Anatomical wood variation of Buddleja cordata (Buddlejaceae)
Trees(2006)20:253–261
DOI10.1007/s00468-005-0007-5
ORIGINAL ARTICLE
Silvia Aguilar-Rodr′?guez·Teresa Terrazas·
Lauro L′o pez-Mata
Anatomical wood variation of Buddleja cordata(Buddlejaceae) along its natural range in Mexico
Received:30June2003/Accepted:6May2005/Published online:25December2005
C Springer-Verlag2006
Abstract Buddleja cordata is an evergreen species of wide distribution in Mexico that is represented by shrubs and trees.Wood variability of B.cordata was evaluated in rela-tion to plant size as well as latitude,altitude,soils,and cli-matic data.Canonical correlation analysis(CCA)showed that two canonical correlations are signi?cant(Wilks’λ, p<0.0001)and explained76%of total variance.Redun-dancy analysis revealed that the?rst pair of canonical vari-ates are signi?cant,thus the canonical variate,named dis-tribution,represents a gradient of maximum temperature of the warmest period,annual temperature range,and latitude in its area of distribution;whereas the canonical variate named wood represents vessel density,?ber length,and plant size,best associated to the environmental gradient. Vessel density expressed by its distribution in latewood and porosity type showed that ring-porosity is common in individuals from high latitudes.Temperatures below zero or lack of rainfall during several months might induce poros-ity variability in B.cordata as suggested by CCA,but was not related to phenology since the species is evergreen along its latitudinal range.Plant size was also in?uenced by extreme temperature and rainfall.Shorter plants are dis-tributed in the northern population or driest sites located in north-central Mexico,and in addition,?ber length fol-lowed an allometric relation with individuals’height.Wood characters in B.cordata as for simple perforation plate,he-lical thickenings,type of intervascular and vessel-ray pits, scanty paratracheal parenchyma,and heterogeneous type IIB rays were not correlated with plant size,climate,and S.Aguilar-Rodr′?guez
Unidad de Morfolog′?a y Funci′o n,FES-Iztacala,
Universidad Nacional Aut′o noma de M′e xico,
Estado de M′e xico54090,M′e xico
T.Terrazas( )·L.L′o pez-Mata
Programa en Bot′a nica,Colegio de Postgraduados,
Montecillo,
Estado de M′e xico56230,M′e xico
e-mail:winchi@.colpos.mx
Tel.:+52-580-45947
Fax:+52-595-9520247soil parameters or species distribution.These features are common with other species of Buddleja.
Keywords Allometric variation.Buddleja cordata. Canonical correlation.Diffuse porosity.Mexico.
Ring porosity
Introduction
Buddleja cordata Humb.Bonpl.&Kunth is an evergreen, dioecious species that varies from shrubs to trees rang-ing from1to20m height and10to45cm in diameter at trunk base(Norman2000;Rzedowski and Rzedowski 2001);however some individual trees are able to reach more than80cm in diameter at breast height(Aguilar-Rodr′?guez,unpubl.data).B.cordata is widely distributed in Mexico from the northern states of Sinaloa,Chihuahua, and Durango to the southern states in Oaxaca and Chiapas (Fig.1).This species grows in a wide range of habitats;
e.g.,pine-oak,?r,and cloud forests,as well as in valleys of high elevation,in desert scrub with cacti and mesquite,and in chaparral with juniper(Norman1966,2000).Moreover, B.cordata is favored by disturbance,frequently occupy-ing the openings of these plant communities or associated with crop?elds.Due to the wide distribution and ecolog-ical tolerances that B.cordata possesses,the potential for variation among morphological,physiological and genetic traits is extremely high.As a consequence,B.cordata is an outstanding system to evaluate the hypotheses that have been postulated for some genera and families in relation to wood variability associated to habitat and habit features (Baas1986;Noshiro and Baas2000).
The in?uence of latitude,temperature,and humidity on secondary xylem structure has been studied in different taxa (Chalk1983),mostly at family or genus levels(Carlquist 1966;Dickison et al.1978;van der Oever et al.1981;Baas 1986;Carlquist and Hoekman1985;Dickison and Phend 1985;Rury1985;Noshiro and Baas1998).However, research at the species level is insuf?cient due to different tendencies in relation to tracheary element features
254
Fig.1.Distribution of B.cordata in Mexico and provenances sam-pled for this study
(Sastrapradja and Lamoureux 1969;Parameswaran and Conrad 1982;Li-Hsia and Tseng-Chieng 1986;Xinying et al.1988;Wilkins and Papassotiriou 1989;Lindorf 1997;Noshiro and Baas 2000;Arias and Terrazas 2001).Soil in-teraction with wood variability is poorly documented in the literature.Yaltirik (1970)found that a high concentration of calcium ions is associated with shorter vessel elements in Acer trees growing in areas with alkaline soils.D¨u nisch and Bauch (1994)evaluated the in?uence of mineral elements in Picea abies wood development,detecting that fertilized trees during lower rainfall showed an increment
of periclinal divisions in the vascular cambium.Villagra and Roig-Ju?n ent (1997)compared wood features of two species growing under different edaphic conditions and found an increment in the number of narrower vessels in Prosopis argentinean related to a water de?cit in sandy soils.Other studies had focused on the in?uence of ?ooded soils on wood structure (Sidiyasa and Baas 1998;Y′a ?n
ez-Espinosa et al.2001).The ability of B.cordata to inhabit a wide distribution range raises the following question:are there anatomical wood differences related to latitude,elevation,climate (temperature and rainfall),and soil features or to plant size?
