Skip to main content

Full text of "Proceedings of the Southern Pasture and Forage Crop Improvement Conference"

See other formats 


Historic,  Archive  Document 

Do  not  assume  content  reflects  current 
scientific  knowledge,  policies,  or  practices. 


(fiO, /? 

So  V3 


United  States 
Department  of 
Agriculture 


Agricultural 

Research 

Service 


December  1988 


Proceedings  of  the 
44th  Southern  Pasture 
and  Forage  Crop 
improvement 
Conference 

Held  at  Lexington,  Kentucky 
May  10-12,1983 


The  papers  presented  in  this  report  are  repro 
duced  as  supplied  by  the  authors  and  do  not 
necessarily  reflect  the  views  and  opinions  of 
the  United  States  Department  of  Agriculture. 
Mention  of  particular  tradenames,  products  or 
companies  do  not  imply  preferential  recommenda- 
tion by  USDA  over  comparable  products  or 
services.  Specific  questions  regarding 
information  presented  in  particular 
manuscripts  should  be  directed  to  the 
manuscript  author. 

Ihe  44th  proceedings  contents  were  collated, 
edited  and  submitted  for  ARS  publication  by 
D.  P.  Belesky,  Appalachian  Soil  and  Water 
Conservation  Research  Laboratory,  P.0.  Box  867, 
Airport  Road,  Beckley,  WV  26802. 

Copies  of  this  report  can  be  purchased  from 
National  Technical  Information  Service,  5285 
Port  Royal  Road,  Springfield,  VA  22161. 


PREFACE 

Through  the  years,  the  Southern  Pasture  and 
Forage  Crop  Improvement  Conference  (SPFCIC)  has 
come  to  be  regarded  as  the  nation's  best 
regional  forage  conference,  and  the  44th 
certainly  did  not  tarnish  that  image.  Credit 
for  its  overwhelming  success  must  be  attributed 
to  many  people,  but  we  must  be  especially 
grateful  to  the  local  planning  committee,  to 
the  program  chairs,  and  to  those  who  presented 
papers.  It  is  impossible  to  measure  the  impact 
of  a conference  such  as  this,  but  it 
unquestionably  is  immense.  Hopefully,  through 
these  proceedings,  the  44th  SPFCIC  will 
continue  to  benefit  forage  scientists  for 
years  to  come. 

These  proceedings  include  the  papers  and 
reports  presented  at  the  44th  meeting  of  the 
Southern  Pasture  and  Forage  Crop  Improvement 
Conference  (SPFCIC)  held  May  10-12,  1988  in 
Lexington,  Kentucky,  with  the  University  of 
Kentucky  as  host.  Papers  presented  in  those 
sessions  appear  on  the  following  pages. 

Minutes  of  business  meetings  of  the  Executive 
Committee  and  the  respective  work  groups  are 
also  included  in  the  44th  Proceedings. 


Don  Ball 

Chairman,  44th  SPFCIC 


i 


CONTENTS 

The  Role  of  Cooperative  and  Collaborative 
Research  in  Forages 

Cooperative  (Interdepartmental)  Research:  Pros 
and  Cons  1 

J.C.  Burns,  D.S.  Fisher,  and  K.R.  Pond 

Regional  and  Multisite  Cooperative  and 
Collaborative  Research  6 
J.W.  Holloway 

Grazing  Experiments  - Objectives,  Compromises 
and  Appropriate  Designs 

Objectives  of  Grazing  Research  and  Their 
Implications  for  Experimental  Design  7 
A.G.  Matches 

Advantages  of  a Non-Replicated  Multiple  Grazing 
Intensity  Approach  for  Grazing  Trials  11 
David  I.  Bransby 

Statistical  Aspects  of  Grazing  Experiment  15 
J.  Wanzer  Drane 


Forage  Breeders  Information  Exchange  Group 

Breeding  Self  Pollinated  Forage  Legumes  22 
D.D.  Baltensperger 

Tannins  in  Forage  Legumes  and  Implications  for 
a Breeding  Program  25 
K.H.  Queensberry 

History  of  Cool  Season  Grass  Breeding  in  the 
Southeast  29 
J.F.  Pedersen 

Forage  Germplasm  Evaluation  33 
M.A.  Hussey  and  D.I.  Bransby 

Ecology  and  Physiology  Information  Exchange 
Group 

Influence  of  the  Fungal  Endophyte  on  Physiology 
and  Ecology  of  Tall  Fescue  37 
C.S.  Hove land 

The  Ecology  and  Physiology  of  Cool-Season 
Forages  Under  Intensive  Rotational  Grazing 
Systems  41 

C. T.  Dougherty 

Ecology  and  Physiology  of  Warm  Season  Forages 
in  Intensive  Rotational  Grazing  Systems  44 
F.M.  Rouquette,  Jr. 

Carbofuran  for  Forage  Establishment:  An  Update  48 

D. D.  Wolf 


Forage  Utilization  Information  Exchange  Group 

Use  of  NIRS  to  Predict  Botanical  Composition 
of  Forage  Mixtures  51 

N.S.  Hill,  J.A.  Stuedemann,  and  G.O.  Ware 

Forage  Intake  as  Influenced  by  Sward 
Characteristics  55 
T.D.A.  Forbes 

Biological,  Practical  and  Statistical 
Considerations  Associated  With  Measuring 
Forage  Availability  in  Grazing  Trials  59 
David  I.  Bransby  and  G.  Peter  Clarke 

Effect  of  Endophyte  Level  of  Tall  Fescue  on 
Subsequent  Feedlot  Performance  of  Steers  64 
N.  Andy  Cole 

Forage  Management  in  an  Integrated  Beef- 
Forage  System  in  Arkansas,  A Total  Farm 
Management  Approach  69 
B.J.  Hankins 


General  Business  and  Information  Exchange 
Group  Meetings 

Minutes  of  the  Business  Meeting  44th  SPFCIC 
May  12,  1988  Lexington,  Kentucky  72 

Minutes  of  the  Executive  Committee  44th  SPFCIC 
May  10,  1988  Lexington,  Kentucky  72 

Resolution  Adopted  Unanimously  by  the  44th 
Annual  Southern  Pasture  and  Forage  Crop 
Improvement  Conference  73 

Southern  Pasture  and  Forage.  Crop 
Improvement  Conference  Executive 
Committee  1989  73 

SPFCIC  Breeders  Work  Group,  May  11,  1988  74 

Ecology  and  Physiology  Work  Group  Business 
Meeting  74 

Minutes  of  the  SPFCIC  Forage  Utilization 
Work  Group  Business  Meeting,  Lexington,  KY 
May  11,  1988  75 

Extension  Work  Group  Minutes  76 

Southern  Pasture  and  Forage  Crop  Improvement 
Conference  1988  Financial  Statement  76 

44th  SPFCIC  Registrants  77 


THE  ROLE  OF  COOPERATIVE  AND  COLLABORATIVE 
RESEARCH  IN  FORAGES 

COOPERATIVE  (INTERDEPARTMENTAL)  RESEARCH:  PROS 
AND  CONS1 

J.C.  Burns,  D.  S.  Fisher,  and  K.  R.  Pond^ 
INTRODUCTION 

The  terras  cooperative  or  team  research,  as  well 
as  interdepartmental,  multidisciplinary  or 
joint  research,  are  frequently  used  inter- 
changeably, inferring  that  two  or  more  indi- 
viduals are  working  together  with  shared  re- 
search objectives.  These  individuals  generally 
provide  different  expertise  (disciplines)  per- 
mitting additional  dimensions  of  a particular 
problem  to  be  studied.  The  wisdom  of  inter- 
departmental research  is  that  results  may  be 
more  encompassing  thereby  achieving  greater 
knowledge  with  an  increased  efficiency. 

Cooperative  research  involves  a dynamic  matrix 
of  scientific  and  technical  personnel  that 
generate  intentions,  commitments,  and  expecta- 
tions in  both  researchers  and  administrators. 
The  same  matrix  includes  facilities  and  the 
source  and  quantity  of  funding  necessary  to 
drive  the  total  effort.  This  matrix  is  in 
constant  flux  because  of  individual  differen- 
ces, shifts  in  funding  and  constant  changes 
that  occur  in  federal  and  state  regulations. 

The  aspects  of  intention  and  commitment  in 
cooperative  research  and  the  subsequent  advan- 
tages and  disadvantages  of  such  cooperation  are 
discussed. 


THE  INTENTION 
Scientist 

The  first  step  in  the  formation  of  cooperative 
research  generally  resides,  and  properly  so, 
with  scientists  sharing  ideas  and  developing  an 
image  of  what  might  be  achieved  through  coop- 
eration. Following  discussions  (generally  sev- 
eral over  time)  intentions  develop  and  are 
shared  regarding  future  cooperation  on  a spe- 
cific objective.  Such  intentions  imply  only 
what  each  individual  has  in  mind  to  do  or  to 
bring  about  as  a result  of  his/her  involvement. 


^Cooperative  investigation  of  the  United  States 
Department  of  Agriculture,  Agricultural  Re- 
search Service,  and  the  North  Carolina  Agricul- 
tural Research  Service,  Raleigh,  NC.  Paper  No. 

11602  of  the  Journal  Series  of  the  North 
Carolina  Agricultural  Research  Service, 

Raleigh,  North  Carolina.  27695-7601. 

^Plant  Physiologist,  USDA,  ARS  and  Professor, 
Dept,  of  Crop  Science  and  Animal  Science;  Plant 
Physiologist,  USDA,  ARS  and  Assistant  Profes- 
sor, Dept,  of  Crop  Science;  Associate  Profes- 
sor of  Animal  Science,  North  Carolina  State 
Univ.  Raleigh,  N.C. 


Further  discussions  may  ensue  which  reaffirm 
the  previous  intentions  and  usually  generate 
even  greater  interest.  At  this  point  no 
commitment  may  be  intended  or  felt  by  any  dis- 
cussant. However,  intention  levels  have  been 
generated  and  the  potential  cooperators  must 
begin  discernment  between  intention  and  poten- 
tial commitment.  Discernment  is  an  individual 
process  involving  intuition  or  may  only  amount 
to  a guess.  This  interest  category  of  inten- 
tion seems  best  represented  by  the  designation 
"cooperator"  on  many  research  proposals. 

Administration 

In  most  situations  involving  cooperative  re- 
search, assistance  from  the  administration  at 
the  work  unit  or  departmental  level  is  essen- 
tial. Occasionally,  involvement  of  the  admini- 
stration at  a higher  level  is  required.  Admin- 
istrative involvement  is  needed  if  the  magni- 
tude of  the  cooperative  effort  requires  addi- 
tional funding,  realignment  of  or  addition  of 
facilities.  Another  area  involves  relaxing  of 
the  operational  policies  among  the  cooperator's 
unit  (such  as  a work  unit  within  the  federal 
system  or  a department  within  the  state  system) 
so  work  can  be  done  with  minimal  effort.  Rig- 
orous unit  policies  can  be  exhaustive  when 
encountered  daily. 

