tree branches and roots and between the surrounding trees. In CMK the growth is less but more homogeneous. This community has a shorter distance between trees and also a smaller distance variance. 101Qlleat¥6 - V.9, n° 2 (juny 1985) &._CONCLUSLONS. his advices and help. tam grateful also to Yaura Gutiérrez and Helena car who help The rads grovth at 25 years of age Pinus uncinata species presents significative dif- ferences in front of the studied factors: the plant community, the stem radius orien- tation and the tree itself. We should count on this variability in studies on second- ary growth and the stem bionass and when making the data interpretation. me to translate the text into English. The tree factor appears as the most signi- ficative one because genetic infornation is particular for each one of them and because of the site heterogeneity. For tits reason it would be necessary to sample a larger nunber of trees rather than to pick up more core samples per tree. ‘The radbis orientation and plant communi- ty factors have a significative interac- tion, Therefore their interpretation has to be done at the same time because the location and light exposure of the latter condition @ strategy and disposition of the tree's assimilatory surface and conse~ quenthy the radial growth The differences between communities are due not only to their exposure but also to the trophic quality of their soilsmepi for cMx is 4,5 and 6 for CPF while in CCP and cEsQ it is basic . The first Two soils are more oligotrophic. 9. REFERENCES, /A/ HICKS, C.R.: "Pundamental concepts in the design of experiments", Holt, Reine- hart and Winston, New York. (1973) /2/ MOREY, R.C.: "Como crecen 10s arboies". Omega - Barcelona (1977). 73/ CIOMERMANS, MLA; BROWN, C.b.: “Trees- structure and function". New York Heidelberg-serlin. Springer. (1971). 10._ACKNOWLEDGMENTS T wish to thank specially Doctor Albert Prat from the "Escuela Técnica Superior de Ingenieros Industriales de Barcelona” for 102Qliest#i6 - V9, n° 2 (juny 1985) 1G oo CADI-MOTXERO Fig. 1.- vocation of the samples sites in the North-zQltests6- v. 9, n2 2 (juny 1985) MEAN RADIAL GROWTH OF 25 YEAR OLD TREES (nm 1 tor Tt Ht Lyf i]? N|S|E]w N{S/E|W N}S[E|W CEsa@ cMx CPF Fig. 2.-eon radiol growth of 25 yer old trees. N,S,E ond W are the four orientations of the somple in the tree stem, ond CESQ, CCP, OK and CPF ‘ore the four sonpled sites. tog