Materials and methods
Forty-two wood samples of B.cordata from 17natural provenances were collected from Chiapas (16?47 N)to Chihuahua (27?54 N)Mexico (Fig.1)in latitude and from 1350to 3040m in elevation (Table 1).The climate where B.cordata inhabits is de?ned as temperate with summer rainfall and dry season during part of the winter and spring,but slight rains may occur during the winter.Mean annual temperature varies from 12to 19?C and annual rainfall from 544to 1286mm (Garc′?a 1973;Quintas 2000).This species also grows in a dry climate where thorn–shrub commu-nities develop (Norman 2000)with mean annual rainfall lower than 500mm (Garc′?a 1973).The predominant veg-etation at the collection sites corresponded to oak forest,pine-oak and xeric thorn (Rzedowski 1986).At most sites litosol soils prevailed with exposed outcrop rocks.
Table 1Buddeja cordata plant features,climate and soil pa-rameters.Provenances arranged in descending latitudinal order.Acronyms:height (H ,m),latitude (LAT,north),longitude (LONG,west),elevation (E ,m),mean annual temperature (MAT),isothermal-ity (ISO),temperature seasonality (TS),maximum temperature of the warmest period (MTWP),minimum temperature of the coldest pe-riod (MTCP),annual temperature range (ATR),annual rainfall (AR),rainfall seasonality (RS),rainfall of the driest quarter (RDQ),elec-trical conductivity (EC,Mmhos/cm),nitrogen (N,%),phosphorus (P,ppm)
Provenance Plant and location Climate
Soil H LAT
LONG E MAT ISO TS MTWP MTCP ATR AR RS RDQ EC N P Chihuahua 14,327?54 06
107?56 52 2014130.551.7429.9?2.73383694490.160.7024Chihuahua 23,427?54 05 107?56 51 2014130.551.7429.9?2.73383694490.160.0616Durango 14,325?18 73 104?38 30 1350190.591.5935.5 2.23339110400.160.2613Durango 2
1.5,1.5,1.5
24?04 15 104?39 28 2084160.601.3130.20.92955310800.160.127Durango-Zacatecas 2.5,4,523?44 72 104?00 53 1938170.621.230.7 2.72849211100.160.2616Zacatecas 3,5,323?11 30 102?53 05 2336160.611.0328.1 2.6254229400.160.1911Aguascalientes 3,322?24 43 102?16 91 1983170.630.9930.3 3.1274439100.160.2116San Luis Potos′?4,322?15 02 101?06 42 2073170.630.9330.0 4.22636877270.140.2518Quer′e taro 1 2.5,620?22 100?16 2330150.640.7126.8 4.12374498320.120.2737Quer′e taro 23,220?22 100?16 2330150.640.7126.8 4.12374498320.110.2941Tlaxcala 7,7,519?35 98?37 2630130.680.6324.1 1.023.166283260.150.46108M′e xico 4,419?30 98?47 2460150.680.7326.30.9256478500.100.237.8Puebla 15,15,719?21
98?39
3040120.680.4720.6 2.61895088310.130.1643Puebla-Oaxaca 8,818?31 22 97?27 15 1760180.670.6428.6 6.7225449200.170.024Oaxaca 10,8,817?33 25 97?23 1 2333150.690.5026.3 4.8217019100.190.2616Chiapas 18,10,916?50 29 92?41 382219140.710.4823.0 5.018128673850.281.1545Chiapas 210,8,9
16?47 30
92?35 49
2700
12
0.74
0.39
20.8
3.2
18
1278
72
88
0.07
0.21
2
255
In each provenance samples of two to three healthy trees were collected(Table1).Samples,including wood and bark,were cut at breast height(1.3m)for trees and at the base of the thickest branch for shrubs.Only mature and reproductive individuals were collected to assure that ju-venile wood was excluded.Samples were?xed and stored with glycerin–ethanol–water(1:2:3)solution until section-ing.Herbarium vouchers were deposited in CHAPA and IZTA.Height and diameter for each individual tree or shrub collected were recorded.Soil samples15–20cm deep were collected in dark-plastic bags and transported to the labora-tory for further analysis.Each site was georeferenced with a GPS Magellan model Map410.