Enthusiasm  and  encouragement  from  the  admini- 
stration for  interdepartmental  research  is 
generally  forthcoming  and  the  intention  to 
assist  with  such  problems  generally  conveyed. 
Such  intentions  again  imply  little  more  than 
what  the  administration  might  do  to  enhance  the 
general  process.  This  encouragement  should  be 
taken  as  such,  but  is  frequently  misinterpreted 
by  scientists  as  commitment.  This  is  clearly  an 
erroneous  assumption,  and  the  result  of  opti- 
mistic and  enthusiastic  minds  oriented  toward 
research  achievement.  Generally,  no  effort  is 
made  to  clarify  intention  vs.  commitment,  as 
both  researchers  and  administrators  would 
rather  allow  the  truth  to  remain  in  the  gray 
area. 


THE  COMMITMENT 

The  heart  of  continuous,  fruitful  cooperative 
research  resides  in  the  commitment  of  both  the 
cooperators  and  administrators  to  achieve  the 
identified  objectives.  Commitment  infers  an 
agreement  (or  pledge)  to  contribute  as  appro- 
priate to  the  success  of  the  total  effort. 

This  is  a step  beyond  "good  intentions"  as  an 
obligation  has  been  accepted.  The  obligation 
binds  one  to  a specific  course  of  action,  but 
without  tangible  penalties  for  nonfulfillment. 
Ultimately,  the  penalty  is  incomplete  data 
leaving  the  team  short  of  their  objectives  and 
the  reneging  individual  labeled  as  a poor  co- 
operator. 

The  status  of  a cooperator  with  commitment  is 
generally  categorized  on  many  research  projects 
as  a co-leader.  It  implies  a contribution  of 
the  cooperator's  time,  funds,  technical  sup- 


1 


port,  or  facilities  either  singly  or  in  all 
combinations  to  achieve  the  objectives  of  the 
research. 

THE  EXPECTATION 

Expectation  is  the  perception  that  an  event  has 
a reasonable  probability  of  occurring.  Expec- 
tations are  essential  in  sustaining  cooperative 
research,  but  are  different  for  scientists  and 
administrators. 

Scientist 

Expectations  that  a scientist  forms  when  enter- 
ing cooperative  studies  begin  early  during 
initial  discussions  and  become  clearer  as  dis- 
cussions continue.  Expectations  are  initially 
based  on  the  opportunity  to  achieve  knowledge 
and  insight  which  is  less  possible  through  in- 
dependent research.  During  the  course  of  coop- 
erative research,  there  is  a gradual  shift, 
probably  unintentional,  in  expectations.  The 
initial  drive  to  acquire  new  knowledge  becomes 
secondary  to  reward  expectations  which  include 
publication  opportunity,  program  enhancement, 
monetary  compensation  and  peer  recognition. 

This  array  of  expectations  will  develop  regard- 
less of  the  individual's  degree  of  involvement 
(i.e.,  intentions  or  commitment)  in  the  pro- 
ject. A contributing  member  with  intention, 
but  without  commitment,  will  usually  find  his/- 
her expectations  frustrated.  Unfortunately, 
intentions  generally  result  in  minor,  if  any, 
contribution  thus  subsequent  recognition  is 
usually  nonexistent.  At  the  same  time,  expecta- 
tions can  be  sufficiently  high  that  even  the 
committed  person  (i.e.,  one  making  appreciable 
contribution)  can  experience  frustration  even 
though  recognition  was  reasonable  and  shared  by 
the  total  team.  In  this  case,  recognition 
received  did  not  meet  the  level  of  expectation. 
This  must  be  guarded  against  by  communication 
among  scientists  and  administrators  in  main- 
taining continuing  cooperative  projects. 

Administration 

The  concept  of  cooperative  or  team  research 
appears  widely  accepted  and  desirable  among 
administrators  with  some  feeling  that  in  cer- 
tain areas  it  is  essential.  This  has  probably 
resulted  from  the  realization  by  scientists  and 
administrators  that  much  of  the  knowledge  to  be 
gained  in  biology  (plants  or  animals)  resides 
in  studying  interactions.  The  effective  study 
of  many  biological  systems  results  in  complex 
experiments  that  require  expertise  from  several 
disciplines  and/or  research  projects. 

Surprisingly,  many  administrators  admit  to  not 
having  any  formula  to  initiate  or  sustain  ac- 
tivity in  cooperative  studies.  The  tendency  is 
to  remain  insulated  as  much  as  possible  from 
the  practical  aspects  of  the  process,  but  pres- 
ent it  as  part  of  the  unit's  program  when 
viewed  from  the  outside.  An  administrator's 
expectation  from  cooperative  studies  is  that  he 
will  be  recognized  by  peers,  and  upper  admini- 


stration, as  a talented  and  successful  leader. 

In  this  case  the  expectation  is  based  on  the 
process  and  not  the  value  or  accomplishments  of 
the  research  effort.  Consequently,  administra- 
tor's expectations  from  cooperative  research 
are  probably  met  more  often  than  are  scien- 
tist 's. 


THE  REALITY 

The  conduct  on  a daily  basis  of  interdepart- 
mental research,  by  necessity,  involves  asso- 
ciation among  a wide  array  of  personalities. 
Attitudes,  feelings  and  egos  of  individuals 
ranging  from  secretarial  to  technical  to  scien- 
tific all  operate  within  the  confines  of  the 
available  facilities  at  the  existing  level  of 
funding.  These  interactions  vary  widely  from 
project  to  project  both  within  and  among  work 
units.  Furthermore,  each  cooperative  project 
generally  resides  within  a unit  much  larger 
than  the  cooperating  team.  In  addition,  indi- 
viduals who  compose  the  team  are  frequently 
from  units  or  departments  composed  of  many  in- 
dividuals who  directly  impact  the  success  of 
the  team  without  direct  contribution.  Variables 
operating  inside  and  outside  the  cooperative 
project  lead  to  a number  of  advantages  and 
disadvantages  that  can  be  generalized  to  cover 
most  cooperative  or  team  research  efforts. 

These  are  listed  below  and  discussed  as  pros 
and  cons  of  interdepartmental  research.  It 
should  be  noted  that  not  all  points  are  intend- 
ed to  apply  to  every  research  effort,  but  many 
can  be  identified  as  applying  to  most  coopera- 
tive research  ventures. 

Pros  of  Interdepartmental  Research 

1)  Focus  of  More  Scientific  Expertise. 
Scientists  from  several  disciplines  can  bring 
more  scientific  expertise  together  to  delineate 
the  problems  and  the  appropriate  methodology  to 
arrive  at  the  measurements  required  to  meet  the 
research  objectives. 

2)  Study  Multiple  Factors. 

Joint  effort  permits  consideration  of  more 
factors  that  have  potential  bearing  on  the 
research  objectives.  In  individual  research, 
factors  known  to  impact  the  research  objectives 
may  lie  outside  of  the  expertise  of  the  scien- 
tist, and  are  frequently  ignored  or  held  con- 
stant. Team  research  permits  such  factors  to  be 
studied . 

3)  Detailed  Measurements. 

Each  cooperator  has  responsibility  for  his/her 
area  of  interest.  This  concentration  permits 
detailed  measurements  to  adequately  character- 
ize the  changes  in  the  factor(s)  that  have  a 
bearing  on  the  response(s)  of  interest. 

4)  Common  Experimental  Units. 

Measurements  on  or  relative  to  the  same  experi- 
mental unit  provide  the  opportunity  to  examine 
general  associations  of  cause  and  effect  re- 
sponse between  or  among  a number  of  measures. 
Cost  efficiency  seems  inherent  in  such  studies 


2 


since  creating  a particular  biological  state 
occurs  once  for  all  measurements  as  opposed  to 
recreating  (if  possible)  the  state  if  measure- 
ments were  taken  independently.  This  efficiency 
may  not  always  exist  because  more  expense  may 
be  involved  in  personnel,  equipment  or  facili- 
ties to  obtain  all  measurements  simultaneously. 
It  is  possible  that  part  of  the  measurements 
may  never  be  obtained  if  left  for  separate  ex- 
periments. Knowledge  obtained  in  the  latter 
case  would  not  be  as  complete  or  as  beneficial 
as  when  all  measurements  are  taken  simultane- 
ously. 

5)  Potential  for  Major  Research  Advancement. 

The  proper  mix  of  skilled  scientists  in  a re- 
search setting  increases  through  observation, 
measurements  and  integration,  the  probability 
of  obtaining  new  and  important  information  that 
will  advance  understanding  of  the  processes 
being  studied. 

6)  Continuous  vs  Discrete  Data. 

Frequently  individual  studies  will  be  frag- 
mented because  of  the  number  of  factors  that 
must  be  measured  and  the  frequency  that  meas- 
urements may  be  required.  With  sufficient 
scientific,  technical,  and  instrumentation  ca- 
pability, frequent  measurements  permit  trend- 
type  analyses  resulting  in  a descriptive  rela- 
tionship as  opposed  to  a selected  number  of 
discrete  samplings  which  are  difficult  or  im- 
possible to  integrate  or  relate. 

7)  Recognition 

Recognized  by  peers  as  a member  of  a research 
team  involved  in  pioneering  work  can  be  ex- 
tremely rewarding.  The  work  being  published  is 
seen  as  a step  or  two  removed  from  traditional 
data  that  dominate  some  present  day  journals. 
Both  personal  satisfaction  and  monetary  bene- 
fits can  occur.  Recognition  may  include  in- 
vited presentations,  awards,  and  local  respect 
within  the  work  unit  or  department. 

Cons  of  Interdepartmental  Research 

1)  Increased  Cost 

The  number  of  factors  measured  and  the  detailed 
measurements  (physical,  chemical  and  biologi- 
cal) generally  require  an  increase  in  personnel 
(technical  and  support),  instrumentation,  phy- 
sical facilities  and  operational  funding.  This 
increase  may  not  be  forthcoming  because  the 
administration's  intentions  did  not  move  to 
commitment.  Consequently,  the  research  effort 
is  soon  underfunded,  the  cooperators  become 
discouraged,  and  fulfillment  of  the  objectives 
remain  incomplete. 

2)  Scope  Expanded  and  Efforts  Diluted. 

As  a member  of  a team,  each  scientist  becomes 
exposed  to  the  other  disciplines.  The  interac- 
tion of  factors  among  the  disciplines  draws  the 
attention  of  all  the  scientists  and  they  begin 
to  integrate  the  various  responses.  The  ten- 
dency is  to  shift  from  a narrow  approach  fre- 
quently followed  by  an  expert,  to  a wider  view 
taken  by  a generalist.  Focused  efforts  are 


diluted  by  time  commitment  to  the  team.  In 
personal  research  an  individual  can  make  major 
decisions  on  the  spur  of  a moment  and  move 
forward.  In  team  research  many  decisions  re- 
quire discussion.  Schedules  and  rescheduling 
for  joint  meetings  and  subsequent  discussions 
to  arrive  at  decisions,  although  necessary,  can 
be  extremely  time  consuming  and  exhausting. 