Sections20μm thick were obtained with a sliding mi-crotome,stained with safranin,and mounted in synthetic resin(Johansen1940).Macerations were prepared using Jeffrey’s solution(Berlyn and Miksche1976).Temporary slides were prepared to gather data on vessel element and ?ber lengths,as well as to corroborate the occurrence of vascular tracheids.Species wood description followed IAWA recommendations(IAWA Committee1989).Due to vessel diameter dimorphism two categories were estab-lished and thus wide and narrow vessels were measured separately.In ring and semiring-porous woods,wide ves-sels corresponded to earlywood and narrow vessels to late-wood,but in diffuse-porous,they were intermixed in early and latewood.Vessel density(number of vessels/mm2),was counted only in latewood.Intervascular pit diameter,?ber total diameter,lumen diameter,and a single wall thick-ness,as well as ray height and width were also measured in cross and tangential sections.For each wood character 25measurements or counts were gathered per individual. All measurements were obtained using an image analyzer IMAGE-Pro Plus version3.1(Media Cybernetics1997) adapted to an Olympus BX-50microscope with a Hitachi KP-D51color digital camera and a screen resolution of 640×480pixels.The semiautomatic option and the proper magni?cation lens(e.g.,10x objective for vessel and?ber length,20x or40x for vessel lumen diameter,and100x for intervascular pit diameter and a single wall thickness)were used to obtain the precision needed.Fiber length/vessel element length ratio was calculated(Chattaway1936)to evaluate the amount of intrusive growth.In the species wood description,values represent the mean and the stan-dard deviation of all provenances.
Based on georeferenced locations for each population and the nearest weather stations,18climate variables were generated by Bioclim version2.0(Busby1986).Soil sam-ples were analyzed for?ve chemical parameters,texture, and electrical conductivity for each site following the stan-dard analytical methods of the Soil Fertility Laboratory at Colegio de Postgraduados(Etchevers1988).Selected cli-mate and soil variables are provided in Table1. Multivariate data analyses were performed with SAS (1989).Data distribution did not meet the assumptions of multivariate normality,thus measurements were trans-formed to logarithm,square root,or arc cosine accord-ingly(Zar1999).Porosity type is a categorical variable and was not included in the multivariate analysis,thus Spearman correlation was applied with latitude and cli-matic variables.Principal component analysis(PCA)was applied as a exploratory technique to detect a data sub-set that characterizes the variability among wood variables and plant size as well as climatic,soil,and location pa-rameters.A canonical correlation analysis(CCA)was ap-plied to determine the association between wood variables and plant size with environmental variables and to best explain the results of PCA analysis.The CCA is a tech-nique for analyzing two or more sets of variables,which organizes sampling entities(wood characters)along pairs of continuous ecological gradients.This analysis allows identifying those variables that have the highest contribu-tion within each canonical variate(Tabachnick and Fidell 1989).
Results
Anatomical description
Growth rings are conspicuous or inconspicuous.In ring-porous wood,earlywood has wider vessels,mostly soli-tary or in groups of two to six vessels;in some prove-nances narrower vessels are present at the beginning of earlywood(Fig.2A);latewood shows vessels grouped in radial rows of two to ten vessels(Figs.2A–C)or in clusters in oblique or tangential distribution(Fig.2D).In diffuse-porous wood,there are few solitary vessels,mostly dis-tributed in radial rows of two to seven vessels(Fig.2E–G) or in clusters of4–6vessels mainly with tangential distribu-tion(Fig.2H).Pores are oval and rarely angular.Tangential diameter of wider vessels is67±18μm and of narrower vessels36±9μm.Most provenances have a bimodal dis-tribution in vessel diameter(Fig.2).Vessel element length has a mean of353±77μm.Perforations plates are sim-ple(Fig.3A).Intervascular pitting is alternate;pits are slightly crowded,oval in tangential outline,with diameter of8.0±0.9μm,and oval to lenticular aperture(Fig.3B). Vessel-ray pits have reduced borders or simple,diameter is similar to intervascular pits(Fig.3C).Helical thickenings are?ne,close-spaced,occurring in wide and narrow ves-sel elements(Fig.3D).Vascular tracheids have a mean of 330±70μm.Ground tissue is composed of nonseptate lib-riform?bers,but some septate?bers are scattered in ground tissue.Fiber length has a mean of865±164μm;with?ber diameter of17±0.5μm,lumen diameter of10.7±2μm, and wall thickness of3.0±0.5μm.F/V ratio is2.2–3.0. Parenchyma is scanty paratracheal,with two to four cells per strand without deposits.Rays are heterogeneous with 4–8/mm.Uniseriate rays are few with two to six cells high. Multiseriate rays have two to three cells wide(Fig.3D), mostly two(37±6μm wide);total height has a mean of 423±136μm;multiseriate rays have a central portion of entirely procumbent cells and uniseriate marginal exten-sions of upright and square cells,commonly one to two cells.Cellular ray inclusions are absent.
256
Fig.2.Porosity of B.cordata .A–D.Ring-porous.A–C.Ves-sels in latewood mostly in radial rows.A .Chihuahua 1(Aguilar 258).B .Durango 2(Aguilar 261).C .Zacatecas (Aguilar 263).D .Vessels in latewood in oblique or tangential dis-tribution,Oaxaca (Aguilar 286).E–H.Diffuse-porous.E–G.Most vessels in radial rows and few solitary.E .Quer′e taro 1(Aguilar 267).F .Quer′e taro 2(Aguilar 268).G .Chiapas 1(Aguilar 282).H .Vessels with tangential distribution,Tlaxcala (Aguilar 281).Scale bar =200μ
m
Fig.3 B.cordata wood.A .Simple perforation plate
(Aguilar 263).B .Intervascular pits (Aguilar 258).C .Vessel-ray pits (Aguilar 267).D .Biseriate rays (Aguilar 281).Scale bar A =30μm,B-D =50μm
Statistical analyses
Table 2shows mean (±SD)values for the anatomical quantitative characters and F /V ratio ordered along the lati-tudinal gradient from north to south for all provenances.Porosity has a negative correlation with latitude (r s =
?0.75,p <0.0004),temperature seasonality (r s =?0.75,p <0.0003),maximum temperature of the warmest pe-riod (r s =?0.67,p <0.002),and annual temperature range (r s =?0.73,p <0.0005),and is positively associated with isothermality (r s =0.74,p <0.0005).No correlation was detected with plant size,mean annual temperature,min-
257
Table2Means and1SD for wood characters by provenance.Vessel density(vede),diameter of wide vessels(diwv),diameter of narrow vessel(dinv),vessel element length(vele),intervascular pit diame-ter(ipid),?ber lumen(?lu),?ber diameter(?di),?ber wall thickness (?wt),?ber length(?le),ray height(rahe),ray width(rawi),libriform ?ber length/vessel element length(F/V)
Provenance Vede diwv
μm dinv
μm
vele
μm
ipid
μm
?lu
μm
?di
μm
?wt
μm
?le
μm
rahe
μm
rawi
μm
F/V
Chihuahua116±476±1442±9378±699.1±0.1510±215±2 2.8±0.05851±103311±9839±6 2.25 Chihuahua213±481±1440±9365±889.0±0.0911±317±3 2.9±0.06988±86317±8037±6 2.71 Durango116±569±1539±10297±608.0±0.0911±317±3 2.9±0.06755±79409±9937±4 2.54 Durango219±468±1236±8249±527.3±0.089±215±3 2.7±0.08603±110359±7934±5 2.42 Durango-Zacatecas24±856±1832±8303±667.3±0.079±215±2 3.0±0.05793±114361±7332±7 2.62 Zacatecas17±457±1731±6294±838.2±0.089±215±2 2.9±0.05759±122386±8729±5 2.58 Aguascalientes21±551±1330±6280±688.3±0.079±215±2 2.9±0.06842±113411±8430±6 3.01 San Luis Potos′?22±365±1338±6290±489.5±0.1112±217±2 2.5±0.05803±91351±12433±5 2.77 Quer′e taro115±452±1128±5306±678.6±0.1010±216±2 3.1±0.05841±109387±7343±8 2.75 Quer′e taro212±472±1636±9276±658.2±0.0811±217±2 2.9±0.05813±110398±7141±8 2.94 Tlaxcala13±459±1433±6340±668.6±0.0612±218±2 3.0±0.05936±109394±8034±5 2.75 M′e xico12±338±625±6360±568.4±0.129±213±2 2.3±0.07960±82372±9428±4 2.67 Puebla7±280±1540±8341±629.2±0.1413±320±3 3.5±0.05860±97573±14548±10 2.52 Puebla-Oaxaca10±375±1338±5344±748.5±0.0511±218±2 3.4±0.08877±96507±18143±8 2.55 Oaxaca10±386±1337±6308±688.5±0.0511±117±1 3.2±0.05894±107493±11041±11 2.90 Chiapas17±270±1641±8377±728.8±0.0614±220±2 3.3±0.051081±150550±15644±7 2.87 Chiapas26±281±1444±8415±818.5±0.0911±118±1 3.4±0.041090±118534±12833±6 2.63
imum temperature of the coldest period,annual rain-fall,rainfall seasonality,and rainfall of the driest quarter (p>0.05).
PCA revealed that three components accounted for63.5% of the total variation of the secondary xylem of B.cordata wood.The?rst component explained41.3%of the varia-tion and13variables showed loadings greater than0.200. The highest loadings were for four variables:vessel den-sity,plant height,as well as maximum temperature of the warmest period(MTWP),annual temperature range(ATR), and isothermality(ISO)(Table3).The second component explained11.5%of the residual variation and the variables, minimum temperature of the coldest period(MTCP)and rainfall of the driest quarter(RDQ)showed the highest loadings.In the third component10.6%of the remaining variation was explained by the soil variables electrical con-ductivity,phosphorus and nitrogen(Table3).