3)  Personality  Conflicts. 

The  opportunity  for  personality  conflicts  are 
greater  and  results  far  more  damaging  when  in- 
dividuals work  in  close  proximity  and  the  work 
conducted  by  one  affects  the  work  of  another. 
Some  individuals  can  handle  changing  situations 
and  work  through  problems  in  a consistent  and 
reasonable  manner.  Other  people  react  in  ex- 
treme ways  to  stress,  or  to  other  people,  and 
can  keep  a research  team  in  turmoil.  Clearly, 
not  all  people  are  emotionally  or  mentally 
suited  for  cooperative  research.  Many  times 
individuals  want  to  be  a part  of  a team  but 
fail  to  realize  their  limitations.  A small 
amount  of  unrest  can  nullify  major  efforts  to 
maintain  a smooth  cooperative  relationship. 

4)  Proper  Representation. 

The  concept  of  cooperative  or  team  research 
places  all  individuals  as  subservient  to  the 
team.  However,  in  most  research  structures 
such  a team  would  lack  administrative  support 
and  lose  funding.  Most  successful  teams  have  a 
spokesperson  who  promotes  the  team  at  the  right 
times  and  in  the  right  places  to  keep  the 
team's  activity  in  the  forefront.  This  may 
require  a spokesperson  in  each  team  members' 
work  unit  and  is  especially  important  in  sys- 
tems using  zero  base  budgeting. 

5)  High  Risk. 

Cooperative  research  has  a high  risk  component 
because  of  the  fragile  nature  of  the  research 
arrangement.  Productive  cooperative  research 
functions  in  much  the  same  way  as  does  negotia- 
tion. 

Negotiation  continues  and  is  successful  if  both 
sides  believe  they  are  winners.  If  one  side  is 
losing,  negotiations  cease.  In  cooperative 
research,  all  individuals  must  benefit  person- 
ally and  professionally  or  the  arrangement  will 
fail.  One  scientist  is  not  likely  to  serve 
another.  Since  there  may  not  be  tangible  pen- 
alties for  nonfulfillment,  there  are  no  legal 
levers  to  cause  continued  involvement. 

Another  aspect  adding  to  the  risk  factor  is  the 
difficulty  in  discerning  between  intention  and 
commitment.  Not  all  team  members  will  be  com- 
mitted to  the  same  degree  and  some  enter  coop- 
erative studies  as  if  committed,  but  really 
only  have  intentions.  Such  individuals  fail  to 
pull  their  weight  and  are  unlikely  to  contrib- 
ute as  they  agreed.  Although  not  always  evi- 
dent, there  is  general  apprehension  between 
scientists  and  the  administration.  Scientists 
resent  being  directed  but  look  toward  admini- 
stration for  leadership.  Administration  fre- 
quently fails  the  leadership  role  by  leading 


3 


only  with  words  and  intentions  and  remaining 
uncommitted.  "Hot  topics"  can  draw  the  admini- 
stration's attention  while  a research  team 
needs  stability  and  commitment  for  quality 
scientific  efforts. 

6)  Inadequate  Recognition. 

The  difficulty  in  providing  fair  recognition  to 
members  of  a productive  team  may  be  the  biggest 
negative  factor  in  cooperative  research.  The 
complexity  of  cooperation  exceeds  that  encoun- 
tered by  the  independent  researcher.  In  the 
first  place,  the  daily  conduct  of  work  in  a 
team  setting,  especially  when  team  members  are 
housed  in  different  locations  (units,  depart- 
ments, or  geographic  regions), is  more  compli- 
cated. Secondly,  the  nature  of  the  research  and 
the  interactions  examined  may  make  the  data 
more  difficult  to  interpret.  Yet,  the  independ- 
ent scientist  who  focuses  on  a specific 
(narrow)  problem  and  researches  the  area  in 
depth,  can  quickly  (6  to  8 years)  become  an 
"expert"  with  surprising  national  and  inter- 
national recognition.  This  requires  that  the 
individual  be  well-published  and  participate  in 
regional  and  national  meetings.  Such  a person 
wi’i  most  likely  move  quickly  through  the  pro- 
motion levels  in  either  the  federal  or  state 
systems.  Such  advancement  may  exceed  that  of  a 
team  researcher,  but  does  not  necessarily  indi- 
cate a more  accomplished  scientist.  This  is 
partially  associated  with  a noted  advantage  in 
most  promotion  systems  for  single  or  senior 
authored  publications.  Being  listed  as  the 
fourth  author  of  a team  research  paper  may  not 
greatly  help  one's  score  in  many  evaluation 
systems.  Yet,  cooperative  research  requires 
that  authorship  be  rotated  so,  in  fact,  the 
last  author  of  some  papers  may  have  made  as 
much  or  perhaps  even  a more  significant  contri- 
bution to  the  results  than  did  the  senior 
author.  The  designation  of  a team  leader  fur- 
ther complicates  this  process.  Evaluation  of 
authorship  needs  to  be  better  understood  when 
scoring  individuals  involved  in  cooperative  re- 
search. 

A solution  to  the  problem  of  senior  authorship 
in  promotion  and  which  is  generally  favored  by 
administrators,  is  to  maintain  in  addition  to 
the  cooperative  research,  a personal  research 
program.  While  this  can  benefit  the  establish- 
ment and  development  of  a scientist  in  his/her 
discipline  and  ultimately  the  team,  the  concept 
coming  from  the  administrative  level  reflects 
negatively  on  cooperative  research  and  will 
likely  be  reflected  in  subsequent  peer  evalua- 
tions. Furthermore,  it  places  an  unfair  burden 
on  the  individual  participating  in  cooperative 
research.  In  such  situations,  personal  re- 
search competes  strongly  with  the  cooperative 
effort  for  time  and  funds  and  hinders  the 
fruitfulness  of  cooperative  research.  Another 
approach  is  to  encourage  scientists  to  publish 
their  segment  of  a joint  project  prior  to  its 
integration  into  the  team  findings.  This 
causes  problems  with  co-authorship  and  unrea- 
sonable delays,  creating  frustration  and  loss 
of  team  interest.  Further,  prior  publication 


will  likely  relegate  publication  of  the  team's 
findings  to  lesser  vehicles  such  as  final  re- 
ports or  local  (regional,  state)  bulletins. 

The  reality  is  that  scientists  with  the  need  to 
have  his/her  own  program,  reduces  commitment  to 
the  cooperative  project  to  a level  of  inten- 
tion. While  the  expectations  of  administration 
are  met  by  this  solution,  they  fail  to  realize 
the  detrimental  effect  on  cooperative  research. 

Many  research  organizations  rely  on  a "peer 
evaluation"  system  to  initiate  or  determine 
promotion.  Use  of  peer  evaluation  appeals  to 
many  administrators  because  it  conveys  the 
notion  that  people  of  equal  knowledge,  experi- 
ence, etc.  are  making  the  evaluation.  The 
fault  in  most  peer  evaluation  systems  is  that 
the  peer  aspect  is  only  relevant  to  the  extent 
that  all  the  evaluators  are  other  scientists, 
or  depending  on  the  system,  just  colleagues 
(extension,  teachers).  Although  exceptions  do 
exist,  few  of  the  other  scientists  will  be  of 
the  same  discipline  creating  a problem  of  per- 
spective. For  example,  scientists  who  are 
trained  and  function  in  a laboratory  (plant  or 
animal)  or  greenhouse  setting  have  problems  ap- 
preciating the  need,  difficulty  and  complexity 
of  a five-year  study  with  a perennial  crop. 
Likewise,  scientists  who  must  conduct  long-term 
studies  have  difficulty  scoring  short-term 
laboratory  experiments. 

Another  aspect  of  team  member  recognition 
arises  when  the  scientists  are  located  in  dif- 
ferent units  or  different  departments.  Members 
of  interdepartmental  teams  may  be  differential- 
ly promoted  because  the  peer  structure  differs 
for  each  member  of  the  team,  and  the  philosophy 
for  promotion  can  vary  among  units.  Such  dif- 
ferentials can  become  obstacles  to  interdepart- 
mental research. 

The  peer  evaluation  concept  permits  minimal 
administrative  input  and  takes  away  one  of  the 
management  tools  important  to  an  administrator 
or  supervisor  in  developing  a sound,  productive 
unit.  This  problem  is  somewhat  alleviated  in 
systems  where  the  unit  supervisor  can  present 
the  scientist's  case  to  the  higher  administra- 
tion following  peer  evaluation.  Also,  discrep- 
ancies created  by  peer  evaluation  can  be 
smoothed  out  somewhat  in  systems  where  annual 
evaluation  and  salary  increases  are  determined 
mainly  by  the  unit  supervisor. 

SUMMARY 

Team  research  is  a dynamic  matrix  involving  the 
interaction  of  individual  traits  of  scientific, 
technical  and  support  personnel.  These  must  be 
integrated  into  the  capabilities  of  the  re- 
search facilities.  Special  consideration  needs 
to  be  given  by  all  parties  initiating  coopera- 
tive research  to  separate  intentions  from  com- 
mitment as  cooperative  teams  are  formed. 

Success  in  cooperative  research  can  be  extreme- 
ly fulfilling  and  rewarding.  However,  problems 


4 


can  and  do  arise  and  these  should  be  considered 
carefully  as  they  can  have  major  consequences 
on  people's  careers.  Most  vulnerable  to  the 


pitfalls  of  cooperative  research  are  young 
scientists  striving  to  become  established  in 
their  disciplines. 


Figure  1. 

State  agricultural  experiment  stations  in  the 
Southern  region:  stars,  main  stations;  closed 
circles,  substations. 


5 


REGIONAL  AND  MULTISITE  COOPERATIVE  AND 
COLLABORATIVE  RESEARCH 

J.W.  Holloway^ 


THE  REASON  FOR  COOPERATIVE  AND  COLLABORATIVE 
RESEARCH 

The  most  pervasive  problems  with  agricultural 
research  arise  from  the  intrinsic  nature  of  the 
research  and  are  results  of  the  goals  of  re- 
search. The  goals  are:  (1)  development  of 
unifying  theorems  "tying  together"  observations 
previously  thought  to  be  unrelated,  and 
(2)  development  of  broadly  applicable  tech- 
nology. The  problems  result  from  the  fact  that 
agriculture  is  performed  on  a shallow  layer  of 
the  earth's  crust  characterized  by  disconti- 
nuities (topographical,  political  and  physio- 
graphical)  and  continuous  expanses.  The 
problem,  therefore  is  logistical  in  nature  in 
that  great  expanses  must  be  encompassed  to 
study  the  variety  of  conditions  underwhich  agri- 
culture is  performed. 