Figure4shows the biplot of the two?rst components for the individuals of provenance.Along the?rst component, shorter individuals(<8m height)with more abundant ves-sels in the latewood(12–24vessels/mm2),shorter?bers (<800μm),narrower?ber diameter(15–17μm),and lower rays(311–317μm)were grouped in the left-hand re-gion.Also individuals with mostly ring to semiring-porous wood were distributed in those localities with MTWP vary-ing from28to35?C and a high ATR(26–33?C).Taller individuals(>8m height)with diffuse-porous wood and fewer vessels(6–10vessels/mm2),tending to show longer ?bers(860–1090μm),wider?ber diameter(18–20μm) and taller rays(>534μm)were grouped in the right-hand region of this axis.The provenances in which those indi-viduals are distributed have MTWP varying from21–23?C and an ATR of18?C.Along the second component north-ern provenances are separated by MTCP–3?C and com-monly zero rainfall during the driest quarter(RDQ),except for two localities of Chihuahua with49mm during that quarter.
The CCA showed that two canonical correlation coef-?cients are statistically signi?cant(Wilks’λ,p<0.0001, n=5).The?rst canonical correlation explained66%of the total variance and the second contributed8%of the remain-ing variance.The?rst pair of canonical variates suggests that an increase in annual temperature range,maximum temperature of the warmest period,a decrease in rainfall as well as an increment in latitude have an effect on?ber length,vessel density and plant size(Table4). Discussion
The PCA revealed that18of the32variables included in this analysis had the highest loading.Plant size together with?ve climate variables,latitude and three soil parame-ters explained the highest percentage of variation present in Buddleja cordata wood,especially vessel density as well as ?ber diameter,?ber length and ray height.Although most populations of B.cordata grow in temperate climates,vari-ability exists in rainfall and temperature along the year and the analyses suggest that plant size and some wood fea-tures(vessel density,?ber size and ray height)are affected by these changes.Individuals of B.cordata collected from the most northern provenances were shrubs which grow under a temperate subhumid climate with summer rainfall and mean annual temperature of approximately13?C,with minimum temperature of the coldest period from Decem-ber to April below0?C,and annual rainfall of800mm; these provenances possess ring-porous wood and in the latewood vessels tend to show longer and more abundant
258
Table3Eigenvectors of PCA for wood,height,and provenance variables for each plant collected
Variables PRIN1PRIN2PRIN3 Variation explained(%)41.3411.5510.61 Eigenvalue13.22 3.69 3.39 Vessel density?.241a?.0260.081 Diameter of wide vessel0.1160.176?.107 Diameter of narrow vessels0.1170.250?.055 Vessel element length0.1680.138?.123 Intervessel pit diameter0.1370.1980.198 Fiber lumen0.2020.1120.157 Fiber wall thickness0.167?.111?.026 Fiber diameter0.2260.0490.122 Fiber length0.2080.088?.019 Ray height0.206?.172?.005 Ray width0.1270.0430.174 Plant height0.233a?.0210.024 Latitude?.2290.2190.019 Elevation0.175?.001?.122 PH?.165?.1750.324 Electrical conductivity0.0020.0350.428a Organic matter0.1020.1560.361 Nitrogen0.1050.1680.371a Phosphorus0.0410.0590.390a Potassium0.106?.2030.284 Sand?.1490.2120.126 Slime0.141?.231?.078 Clay0.099?.108?.079 Mean annual temperature?.175?.2520.155 Isothermality0.234a?.204?.014 Temperature seasonality?.2200.253?.010 Maximum temperature of the
warmest period
?.249a0.0090.079
Minimum temperature of the
coldest period
0.095?.404a0.116 Annual temperature range?.245a0.164?.005 Annual rainfall0.2250.179?.062 Rainfall seasonality?.211?.104?.013 Rainfall of the driest quarter0.1840.300a0.007
a Indicates the highest values for each component
radial groups with shorter and narrower?bers.In contrast, B.cordata trees collected in most southern provenances grow in temperate subhumid climates with summer rain-fall,but with mean annual temperature of14?C,with low-annual temperature range of18?C along the year,and an-nual rainfall reaches1300mm with more than80mm of rain during the driest quarter.In these southern provenances individuals commonly possess diffuse porosity and shorter vessel groups,as well as longer and wider?bers.Ring and semiring-porous wood were also observed in some sites from the Central highlands like Zacatecas and Northeast-ern states.These provenances are characterized by a drier climate with the highest mean annual temperature(17?C), a lower annual rainfall(<500mm),and zero rainfall during the driest quarter.