The  goals  of  agricultural  research  can  be. viewed 
as  being  accomplished  in  a two  prong  approach: 

(1)  Observe  and  evaluate  interactions.  If 
none  exist,  the  logistical  problem  is  minimized. 

(2)  Explain  in  terms  of  fundamental  princi- 
ples. These  principles  are  developed  at  the 
order  just  lower  than  that  of  interest  in 
hierarchies  such  as:  ecosystems,  populations, 
organisms,  systems,  organs,  tissues,  cells, 
organelles,  molecules,  atoms. 

Since  logistical  problems  are  minimized  at  the 
lower  ends  of  these  hierarchies,  more  progress 
has  been  made  at  these  levels.  The  machinery, 
however,  is  available  for  studying  the  higher 
levels  and  is  the  network  of  experiment  stations 
present  in  most  states  (Figure  1).  This  experi- 
mentation is  more  expensive  than  we  like  to  admit 
and  is  often  subsidized  in  ways  we  don't  like  to 
discuss.  This  subsidy  is  often  in  the  form  of 
receipts  from  produce. 


ORGANIZATIONAL  FEATURES 

One  organization  of  Experiment  Stations  is  the 
substation  approach  in  which  the  scientists  are 
located  at  the  main  station  and  support  staff  are 
located  at  substations  in  different  chains  of 
command.  In  order  for  research  to  result  from 
this  organization,  cooperation  (association  for 
common  benefit)  must  occur.  This  arrangement 
is  sometimes  ineffective  because  the  goals  of  the 
two  parties  are  different,  although  never  stated 
to  be  different.  That  is,  the  goal  of  the 
scientist  is  to  perform  fundamental  research, 
publish,  attain  tenure  and  advance  in  grade, 
whereas,  the  goals  of  the  support  personnel  often 


1/Resident  Director  of  Research,  Texas  Agricul- 
tural Experiment  Station,  Uvalde  Center. 


include  performance  of  production  research,  sale 
of  produce,  payment  of  bills  and  continuance  of 
existence.  Some  considerations  in  performing 
cooperative  research  are:  (1)  remember  the  chain 
of  command,  (2)  involve  the  superintendent  and 
support  staff  at  all  levels  of  research  (project 
development  and  execution),  and  (3)  communicate 
frequently  (confront  emerging  issues). 

Another  organization  selected  by  some  states  is 
called  the  center  approach  and  involves 
scientists  located  both  at  the  main  station  and 
at  outlying  stations.  These  scientists  are 
similar  in  some  respects  but  have  different  per- 
spectives. At  the  main  station,  scientists  are 
closer  to  the  industry  and  are  preoccupied  with 
the  clientele.  In  order  for  the  goals  of  re- 
search to  be  accomplished,  these  scientists  must 
collaborate  (joint  intellectual  effort).  The 
needs  required  by  each  type  of  scientist  from  the 
collaborative  effort  are  different.  Scientists 
at  the  main  station  need  a site  for  research,  on- 
site support  and  to  gain  a feel  for  area  agri- 
culture and  problems  whereas,  scientists  at 
outlying  stations  need  rapport,  discipline  con- 
tact and  graduate  student  interaction  from  the 
collaborative  effort.  Some  considerations  to 
keep  in  mind  in  collaboration  are: 

1.  In  project  development, 

a.  Develop  your  objectives  so  that  success- 
ful execution  is  not  dependent  upon 
activities  of  your  colleague 

b.  Define  the  specific  objectives  you  desire 
from  the  collaboration 

c.  Determine  areas  of  overlap 

d.  Delineate  responsibilities 

e.  Determine  titles,  dates  and  authors  of 
manuscripts 

f.  Begin  small  to  develop  trust 

2.  In  project  execution, 

a.  Maintain  enthusiasm 

b.  Communicate  often 

c.  Remember  your  collaborators 

d.  Publish  (follow  through) 


CONCLUSION 

Because  of  logistical  and  communication  problems, 
cooperative/collaborative  research  is  difficult 
but  necessary  to  solve  the  goals  of  research  at 
the  resolution  necessary  to  the  industry.  To 
pursue  these  goals  (to  develop  unifying  theorems 
and  broadly  applicable  technology),  research  at 
each  site  must  fit  into  a statewide,  regional  or 
national  matrix  and,  therefore,  both  coopera- 
tion and  collaboration  are  required. 


6 


GRAZING  EXPERIMENTS  - OBJECTIVES,  COMPROMISES 
AND  APPROPRIATE  DESIGNS 

OBJECTIVES  OF  GRAZING  RESEARCH  AND  THEIR 
IMPLICATIONS  FOR  EXPERIMENTAL  DESIGN 

A.  G.  Matches  — ^ 

A complete  treatise  on  this  topic  is  not 
possible  within  the  time  frame  available.  In 
fact,  a group  has  met  the  past  two  days 
discussing  this  topic  in  some  detail  in 
preparation  for  a symposium  at  a later  date. 

I will  briefly  touch  upon  thoughts  that  I have 
on  these  topics,  information  gleaned  from  the 
literature,  and  conclude  with  a quotation  from 
Gill  which  I believe  is  very  appropriate  for 
the  discussion  of  Dr.  Bransby  which  follows. 


WHY  GRAZING  EXPERIMENTS? 

More  than  one  research  administrator  and 
experiment  station  director  has,  over  the 
years,  questioned  whether  grazing  experiments 
are  really  necessary.  From  us  they  have 
become  aware  of  several  "truisms": 

1.  High  costs  per  treatment  as  compared 
to  other  field  experiments. 

a.  Sizeable  land  requirements 

b.  Large  investment  in  livestock 

c.  Fencing  and  watering  facilities 

d.  Maintenance  cost 

- physical  facilities 

- animal  health 

- fertilizer 

- weed  control  in  experimental 
area 

- irrigation  (where  used) 

e.  Duration  of  experiments  usually  3 
to  6 years 

2.  Quite  labor  intensive  if  conducted 
properly. 

a.  Animal  management 

b.  Animal  health 

c.  Monitoring  to  characterize  the 
sward  over  time 

3.  Involve  complex  relationships  which 
are  not  easily  delineated. 

a.  Plant/animal 

b.  Soi 1 /plant/animal 

c.  Management/. .. .etc. 

4.  High  experimental  variability  not 
uncommon. 

a.  Two  biological  systems  imposed  on 
each  other 

b.  Less  statistical  sensitivity  than 
in  many  other  kinds  of  research 

\_/  Department  of  Plant  and  Soil  Science,  Texas 
Tech  University,  Lubbock,  TX  79409 


One  might  conclude  from  the  above  that  many  of 
us  have  selected  careers  which  are  constantly 
in  jeopardy  in  one  way  or  another. 

PURPOSE  OF  GRAZING  EXPERIMENTS: 

Short-term  and  long-term  goals  may  be  achieved 
from  conducting  good  grazing  trials.  In  the 
short  term  (mainly  a number  of  separate 
experiments),  grazing  research  can: 

1.  Depict  the  animal  influence  on  sward 
and  soi 1 . 

a.  Herbage  dry  matter  production 

b.  Influences  on  plant  morphology 

c.  Regrowth  potential  under  grazing 

d.  Changes  in  botanical  composition 

e.  Persistence 

f.  Soil  compaction 

g.  Soil  conservation 

h.  Nutrient  transfer 

i.  etc. 

2.  Depict  the  sward  influence  on 
animals . 

a.  Output  per  animal  as  influenced 

by: 

- herbage  availability 
(fertilization,  species, 
stocking  rates,  etc.) 

- canopy  structure 

- forage  qual ity 

- plant  morphology 

- anti-quality  factors 

- etc. 

b.  Animal  health 

- anti-quality  factors 

- morphological  factors 

3.  Identify  pasture/management 
components  for  pasture/livestock 
systems . 

In  the  long  term,  grazing  trials  provide  the 
basis  for  designing  and  testing  pasture 
systems  for: 

1.  Specific  classes  of  livestock. 

2.  Specific  environmental  constraints. 

3.  Specific  types  of  markets. 

4.  Different  economic  scenarios. 

Certainly  expenditures  on  grazing  research  are 
justified.  However,  Morley  (1978)  has 
emphasized  the  need  for  a systems  approach  in 
grassland  evaluation.  He  suggests  that  the 
greatest  progress  will  arise  from  the 
combination  of  mathematical  modeling  (systems 
analysis)  and  field  experimentation.  Based  on 
my  review  of  over  70  IJ.S.  grazing  research 
papers  that  have  been  published  in  the  past  20 
years,  it  is  apparent  that  we  should  give  more 
attention  to  describing  functions  related  to 
the  plant/animal  interface.  Such  information 
in  the  hand  of  modelers  will  ultimately  pave 
the  way  for  us  to  plan  even  more  comprehensive 
grazing  trials  in  the  future. 


7 


DESIGNING  THE  EXPERIMENT: 

Designing  research  should  include  the 
following  three  important  steps  (Cook  and 
Stubhendieck , 1986): 

1.  Clearly  define  and  prioritize  the 
objectives  of  the  research  and  give 
the  hypothesis.  The  null  hypothesis 
(no  differences  among  treatments)  is 
most  commonly  used,  but  in  some  cases 
the  alternative  hypothesis  may  be 
more  appropriate. 

2.  Describe  the  experimental  material, 
treatments  to  be  investigated,  and 
conditions  under  which  treatments 
will  be  compared.  These  all 
influence  the  selection  of  the  most 
appropriate  experimental  design. 

3.  Describe  the  measurements  to  be 
recorded,  the  precision  desired,  and 
the  type  of  conclusions  to  be  drawn 
(how  are  results  to  be  applied). 

Too  often,  pasture  plantings  are  made  before 
the  researcher  has  given  adequate  thought  to 
the  above  points.  Consequently,  the 
experimental  design  selected  when  planting  may 
not  be  entirely  appropriate  to  meet  the 
objectives  and  intended  application  of 
results.  In  planning  experiments,  I find  that 
keying  out  the  analysis  of  variance  for 
different  designs  and  combinations  of 
treatments  is  very  helpful  for  selecting  the 
final  design  and  make-up  of  the  experiment. 

In  some  cases  following  this  exercise,  I have 
decided  that  with  the  resources  available,  the 
intended  research  could  not  be  effectively 
conducted . 


EXPERIMENTAL  DESIGN:  COMMENTS  AND 
CONSIDERATIONS: 

My  comments  are  limited  to  grazing  experiments 
where  both  animal  production  and  plant 
responses  to  grazing  (objectives  1 and  2 
above)  are  investigated  concurrently  within 
the  same  experimental  pastures. 

Two  types  of  trials  that  are  in  common  use  are 
continuous  trials  and  change-over  trials 
(Lucas,  1959).  In  continuous  trials,  animals 
remain  on  the  same  treatment  for  the  duration 
of  the  experiment.  In  change-over  trials, 
animals  are  subjected  to  at  least  two  or  more 
experimental  treatments  during  the  course  of 
investigation.  Most  grazing  trials  are  of  the 
continuous  type  and  utilize  a randomized 
complete  block  (RCB)  or  a split-plot  (SP) 
design.  Because  so  few  treatments  are  usually 
investigated  in  grazing  trials  and  because  of 
certain  statistical  limitations,  other 
experimental  designs  such  as  the  latin  square 
and  lattice  are  used  less  frequently. 

The  impact,  of  experimental  design  on  the 
application  of  grazing  results  is  covered  by 


Brown  and  Waller  (1986),  especially  in  respect 
to  replicated  vs.  unreplicated  experiments. 
Walker  and  Richardson  (1986)  also  discuss  the 
matter  of  replications  in  grazing  studies. 
Replications  are  necessary  in  order  to  perform 
tests  of  significance  of  treatment  effects 
related  to  land  area  in  the  analysis  of 
variance.  Brown  and  Waller  state  that 
"experimental  design  of  comparative  range  and 
pasture  grazing  trials  should  include 
sufficient  replication  of  land,  animals,  and 
time  to  properly  estimate  variance  at  an 
acceptable  level  of  precision  for 
characterization  or  inference."  In  their 
opinion,  unreplicated  pasture  studies  can 
serve  as  a screening  trial  for  several 
treatments  which  may  be  included  in  replicated 
trials  which  follow.  In  the  following  paper, 
regression  techniques  which  do  not  require 
field  replications  are  discussed  by  Dr. 

Bransby . 

As  stated  by  Walker  and  Richardson  (1986)  from 
a reference  source,  "an  experimental  unit  or 
experimental  plot  is  the  unit  of  material  to 
which  one  application  of  a treatment  is 
applied."  In  grazing  experiments  where  animal 
performance  is  measured,  the  experimental  unit 
is  the  pasture  (Brown  and  Waller,  1986;  Cook 
and  Stubhendieck,  1986;  Morley,  1978). 
Therefore,  the  animal -to-ani mal  source  of 
variation  may  not  be  considered  by  some  as  an 
appropriate  error  component  for  testing 
resulting  treatment  differences  in  animal 
performance. 

Where  rotational  grazing  is  followed  in 
grazing  experiments,  some  researchers  move 
animals  among  replications  to  achieve 
rotational  grazing.  But  as  explained  by  Mott 
(1959),  weight  increase  for  an  animal  over 
each  weigh  period  (W  ) is  represented  by  the 
following  equation  for  four  weighings: 

Weight  increase=[W9 (e)  - W, (e)]  + 

[Wo (e ) - w;(e)j  + 

[Wj(e)  - W‘(e)] 

where  "e"  is  the  error  associated  with  each 
weighing.  Therefore,  in  rotating  animals 
among  repl ications,  the  errors  of  each 
weighing  are  accumulated  and  this  inflates  the 
error  term  in  the  analysis  of  variance.  Mott 
recommended  that  each  pasture  be  subdivided 
into  paddocks  and  animals  rotated  within  each 
pasture.  Then,  over  the  four  weigh  periods, 
all  but  the  weighing  errors  on  the  first  and 
last  weighing  would  cancel  out  as  follows: 

Weight  increase  = [W^(e)  - W^(e)] 

Matches  (1969)  showed  a hypothetical  situation 
where  rotating  animals  among  replications 
might  result  in  a true  interaction  of 
"replication  x treatments".  This  would 
further  inflate  the  experimental  error  in  a 
simple  RCB  design  experiment  since  the  RxT 
interaction  is  normally  the  error  term  for 
testing  treatment  differences 


8 


Change-over  trials  are  of  two  types, 
rotational  and  switch-back  or  reversal.  When 
used  properly,  change-over  trials  may  reduce 
experimental  errors  associated  with  the 
variability  among  animals  (Lucas  1960). 
Normally,  change-over  trials  are  used  only 
where  treatment  effects  do  not  have  a strong 
carry-over  effect  on  the  animal  (Gill,  1981; 
Lucas,  1963).  Because  pasture  treatments 
usually  have  large  carry-over  effects  on 
animals,  change-over  trials  are  not  generally 
recommended  for  use  in  grazing  trials. 


COMPONENTS  VS.  PASTURE  SYSTEMS: 

Pasture  systems  consisting  of  separate 
pastures  of  different  forages  or  management 
schemes  are  useful  for  extending  grazing  and 
providing  higher  quality  herbage  throughout 
the  grazing  season  (Matches,  et  a 1 . , 1974; 
Matches,  1981).  Components  of  such  systems 
are  generally  first  evaluated  "individually" 
in  conventional  grazing  trials.  Sometimes, 
researchers  are  tempted  to  use  component  data 
from  several  experiments  to  project  what 
animal  performance  would  be  if  different 
components  were  grouped  to  form  pasture 
systems.  Projected  animal  performance  (daily 
gain/head  and  gain/ha)  likely  will  be 
inaccurate.  For  example,  Matches,  et  al. 
(1974)  and  Matches  (1981)  found  that  eight 
forage  components  had  a spread  of  459  g for 
daily  gain  and  295  kg  for  gain/ha.  In 
comparison,  under  grazing  the  spread  among  ten 
pasture  systems  comprised  of  the  various 
components  was  only  204  g for  daily  gain  and 
78  kg  for  gain/ha.  Similar  results  were 
reported  from  other  experiments.  Apparently, 
compensatory  gain  responses  (positive  and 
negative)  of  cattle  resulted  in  a leveling-out 
of  gains  in  the  systems.  Therefore, 
projections  of  cattle  performance  should  not 
be  made  based  on  the  combined  results  from 
single  component  grazings. 

Related  to  the  above  is  the  question  of 
whether  animals  should  be  rerandomized  when 
moving  from  one  component  to  another  in  a 
pasture  systems  trial.  This  has  been  done  in 
some  pasture  system  experiments.  Because  the 
experimental  unit  is  a pasture  system 
comprised  of  two  or  more  components,  and 
because  the  system  is  the  variable  of  concern 
when  applying  the  results,  I would  not 
rerandomize  animals  when  animals  are  shifted 
among  components.  Compensatory  gain  effects 
of  animals  would  be  masked  so  that  the  system 
potential  could  not  be  accurately  measured. 


DEALING  WITH  DIFFERENCES  IN  GROWTH  PATTERN 
AMONG  FORAGES: 

With  cultivar  evaluations  and  pasture 
component  research,  it  is  not  unusual  to  have 
pastures  of  different  forages  which  are  not 
all  ready  for  grazing  at  the  same  time  because 
of  inherent  differences  in  time  and  rapidity 
of  initial  growth.  Should  grazing  be  delayed 


until  "all"  entries  have  sufficient  growth  for 
grazing?  Or  should  grazing  begin  on  "each" 
entry  when  its  growth  is  ready?  In  my 
opinion,  researchers  and  producers  should 
"read  the  plants  and  not  the  calendar." 
Delaying  grazing  of  early  growing  entries 
means  lost,  quality  and  lost  animal  gain. 
Certainly,  wise  producers  will  graze  each  of 
their  forages  when  it  is  ready;  consequently, 
if  research  is  to  meet  the  producers'  needs, 
grazing  should  be  initiated  for  each  entry 
when  it  reaches  its  proper  stage  of  growth. 


FUTURE  NEEDS: 

Funding  limitations,  land  avai labi 1 ity , and 
other  factors  often  restrict  the  scope  of  many 
grazing  trials.  Replicating  pastures  is 
expensive  and  minimizes  the  number  of 
treatments  that  can  be  investigated.  We  need 
the  help  of  statisticians  who  can  devise 
experimental  designs  which  will  give  us 
reasonable  precision  and  still  allow  the 
testing  of  more  treatments  per  trial.  Gill 
(1981)  has  suggested  for  feeding  experiments 
that  wider  use  of  regression  techniques, 
multivariate  procedures,  and  response  surface 
designs  might  offer  opportunities  as  more 
flexible  computer  programs  become  available  in 
the  future.  But  he  also  cautions  that 
"perhaps  the  greatest  source  of  inertia 
holding  back  the  wheels  of  progress  in 
statistical  applications  is  the  occasional 
professor  who  insists  that  his  research 
students  ignore  their  modern  instruction  in 
statistical  methodology  and  do  it  his  way." 
Hopefully,  new  statistical  procedures  and  open 
minds  will  give  us  the  tools  to  meet  the 
research  challenges  of  the  future. 


Literature  Cited 

Brown,  M.A.  and  S.S.  Waller.  1986.  The 
impact  of  experimental  design  on  the 
application  of  grazing  research  - an 
exposition.  J.  Range  Manage.  39:197-200. 

Cook,  C.  Wayne  and  James  Stubbendieck  (ed.) 
1986.  Chapter  10,  Experimental  design,  jjn 
Range  research:  basic  problems  and 
techniques.  Soc.  Range  Manage,  p.  251-252. 

Gill,  J.L.  1981.  Evaluation  of  statistical 
design  and  analysis  of  experiments.  J.  Dairy 
Sci.  64:1494-1519. 

Lucas,  H.L.  1959.  Experimental  designs  and 
analyses  for  feeding  efficiency  trials  with 
dairy  cattle.  J_n  C.R.  Hoglund  (ed.) 
Nutritional  and  economic  aspects  of  feed 
utilization  by  dairy  cows.  Iowa  State  College 
Press,  p.  177-192. 

Lucas,  H.L.  1960.  Critical  features  of  good 
dairy  feeding  experiments.  J.  Dairy  Sci. 
43:193-212. 


9 


Lucas,  H.L.  1963.  Special  considerations  in 
the  design  of  grazing  experiments.  In  Range 
research  methods.  IJ.S.D.A.  Misc.  Pub.  No.  940, 
p.  132-137. 

Matches,  Arthur  G.  1969.  Pasture  research 
methods.  J_n  R.F.  Barnes,  D.C.  Clanton,  G.H. 
Gordon,  T.J.  K 1 opfenstei n , and  D.R.  Waldo 
(ed.)  Proc.  nat.  conf.  on  forage  quality 
evaluation  and  utilization.  Neb.  Ctr.  for 
Continuing  Education.  Lincoln,  NE.  11-33. 

Matches,  A.G.,  F.A.  Martz,  and  G.B.  Thompson. 
1974.  Multiple  assignment  tester  animals  for 
pasture-animal  systems.  Agron.  J.  66:719-722. 

Matches,  A.G.  1981.  Theoretical  construction 
of  grazing  systems  from  knowledge  of  component 
humid  pastures.  J_n  J.L.  Wheeler  and  R.D. 
Mochrie  (ed.)  Forage  evaluation:  concepts  and 
techniques.  American  Forage  and  Grassland 
Council  and  CSIRO.  Melbourne,  p.  473-481. 

Morley,  F.H.W.  1978.  Animal  production 
studies  on  grassland.  J_n  L.  't  Mannetje  (ed.) 
Measurement  of  grassland  vegetation  and  animal 
production.  Commonwealth  Agriculture  Bur. 
Pasture  and  Field  Crops.  Hurley,  Berkshire, 
England.  Bull.  52.  p.  103-162. 