Differences in humidity and temperature are re?ected in wood porosity present in the provenances of B.cordata along the distribution range and are not related to species leaf phenology because all individuals are evergreen.The typical ring-porous wood is distinctive of those individ-uals growing in provenances where cambial activity ces-sation occurs,caused by freezing,heat stress or drought (Chalk1983).These factors surely contributed to cam-bial inactivity in the northern and central provenances of B.cordata.True ring-porous wood occurred in the north-ern provenances with temperatures below zero,whereas in the North-Central provenances water stress is an addi-tional factor that contributes to porosity type and vessel trait modi?cations.Bissing(1982)found that porosity type and abundance of solitary vessels varied with water availabil-ity and the results mentioned for B.cordata support this assertion.Higher abundance of vessel grouping is inter-preted as a way to avoid embolisms(Baas1986).Changes in wood porosity not related to B.cordata phenology make this species an interesting system to study cambial activity associated with water conduction and photosynthesis. Plant height and?ber length in B.cordata as shown by CCA supports an allometric relationship also reported for other genera(Rury1985;Zhang1992;Terrazas1994; Noshiro and Baas1998;Terrazas and Loza-Cornejo2003). However,these results differ from those found for plant height and?bers length in three species of Cornus(Noshiro and Baas2000).No tendency was detected for?ber intru-sive growth in relation to B.cordata provenance distribu-tion or plant size.However,a F/V ratio between2.2and 3.0has been interpreted as an adaptation to favor longer ?bers to acquire an optimum mechanical system(Carlquist 1988).Latitude is also an important source of variation closely related with temperature and rainfall as shown by CCA.Chalk(1983)mentioned a tendency to decrease cel-lular element size of secondary xylem as latitude increases. This tendency is con?rmed in B.cordata.On the other hand,?ber length and ray height showed a negative cor-relation with latitude and the four climate variables de-tected by CCA,in contrast to the report for?ber length in Cornus macrophylla(Noshiro and Baas2000).In Acacia melanoxylon there is a positive correlation between vessel element lengths,proportion of?bers,and multiseriate rays with latitude;but a negative correlation with vessel den-stiy and diameter,proportion of uniseriate rays,and axial parenchyma(Wilkins and Papassotiriou1989).Notably,in B.cordata axial parenchyma does not vary as function of latitude or elevation.
The in?uence of elevation on wood has been studied in different taxa,however the results do not show a consistent pattern(Noshiro and Baas2000).In Metrosideros poly-morpha physiological and morphological variation along an altitudinal gradient was found,and size of this species decreases as elevation increases(Cordell et al.1998). Garc′?aXinying et al.(1988)found positive correlations be-tween elevation and vessel element length,vessel diameter,?ber–tracheid length and diameter,and ray height for an elevation range between1000and1800m.In B.cordata
259
Principal Component 2
? ?
6 ?
? ?Chih 1
? ?
5 ?
? ? ?Chih 2
?
4 ?
? ? ? ?
3 ?
? ? ? ?
2 ?
? ? Chis1
?SLP
? 1
? ? ? ? ? Dgo1Tlax
?
Pue
? Chis2
?
? Dgo2
? Qro1
?1 ?
? Zac
?
Qro2
? Mex
Oax
?
Zac-Dgo Ags
?2 ? Pue-Oax
? ? ?
?3 ? ?
?
?????????????????????????????????????????????????????????????????????????????????????????????
?6
?4
?2
2
4
6
8
Principal component 1
Fig.4.Principal component analysis based on wood,climate,and soil variables,plotted for provenances where individuals of B.cordata were collected;see Table 1for provenance information.Provenances:Chih1=Chihuahua 1,
Chih2=Chihuahua 2,Dgo1=Durango 1,Dgo2=Durango 2,Dgo-Zac =Durango-Zacatecas,Zac =Zacatecas,Ags
=Aguascalientes,SLP =San Luis Potos′?,Qro1=Quer′e taro 1,Qro2=Quer′e taro 2,Tlax =Tlaxcala,Mex –M′e xico,Pue =Puebla,Pue-Oax =Puebla-Oaxaca,Chis1=Chiapas 1,Chis2=Chiapas 2
elevation is not an important source of variation according to PCA and CCA,despite an elevation range from 1350m in Durango to 3040m in Puebla,as is also the case in Dodonaea viscosa despite its world-wide distribution (Liu and Noshiro 2003).Moreover,each species attains different strategies to adjust wood anatomy to environmental con-ditions associated with elevation or latitude which may be related to its genetic background.
The soil parameters,electrical conductivity,phosphorus and nitrogen had the highest loadings in the third com-ponent,explaining only 10%of the remaining variance;however they were not recovered by CCA as important variables to explain wood variation in B.cordata .Al-though minerals are associated with cellular division and differentiation in the vascular cambium,this relationship is also affected by environmental conditions (D¨u nisch and Bauch 1994);for example,at high latitudes,low temper-atures delay the carbon and nitrogen cycles in the surface soil (Currie 1999),thus electrical conductivity,phospho-rus and nitrogen may in?uence B.cordata wood through climate.This suggests that B.cordata is able to grow un-der different soil properties without in?uencing wood traits directly.
We conclude that B.cordata wood characters such as the occurrence of simple perforation plates,alternate in-tervascular pits,vessel–ray pits with reduced borders,ves-sel elements with helical thickenings,scanty paratracheal parenchyma,and heterogeneous rays are not signi?cantly in?uenced by plant size,climate,or species distribution.These wood features have been reported as distinctive for Buddleja wood (Aguilar-Rodr′?guez and Terrazas 2001),thus we suggest they have taxonomic value.Moreover,F /V ratio has not been in?uenced by plant size,climate or species distribution.Nonetheless,there is an evident
260
Table4.Canonical correlation analysis(CCA)of environmental variables (distribution)versus wood characters(wood)
First canonical variate Second canonical variate
Correlation Coef?cient Correlation Coef?cient
Wood
a
Diameter of wide vessel?0.2970.088?0.68a0.55
Diameter of narrow vessels?0.310.096?0.5740.426
Fiber length?0.734a0.5380.2840.612
Ray height?0.727a0.528?0.1990.568
Ray width?0.3940.155?0.5350.441
Plant height?0.766a0.586?0.070.591
Percent of variance0.6660.083Total=0.75 Redundancy0.6610.08Total=0.74 Distribution
Latitude0.831a0.691?0.0590.695
Maximum temperature of
the warmest period
0.832a0.6920.0180.693
Annual temperature range0.842a0.710.0320.71
Temperature seasonality0.7660.587?0.1040.598
Annual rainfall?0.768a0.591?0.231a0.644
Percent of variance0.6540.078Total=0.73 Redundancy0.730.08Total=0.81 Canonical correlation0.9820.942
a Indicates the highest values
allometric relationship between plant size and?ber length and ray height.Porosity and vessel grouping varied nega-tively with latitude as well as with temperature and rain-fall distribution along the year.The few studies of wood variation at the species level have revealed contrasting re-sults.Tracheary element size responds differently to en-vironmental factors in which species are distributed,as it has been the case in B.cordata.In some northern prove-nances,low temperatures probably favored B.cordata vas-cular cambium inactivity as PCA and CCA suggested.In these provenances a distinctive ring-porous wood occurs, whilst in more central localities,the lack of rainfall during the driest months of the year probably is the main fac-tor that induces vascular cambium inactivity,as well as a higher variation in ring porosity type and vessel grouping in latewood.To understand the in?uence of climate(e.g.tem-perature and rainfall)on wood features of species widely distributed in the intertropical region,reciprocal transplants among selected provenances of particular species should be conducted.
Acknowledgements This work was partially supported by Con-sejo Nacional de Ciencia y Tecnolog′?a scholarship to Silvia Aguilar-Rodr′?guez(118733)and Colegio de Postgraduados.We appreci-ate the contribution of J.Daniel Tejero D′?ez(Universidad Nacional Aut′o noma de M′e xico-FES Iztacala)for identifying all the material collected for this study and to Araceli Cort′e s Gonz′a lez for dark-room assistance
References
Aguilar-Rodr′?guez S,Terrazas T(2001)Anatom′?a de la madera de Buddleja L.(Buddlejaceae):an′a lisis fen′e tico.Madera y Bosques7:63–85Arias S,Terrazas T(2001)Variaci′o n en la anatom′?a de la madera de Pachycereus pecten-aboriginum(Cactaceae).Anal Inst Biol, Univ Nac Aut M′e x,Ser Bot72:157–169
Baas P(1986)Ecological patterns in xylem anatomy.In:Givnish TJ (ed)On the economy of plant form and function.Cambridge University Press,Cambridge,MA,pp327–352
Berlyn GP,Miksche JP(1976)Botanical microtechnique and cytochemistry.Iowa State University Press,Ames,Iowa Bissing DR(1982)Variation in qualitative anatomical features of the xylem of selected dicotyledonous woods in relation to water availability.Bull Torrey Bot Club68:317–326
Busby JR(1986)Bioclimatic Prediction System(BIOCLIM)User’s Manual Version2.0Australian Biological Resources Study Lea?et
Carlquist S(1966)Wood anatomy of Compositae:a summary,with comments on factors controlling wood evolution.Aliso6:25–44 Carlquist S(1988)Comparative wood anatomy.Systematic ecolog-ical,and evolutionary aspects of dicotyledon wood.Springer, Berlin Heidelberg New York
Carlquist S,Hoekman DA(1985)Wood anatomy of Staphyleaceae: ecology,statistical correlations,and systematics.Flora 177:195–216
Chalk L(1983)The effects of ecological conditions on wood anatomy.In:Metcalfe CR,Chalk L(eds)Anatomy of the dicotyledons,vol.II,2nd edn.Clarendon Press,Oxford,pp 152–156
Chattaway MM(1936)Relation between?bre and cambial initial length in dicotyledonous woods.Trop Woods46:16–20 Cordell S,Goldstein G,Mueller-Dombois D,Webb D,Vitousek PM(1998)Physiological and morphological variation in Metrosideros polymorpha,a dominant Hawaiian tree species, along an altitudinal gradient:the role of phenotypic plasticity.
Oecologia113:188–196
Currie SW(1999)The responsive C and N biogeochemistry of the temperate forest?oor.Trees14:316–320
Dickison WC,Phend KD(1985)Wood anatomy of the Styracaceae: evolutionary and ecological considerations.IAW A Bull ns 6:3–22
Dickison CW,Rury PM,Stebbins GL(1978)Xylem anatomy of Hibbertia(Dilleniaceae)in relation to ecology and evolution.J Arn Arb59:32–49
261
D¨u nisch O,Bauch J(1994)In?uence of mineral elements on wood formation of old growth spruce(Picea abies[L.]Karst.).