Mott,  G.O.  1959.  Intersociety  forage 
evaluation  symposium:  IV.  Animal  variation  and 
measurement  of  forage  quality.  Agron.  J. 
51:223-226. 

Walker,  John  W.  and  Edgar  W.  Richardson. 

1986.  Replication  in  grazing  studies  - why 
bother?  J_n  Proc.  symposium  on  statistical 
analysis  and  modeling  of  grazing  systems. 
Convened  and  compiled  by:  Charles  D.  Bonham, 
Sandra  S.  Coleman,  Clifford  E.  Lewis,  and 
George  W.  Tanner.  39th  Ann.  Meeting  Soc. 

Range  Manage.  Feb.  13,  Kissimmee,  FL.  p.  51- 
58. 


10 


ADVANTAGES  OF  A NON-REPLICATED  MULTIPLE 
GRAZING  INTENSITY  APPROACH  FOR  GRAZING 
TRIALS 

David  I.  Bransby* 

Introduction 

The  procedures  and  design  used  in  any 
grazing  trial  will  depend  largely  on  the 
objectives  of  that  trial.  However, 
financial  and  logistical  constraints  such 
as  limited  labor,  land,  paddocks,  animals 
etc.  are  almost  always  severely 
restricting.  No  matter  what  the 
objectives  of  a grazing  trial  are, 
therefore,  compromise  is  usually 
necessary.  This  compromise  requires 
determination  of  an  acceptable  balance 
between  biological  or  practical  value  and 
applicability  of  results  on  the  one  hand, 
and  on  the  other,  scientific  value.  In 
other  words,  grazing  research  data  should 
ideally  be  applicable  to  the  producer 
and/or  reveal  new  biological  information, 
and  be  scientifically  testable  or 
assigned  a level  of  probability. 

It  might  be  argued  that  component 
research  (as  opposed  to  systems  research) 
need  not  necessarily  be  practically 
applicable.  This  is  true,  but  if 
experimental  designs  and  procedures  that 
have  practical,  biological  and  scientific 
merit  can  be  developed,  this  will  make 
more  efficient  use  of  limited  available 
resources  for  grazing  research. 
Futhermore,  in  animal  production  systems 
such  as  beef  enterprises,  a component  of 
a system  can  represent  a simple  system  in 
itself.  For  example,  many  producers  in 
the  south  buy  calves,  grow  them  out  on 
pastures,  and  sell  them,  and  this  simple 
system  is  simulated  by  a large  proportion 
of  beef  grazing  trials.  It  is  therefore 
desirable  that  data  from  such  trials  be 
well  suited  to  economic  analysis  (2). 

The  objective  of  this  paper  is  to  discuss 
the  biological,  practical  and  scientific 
merit  of  traditional  procedures,  and  an 
alternative  multiple  grazing  intensity 
procedure  for  grazing  trials. 

Traditional  procedures 

The  traditional  approach  to  grazing 
research  in  the  U.S.  has  included  (a)  the 
put-and-take  method  of  stocking,  (b) 
usually  only  one  grazing  intensity  (one 
grazing  pressure,  forage  availability  or 
level  of  forage  on  offer)  for  a whole 
experiment  (occasionally  grazing 


IDepartment  of  Agronomy  and  Soils,  Auburn 
University,  AL  36849. 


intensity  has  been  confounded  with 
treatment),  and  (c)  a randomized  complete 
blocks  design  with  two  or  three 
repl ications . The  objectives  of  the 
put-and-take  method  of  stocking  have  been 
(1)  to  maintain  grazing  intensity 
constant  across  replications  and 
treatments  over  time,  (2)  to  ensure  that 
animal  potential  is  above  pasture 
potential  at  all  times,  and  (3)  to 
facilitate  use  of  average  daily  gain 
(ADG)  as  an  animal  measure  of  forage 
quality,  and  animal  grazing  days  per  unit 
area  (or  average  stocking  rate)  as  a 
measure  of  forage  quantity  (15).  In 
other  words,  the  last  objective 
represents  an  attempt  to  avoid 
confounding  of  forage  quantity  and 
quality  as  it  affects  animal  performance. 
This  concept  clearly  has  considerable 
biological  and  scientific  merit. 

Strengths  and  weaknesses  of  the 
traditional  approach 

The  main  strengths  of  the  traditional 
approach  are  the  "pure"  error  term 
provided  by  replication,  and  the 
flexibility  afforded  by  put-and-take 
stocking  to  adjust  for  the  unknown,  such 
as  unpredictable  weather,  no  previous 
information  on  new  treatments,  and  new 
environments.  On  the  other  hand,  several 
weaknesses  are  apparent,  (a)  The  optimum 
grazing  intensity  is  seldom  (if  ever) 
defined,  and  cannot  be  identified  without 
mutliple  grazing  intensity  (or  stocking 
rate)  research,  (b)  Treatment  x grazing 
intensity  interactions  cannot  be  detected 
if  only  one  grazing  intensity  is  used. 
Since  these  interactions  are  probably 
common  in  grazing  systems,  results  from 
single  grazing  intensity  studies  may 
apply  only  to  the  grazing  intensity  used, 
(c)  Grazing  intensity  (as  determined  by 
kg  available  forage  per  unit  area)  cannot 
be  perfectly  measured,  or  perfectly 
replicated  by  means  of  the  put-and-take 
method  of  stocking.  In  some  cases, 
control  of  forage  availability  may  be 
poor  (16).  This  means  that  forage 
availability  is  confounded  with 
repl ications , treatments,  years  etc.,  in 
which  case  ADG  cannot  be  used  as  a 
measure  of  forage  quality,  and  animal 
grazing  days  per  unit  area  is  not  a 
reliable  measure  of  forage  quantity,  (d) 
Put-and-take  of  animals  is  not  common 
farm  practice  (put-and-take  of  land 
and/or  time  might  be  more  applicable 
here)  and  is  not  well  suited  to  economic 
analysis  (12).  (e)  The  basis  for 
put-and-take  of  animals  appears  to  have 
been  highly  variable  among  workers  (eg. 
kg  forage/animal,  kg  forage/animal /day , 
kg  forage/ha,  forage  height,  residual 


11 


forage  etc);  therefore,  results  may  not 
be  comparable  across  workers. 

The  traditional  approach  could  be 
improved  in  several  ways.  Firstly,  if 
only  one  grazing  intensity  is  used  in  an 
experiment,  forage  availability  should  be 
measured  and  analysed  to  show  the  degree 
to  which  it  was  controlled  by 
put-and-take . It  should  also  be 
expressed  in  a form  which  reflects  ease 
of  prehension  by  the  grazing  animal,  such 
as  height  or  density  (6,7).  Weight  of 
forage  per  animal  unit  may  not  adequately 
reflect  this.  For  example  if  a 2-ha  and 
a 10-ha  pasture  each  contained  2000  kg  of 
forage  and  10  animals,  kg  of 
forage/animal  is  the  same,  but  intake  and 
animal  performance  are  likely  to  be  very 
different.  Secondly,  if  forage 
availability  was  not  successfully 
equalized  across  treatments  and 
replications  by  put-and-take  in  a single 
grazing  intensity  trial,  it  may  help  to 
use  forage  availability  as  a co-variate 
in  an  analysis  of  covariance.  Finally, 
from  a practical  and  economic  viewpoint, 
it  would  probably  be  better  to 
put-and-take  land  and/or  time  instead  of 
individual  animals,  and  to  make  these 
adjustments  relatively  infrequently. 

The  need  for  multiple  grazing  intensity 
trials 

It  is  of  great  value  in  a grazing  trial 
to  apply  several  treatments  at  several 
(preferably  at  least  four)  grazing 
intensities,  which  may  constitute 
different  levels  of  forage  availability 
achieved  with  put-and-take,  or  different 
fixed  stocking  rates  (1,2,5,8,10,13,17). 
Multiple  grazing  intensity  trials  are 
needed  because;  (1)  treatment  x grazing 
intensity  (stocking  rate)  interactions 
can  be  detected;  (2)  optimum  grazing 
intensity  varies  among  treatments;  (3) 
economic  optimum  grazing  intensity  varies 
with  buying  and  selling  price  of  animals 
(3,4,11,14,18);  (4)  it  is  not  necessary 
to  perfectly  repeat  a level  of  forage 
availability  to  use  ADG  as  an  index  of 
forage  quality,  since  regression  lines 
are  used  in  the  analysis,  and  not  single 
points;  and  (5)  it  is  very  important  to 
establish  the  heaviest  grazing  intensity 
a pasture  can  tolerate  without  losing 
stand,  since  stand  longevity  is  probably 
the  trait  most  valued  by  producers  in  a 
perennial  species,  and  under  production 
conditions  there  are  inevitably  going  to 
be  times  when  heavy  grazing  is 
unavoidable.  Financial  and  logistical 
constraints  will  clearly  make  it 
difficult  to  apply  several  treatments  at 
3 or  4 stocking  rates  (preferably  4)  and 


2 or  3 replications.  Flowever,  a 
regression  approach  (which  is  a well 
recognized  statistical  procedure) 
facilitates  statistical  analysis  without 
replication  (9,17,19). 

Analysis  and  interpretation  of  data  from 
non-repl icated , multiple  grazing 
intensity  trials 

To  analyse  and  interpret  data  from 
non-repl icated,  multiple  grazing 
intensity  trials  it  is  necessary  to 
examine  three  regression  relationships; 
ADG  vs.  stocking  rate,  ADG  vs.  forage 
availability  and  forage  availability  vs. 
stocking  rate  (5).  These  regressions  can 
be  developed  from  several  treatments  and 
compared  by  testing  for  differences  among 
slopes  and  intercepts  of  the  lines.  In 
the  broadest  sense,  testing  for 
statistical  differences  among  treatments 
in  a scientific  experiment  involves 
comparing  variation  which  can  be 
accounted  for  (by  treatments,  blocks, 
regression  coefficients,  etc.)  with 
variation  that  cannot  be  explained  (the 
error  term):  if  the  former  is  large 
relative  to  the  latter,  then  statistical 
significance  is  indicated.  In  the  case 
of  a non-repl icated , multiple  grazing 
intensity  design,  deviation  from 
regression  is  the  only  variation  that 
cannot  be  explained,  since  all  other 
variation  is  accounted  for  by  treatments 
and  regression  coefficients.  Viewed  in 
another  way,  each  treatment  can  be 
considered  as  replicated,  but  replicates 
are  confounded  with  grazing  intensity  or 
stocking  rate.  The  data  could  then  be 
analysed  by  analysis  of  covariance,  with 
grazing  intensity  or  stocking  rate  as  the 
covariate.  Differences  between  stocking 
rates  cannot  be  tested,  but  this  is  of 
little  consequence  provided  other 
experimental  variables  are  not  confounded 
with  stocking  rate. 