Holzforschung48:5–14
Etchevers BJD(1988)An′a lisis qu′?micos de suelo y planta.Programa de Edafolog′?a,Colegio de Postgraduados,Montecillo
IAW A Committee(1989)IAW A list of microscopic features for hardwood identi?cation.IAW A Bull ns10:219–332
Garc′?a E(1973)Modi?caciones al sistema de clasi?caci′o n clim′a tica de K¨o ppen.Instituto de Geograf′?a.Universidad Nacional Aut′o noma de M′e xico.M′e xico.Offset Larios,M′e xico,DF Johansen DA(1940)Plant microtechnique.McGraw-Hill,New York Li-Hsia Ch,Tseng-Chieng H(1986)Anatomical study of leaf and stem of Formosan Pittosporum,as an ecological implication.
Tawania23:41–63
Lindorf H(1997)Wood and leaf anatomy in Sessea corybi?ora from an ecological perspective.IAW A J18:157–168
Liu J,Noshiro S(2003)Lack of latitudinal trends in wood anatomy of Dodonaea viscosa(Sapindaceae),a species with a worldwide distribution.Am J Bot90:532–539
Media Cybernetics(1997)Image-Pro Plus reference guide,version
3.1for windows.Silver Spring,Maryland
Norman ME(1966)The genus Buddleia in North America.Gentes Herbarium10:47–116
Norman ME(2000)Buddlejaceae.Flora Neotropica Monograph81.
The New York Botanical Garden,Bronx,New York
Noshiro S,Baas P(1998)Systematic wood anatomy of Cornaceae and allies.IAW A J19:43–97
Noshiro S,Baas P(2000)Latitudinal trends in wood anatomy within species and genera:case study in Cornus s.L.(Cornaceae).Am J Bot87:1495–1506
Parameswaran N,Conrad H(1982)Wood and bark anatomy of Balanites aegyptiaca in relation to ecology and taxonomy.
IAW A Bull ns3:75–88
Quintas I(2000)Eric II.Documentaci′o n de la base de datos climatol′o gica y del programa extractor.Instituto Mexicano de Tecnolog′?a del Agua.Mexico,DF
Rzedowski J(1986)Vegetaci′o n de M′e xico.Limusa,M′e xico,D.F. Rzedowski GC de,Rzedowski J(2001)Flora fanerog′a mica del Valle de M′e xico.Instituto de Ecolog′?a,A.C.y Comisi′o n Nacional para el Conocimiento y Uso de la Biodiversidad.P′a tzcuaro, Mich Rury MP(1985)Systematic and ecological wood anatomy of the Erythroxylaceae.IAW A Bull ns6:365–397
SAS(1989)SAS procedures guide,version6.04.Cary,NC Sastrapradja SD,Lamoureux Ch(1969)Variations in wood anatomy of Hawaiian Metrosideros(Myrtaceae).Ann Bogorienses 5:1–83
Sidiyasa K,Baas P(1998)Ecological and systematic wood anatomy of Alstonia(Apocynaceae).IAW A J19:207–229 Tabachnick BG,Fidell LS(1989)Using multivariate statistics,2nd edn.Harper Collins,New York
Terrazas T(1994)Wood anatomy of Anacardiaceae:ecological and phylogenetic interpretation.Ph.D.Dissertation,University of North Carolina-Chapel Hill,NC
Terrazas T,Loza-Cornejo S(2003)Anatom′?a de la madera y alometr′?a de veinte especies de Stenocereus(Cactaceae).
Anales Inst Biol Univ Nac Aut′o n M′e xico Bot74:195–208
van der Oever L,Baas P,Zandee M(1981)Comparative wood anatomy of Symplocos and latitude and altitude of provenance.
IAW A Bull ns2:3–24
Villagra EP,Roig-Ju?n ent FA(1997)Wood structure of Prosopis alpataco y P.argentina growing under different edaphic conditions.IAW A J18:37–51
Wilkins AP,Papassotiriou S(1989)Wood anatomical variation of Acacia melanoxylon in relation to latitude.IAW A Bull ns 10:201–207
Xinying Z,Liang D,Baas P(1988)The ecological wood anatomy of the Lilacs(Syringa oblata var giraldii)on mount Taibei in Northwestern China.IAW A Bull ns9:24–30
Yaltirik F(1970)Comparison of anatomical characteristics of wood in Turkish maples with relation to the humidity of the sites.J Wood Sci5:43–48
Y′a?n ez-Espinosa L,Terrazas T,L′o pez-Mata L(2001)Effects of ?ooding on wood and bark anatomy of four species in a mangrove forest community.Trees15:91–97
Zhang S-Y(1992)Wood anatomy of the Rosaceae.Rijksherbarium Hortus Botanicus,Leiden
Zar J.(1999)Biostatistical analysis.Prentice Hall,Englewood Cliffs,NJ