Differences  among  treatments  in  the  ADG 
vs.  stocking  rate  regression  (1,8,13,17) 
are  of  little  value  if  they  are  not 
linked  to  forage  availability  because 
separate  effects  of  forage  quality  and 
quantity  on  production  per  animal  cannot 
be  determined  (Fig.  1).  However,  this 
relationship  forms  the  basis  for  economic 
analysis  (3,4,11,14,18).  Differences 
among  treatments  in  the  ADG  vs.  forage 
availability  regressions  represent 
differences  in  forage  quality  (Fig.  2), 
while  differences  in  the  forage 
availability  vs.  stocking  rate  regression 
represent  differences  in  yield  or 
"carrying  capacity"  at  a particular 
forage  availability  (Fig.  3).  These 
three  relationships  provide  a basis  for 


12 


ADG . 


relating  gain/ha  and  profit  to  stocking 
rate  and  forage  availability. 
Optimization  procedures  can  then  be  used 
to  determine  the  level  of  forage 
availability  or  stocking  rate  that 
maximizes  gain/ha  or  profit 
(3,4,11,14,18). 

Furthermore,  inherent  differences  among 
paddocks  are  likely  to  be  expressed 
mainly  in  terms  of  forage  availability. 
Consequently,  the  ADG  vs.  forage 
availability  regression  (Fig.  2)  will 
largely  remove  this  variation  which  will 
appear  as  deviations  from  regression  in 
the  ADG  vs.  stocking  rate  and  available 
forage  vs.  stocking  rate  regressions 
(Fig.  1 and  Fig.  3).  Consequently, 
measurements  of  forage  availability  are 
critical  in  analysis  and  interpretation 
of  data  from  non-replicated,  multiple 
stocking  rate  grazing  trials. 

Conclusion 

The  non-replicated,  multiple  grazing 
intensity  approach  represents  an 
extremely  attractive  compromise  between 
scientific,  practical  and  biological 
needs  in  grazing  trials. 


Relationships  between  ADG  and  stocking  rate  for 
hypothetical  treatments  A and  B,  showing  a typical 
treatment  x stocking  rate  interaction  (different 
slopes) . 


Fig.  2. 

Relationships  between  ADG  and  kg  forage/ha  for  hypothetical 
treatments  A and  B,  showing  a typical  main  effect  (parallel 
lines)  which  can  be  ascribed  to  differences  in  forage  quality 
between  A and  B. 


Fig.  3. 

Relationships  between  kg  forage/ha  and  stocking  rate  for 
hypothetical  treatments  A and  B showing  a typical  main 
effect  (parallel  lines).  This  relationship  indicates 
quantitative  effects:  the  difference  in  stocking  rate  at 
a given  level  of  forage/ha,  or  the  difference  in  forage/ha 
at  a given  stocking  rate. 

References 

1.  Bransby,  D.I.  1984.  A model  for 
predicting  livemass  gain  from  stocking 
rate  and  annual  rainfall.  J.  Grassl . 

Soc.  Sth.  Afr.  1(2):  22-26. 

2.  Bransby,  D.I.  1985.  Modelling 
grazing  intensity  studies.  Proc.  15th 
Int.  Grassl.  Cong.  Kyoto,  Japan: 

1092-1093. 

3.  Bransby,  D.I.  1985.  A model  for 
predicting  long  term  economic  optimum 
stocking  rates  for  beef  cattle  grazing 
dryland  pastures.  J.  Grassl.  Soc.  Sth. 
Afr.  2:  18-20. 


13 


4.  Bransby,  D.I.,  and  B.E.  Conrad. 

1985.  Relating  profit  to  quantity  of 
standing  herbage  in  grazing  intensity 
studies.  Proc.  15th  Int.  Grassl . Cong., 
Kyoto,  Japan:  1151-1152. 

5.  Bransby,  D.I.,  B.E.  Conrad,  H.M. 

Dicks  and  J.W.  Drane.  1988. 

Justification  for  grazing  intensity 
experiments:  analysing  and  interpreting 
grazing  data.  J.  Range  Man.:  in  press. 

6.  Bransby,  D.I.,  A.G.  Matches  and  G.F. 
Krause.  1977.  Disk  meter  for  rapid 
estimation  of  herbage  yield  in  grazing 
trials.  Agron.  J.  69:  393-396. 

7.  Bransby,  D.I.,  and  N.M.  Tainton. 

1977.  The  disc  meter:  possible 
applications  in  grazing  management.  Proc. 
Grassl.  Soc.  Sth.  Afr.  12:  115-118. 

8.  Cowlishaw,  S.J.  1969.  The  carrying 
capacity  of  pastures.  J.  Brit.  Grassl. 
Soc.  24:  207-214. 

9.  Draper,  M.R.  and  H.  Smith.  1966. 
Applied  regression  analysis.  John  Wiley 
and  Sons,  Inc.  New  York. 

10.  Hart,  R.H.  1972.  Forage  yield, 
stocking  rate  and  beef  gains  on  pasture. 
Herb.  Abstr.  42:  345-353. 

11.  Hildreth,  R.J.  and  M.E.  Riewe.  1963. 
Grazing  production  curves  II. 

Determining  the  economic  optimum  stocking 
rate.  Agron.  J.  55:  370-372. 

12.  Jacobs,  V.E.  1974.  Needed:  A 
systems  outlook  in  forage-animal 
research,  p 33-48.  I_n  R.W.  van  Keuren 
(ed.)  Systems  analysis  in  forage  crops 
production  and  utilization.  Crop  Science 
Society  of  America,  Madison,  Wisconsin. 


13.  Jones,  R.J.  and  R.L.  Sandland.  1974. 
The  relation  between  animal  gain  and 
stocking  rate.  J.  Agric.  Science  Cainb. 
83:  606-611. 

14.  McCartor,  M.M.  and  F.M.  Roquette. 
1977.  Grazing  pressure  and  animal 
performance  from  pearl  millet.  Agron.  J. 
69:  983-987. 

15.  Mott,  G.O.  1960.  Grazing  pressure 
and  the  measurement  of  pasture 
production.  Proc.  8th  Int.  Grassl. 
Congr.,  Reading,  England:  606-611. 

16.  Read,  J.C.  and  B.J.  Camp.  1986.  The 
effect  of  the  fungal  endophyte, 

Acremonium  coenophialum  in  tall  fescue  on 
animal  performance,  toxicity  and  stand 
maintenance.  Agron.  J.  78:  848-850. 

17.  Riewe,  M.E.  1961.  Use  of  the 
relationship  of  stocking  rate  to  gain  of 
cattle  in  an  experimental  design  for 
grazing  trials.  Agron.  J.  53:  309-313. 

18.  Riewe,  M.E.  1981.  The  economics  of 
grazing.  In  J.L.  Wheeler  and  R.D.  Mochrie 
(ed.).  Forage  evaluation:  Concepts  and 
techniques.  AFGC,  USA  and  CSIR0, 

Austral ia. 

19.  Snedecor,  G.W.  and  W.G.  Cochran. 

1967.  Statistical  methods.  Iowa  State 
University  Press,  Ames,  Iowa. 


STATISTICAL  ASPECTS  OF  GRAZING  EXPERIMENT 
J . Wanzer  Drane1 


SUMMARY 

Research  designs  available  for  the  study  of 
forage  crop  improvement  and  grazing  trials 
are  many  and  varied.  The  ab i I i ty  to  detect 
treatment  differences  depends  almost  en- 
tirely on  the  foraging  meat-producing  model 
and  the  research  design  actual  ly  used. 

Small  changes  in  the  designs  can  increase 
or  decrease  sinsitivity  of  statistical 
tests  substantial  ly  by  virtue  of  the  mean 
squares  used  as  error  variances.  In  this 
report  examples  of  hypothetical  and  actual 
experiments  are  evaluated  and  compared  from 
the  viewpoint  of  statistical  designs  of 
research  plans. 


should  be  used  and  compared  to  the  upper 
percentage  points  of  an  F with  df  (degrees 
of  freedom)  equal  J-1  and  (I— 1)(J— 1).  But 
this  F cannot  be  calculated  because  there 
is  only  one  observation  per  plot,  and  the 
random  variation  between  plots  cannot  be 
estimated.  What  is  usual  ly  done  (Steel  and 
Torrie,  1980,  p.  195  or  Montgomery,  1984, 
p.  211)  is  to  replace  (2)  with 

F = MST/MSBT  (3) 

by  assuming,  correctly  or  not,  that  the  B 
by  T interaction  is  zero.  Thus,  (3)  is  a 
central  F statistic  if  and  only  if  Tj  and 
BT j j are  both  zero  for  all  combinations  of 
i and  j.  The  hypothesis  of  no  treatment 
effect  is  rejected,  if  the  calculated  var i 
ance  ratio  equals  or  exceeds  the  tabulated 
value  for  a given  error  rate,  say,  0.05  or 
sma  Her. 


I NTRODUCT ION 

In  row  crop  and  grain  experiments,  land  is 
divided  into  plots.  Treatment  variables 
are  applied  to  the  plots,  and  at  the  appr- 
opriate times,  yields  are  measured.  If 
yield  is  measured  as  a single  number  for  a 
plot,  then  random  variability  among  plots 
cannot  be  measured  directly  without  repli- 
cating the  experiment.  Let  us  consider  a 
randomized  block  experiment  wherein  strips 
of  land  cal  led  blocks  are  subdivided  into 
plots  of  nearly  equal  sizes.  In  fact,  they 
are  always  treated,  statistically,  to  be 
exactly  the  same  size.  Following  the  sub- 
division of  the  strips,  treatment  variables 
are  randomly  assigned  to  the  plots  within 
the  b I ocks . 

The  statistical  "effects"  of  the  experiment 
are  assignable  to  Treatments  and  Blocks  and 
their  interactions,  and  in  most  cases,  both 
are  considered  "fixed".  The  linear  additive 
model  and  the  analysis  of  variance  (ANOVA) 
(Table  1)  have  the  following  forms: 

Y j j = M +B  j +T  j +BT  j j +E  j j ( 1 ) 

wherein  M is  the  overall  population  mean; 

B | are  the  block  effects,  and  Tj  are  the 
treatment  effects.  If  the  treatment  effects 
are  not  constant  from  block  to  block,  then 
the  term  BT j j is  added.  Random  vari- 
ation, E | j , is  omnipresent  whether  it  can 
be  measured  or  not. 

Blocks  are  usually  ignored.  To  test  the 
hypothesis  that  there  Is  no  treatment  ef- 
fect or  that  the  treatment  effect  is  con- 
stant across  al I treatments,  the  variance 
ratio 

F = MST/MSE  (2) 

J./  Department  of  Research  Data  Analysis, 

Alabama  Agricultural  Experiment  Station, 
Auburn  University,  AL  36849-5402.  AAES 
Journal  No.  13-881657P. 


If  for  any  reason  it  is  known  or  be  I i eved 
that  the  forgoing  interactions  BT j j are  not 
zero,  then  the  experiment  must  be  designed 
to  include  estimates  of  the  error  variance, 
Ve.  Equation  1 would  be  rewritten  to  re- 
flect the  same,  namely 


Yijk  - M + B i +Tj  +BY  j j +E|<(jj),  (4) 

and  Table  1 would  change  to  Table  2. 


Now,  the  denominator  in  (3)  Is  replaced 
with  the  MSE : 


F(Trtmnt)  = MST/MSE,  and 

F(BxT)  = MSBT/MSE.  (5) 

One  answer  to  the  presence  of  interactions 
in  row  crop  experiments  is  replication 
within  every  treatment,  thereby  allowing 
for  an  estimate  of  the  error  variance.  Even 
though  there  may  be  variabi I i ty  from  plot 
to  plot,  in  general,  there  is  no  great 
concern  expressed  over  it.  This  is  not  true 
in  grazing  experiments. 

When  blocking  is  not  available,  desirable 
or  less  than  artificial,  the  experiment  may 
resolve  to  that  of  one  completely  rando- 
mized over  experimental  units.  Let  us 
consider  one  i I lustrated  completely  in 
Steele  and  Torrie  (1980,  p.  153). 

Treatments  are  appl led  to  the  so i I within 
pots  within  a greenhouse.  Mint  roots 
(plants)  are  planted  in  equal  or  varying 
numbers  in  each  pot.  The  resultant  linear 
additive  model  for  equal  numbers  of  roots 
within  all  pots  i s 

Y = M + T|  + P j ( j ) + R k ( i j ) (6> 

wherein,  M is  the  grand  mean;  Tj  is  treat- 
ment effect,  Pj(j)  is  the  between  Pot  vari- 
ability and  R k ( i j ) * the  between  root  vari- 
ability. Tj  is  externally  imposed  and 
considered  to  be  an  effect  fixed  in  nature. 
Both  P j ( j ) and  Rk ( I j ) • are  random  variables 
and  measure  variation  from  pot  to  pot  and 


15 


from  mint  root  to  root.  The  skeletal  ANOVA 
is  found  in  Table  3. 

Since  the  pot  mean  square  is  the  error  term 
for  testing  one  treatment  against  another, 
the  ability  to  detect  differences  among 
treatments  rests  in  the  number  of  pots 
which  can  be  properly  maintained  within  the 
greenhouse.  However,  the  number  of  plants 
within  a pot  could  be  a factor  of  the 
treatment.  Then,  in  spite  of  our  wishes, 
the  number  of  roots  per  pot  is  dictated  by 
the  treatment  itself  and  can  be  used  as  a 
concomitant  variable  to  measure  root  pres- 
sure or  root  intensity.  If,  in  addition  to 
root  pressure,  the  experiment  were  carried 
out  in  the  open,  not  in  a greenhouse,  then 
a second  concomitant  vari-able  could  also 
be  used,  namely  rainfall.  These  would 
alter  (6)  as  follows: 

Yijk  = A + B 1 N i j + b2w i j + T i 

+Pj(i)+Eijk>  < 7 > 

wherein,  A replaces  M and  is  considered  the 
intercept;  B-j  is  the  regression  coefficient 
for  the  number  of  roots  per  pot,  and  B2  is 
that  for  inches  of  water  (rain)  received  by 
the  pots.  R k ( j j ) is  now  replaced  by  E j j ^ , 
the  residual  error  after  extracting  SS(B-|, 
B2).  SSE  is  used  to  test  A,  B-|,  and  B2, 
wh i le  SSP  remains  the  error  term  to  test  T 
unless  the  following  is  true. 

Suppose  the  soil  within  each  pot  comes  from 
the  same  batch  as  for  every  other  pot  and 
the  only  differences,  practically  speaking, 
come  from  Njj  and  Wjj  which  are  then  sur- 
rogates for  P j ( j ) . Equation  7 makes  one 
more  metamorphic  leap  to  become 

Yijk  = A + BiNij  + b2w i j 

+ Ti+Eijk-  ( 8 ) 

At  this  juncture  MSE  is  the  error  term  for 
all  comparisons.  SSP  has  been  absorbed  by 
SSE  which  is  now  SSR-SS ( B 1 , B2 ) +SSP . 


DESIGNS  FOR  GRAZING  EXPERIMENTS 

Comparisons  with  Standard  Designs 

Each  of  the  foregoing  models  will  be  recast 
into  designs  which  could  be  used  in  grazing 
trials.  Strengths  and  weaknesses  of  each 
will  be  discussed  from  the  viewpoint  of 
efficiency  of  use  of  experimental  materi- 
als. Lastly,  the  statistical  power  of  the 
test  will  a I so  be  discussed. 

Design  1:  Paddocks  are  IJ  in  number.  They 
are  assigned  to  I groups  as  nearly  homoge- 
neous as  possible  and  are  called  Blocks. 

The  paddocks  do  not  have  to  be  contiguously 
arranged.  To  each  of  the  I blocks  J treat- 
ments are  randomly  assigned.  A treatment, 
let  us  remember,  is  a combination  of  exter- 


nally imposed  conditions  or  applications. 

A treatment  could  consist  of  a particular 
combi-nation  of  components  taken  from  a) 
fertilizer  kind  and  rate  applied  to  the 
soil,  b)  species  or  variety  of  grass,  clo- 
ver or  mixture  of  foraging  crops,  c)  sup- 
plemental feeding,  d)  sources  of  water  for 
the  so i I , e)  stocking  rate,  f)  other  pos- 
sible factors.  Stocking  rate  will  be 
treated  separately  from  all  others  because 
the  grazing  animal  is  the  material  which 
gives  us  our  measurement,  and  it  can  also 
be  part  of  the  treatment  itself. 

a)  Among  the  J treatments,  stocking  rate 
is  held  constant.  Then  (1)  is  the  linear 
additive  model.  The  response  variable  is 
total  weight  gained  per  unit  area  and  ani- 
mal to  animal  variation  contributes  nothing 
to  the  experiment.  Treatments  are  tested 
using  MSBT , and  the  power  of  the  test  de- 
pends entirely  on  df  = (I— 1)(J— 1). 


b)  Stocking  rate  is  a factor  of  the  treat- 
ment combination  but  it  is  not  considered 
to  be  an  interval  measure.  Then,  there  is 
no  change  in  the  above  model. 

c)  Stocking  rate  is  considered  an  interval 
measure  and  separate  from  all  other  vari- 
able combinations  which  make  up  a treat- 
ment. Then,  (1)  is  altered  and  becomes 


Yj  j = A + Cf (Sj  j)  + B|  + Tj  + Ej  j (9) 

wherein  A is  the  intercept;  C is  the  re- 
gression coefficient  for  f(Sjj),  the  re- 
sponse function  for  the  intensity  of 
stocking  or  grazing  pressure;  and  E j j is 
the  residual  error.  The  response  variable 
Is  again  total  gain  per  unit  area. 


Design  2:  Paddocks  are  I JK  in  number.  The 
design  is  that  of  Design  1 above  with  K 
repl icates  of  each  block  by  treatment  com- 
bination. The  linear  additive  model  is 
that  of  (4),  and  the  response  variable 
remains  total  gain  per  paddock  or  norma- 
lized to  accepted  units  of  area  (hectares 
or  acres).  Equation  5 would  be  used  for 
testing,  if  stocking  rate  were  considered  a 
nominal  measure,  but  linear  additive  model 
(9)  would  be  used  if  a response  function 
were  used  to  measure  effects  of  varying 
stock i ng  rates . 


Design  3:  Treatments  are  I in  number,  and 
each  is  applied  to  J paddocks  containing 
Njj  animals.  The  linear  additive  models 
are  those  of  (6),  (7)  and  (8)  or  more  gene- 

ral I y , 

Y | j k = A + B1f(SjJ)  + B2g (W | j ) + T | 

+Pj(i)+Eijk  (1Q) 

or 

Y|  Jk  = A + B j f ( S j j)  + Cg ( W , j) 


+ T i + E j j k 


(11) 


16 


depending  on  whether  S j j and  W | j are 
adequate  surrogates  for  paddocks  Pj(i)- 

Design  4:  A Complex  example:  This  hypothe- 
tical experiment  Is  typical  of  grazing 
research  experiments.  It  consists  of  JK 
treatments  composed  of  K stocking  rates 
together  with  J combinations  of  other 
treatment  factors.  The  entire  experiment 
is  repeated  for  I years.  Instead  of  water 
ava i lable  to  the  sol  I , forage  ava i lable  to 
the  grazing  animal  is  measured.  Weight 
gain  is  measured  on  each  animal  from  which 
other  measures  of  production  can  be  calcu- 
lated. Forage  availability  is  taken  as  a 
surrogate  for  paddock  effect  and  its  re- 
sponse to  the  Tj  by  Sk  combination.  Thus, 
we  are  mapping  the  complex  interaction  of 
paddock  by  treatment  by  stocking  rate  into 
forage  availability,  which  in  turn  is  a 
concomitant  regression  variable  with  a 
causal  effect  on  weight  gain  for  the 
animal . 

The  linear  additive  model  is 


Gijkl  = B0  + B1Fijk  + Yi  + Tj  + TYij 
+ sk  + VS  j k + TSjk  + VTS  j jk 
+A|(jjk).  (12) 


wherein  Gjjk|  = weight  gain. 


B0 

B1 


F i jk 
Yi 

TJ 

Sk 

A I ( i j k ) 


I ntercept , 

Regression  coefficient,  as- 
suming, I inear  dependence  be- 
tween forage  availability  and 
we i ght  gain. 

Forage  ava i labi I ity. 

Year  to  year  effect. 

Treatment  (other  than  Sk)  effect. 

Stocking  rate  effect, 

Animal  effect,  and  YT.YS.TS  and 
YTS  are  the  respective  inter- 
actions. 


Both  Tj  and  Sk  are  considered  fixed  effects 
since  they  are  predetermined  by  the  persons 
conducting  the  experiment,  while  both  Yj 
and  A | ( j j k ) are  random  variables. 

Table  4 is  the  skeletal  ANOVA  table  for 
(12)  includes  the  degrees  of  freedom  and 
the  mean  square  error  used  to  test  each 
term . 


Because  the  error  mean  squares  for  Tj  and 
Sj  are  the  YT  and  YS  mean  squares,  respec- 
tively, this  design  has  a major  flaw.  The 
role  of  the  animal  mean  square  is  that  of 
testing  components  of  variance  of  Y and  its 
interactions  and  the  very  important  regres- 
sion coefficient,  B -j  , of  Fjjk.  In  order  to 
gain  respectable  degrees  of  freedom  of 
error  to  be  used  when  testing  Tj  and  Sk  it 
becomes  necessary  to  rerun  the  exact  same 
experiment  which  usually  requires  a period 
of  sever a I years . 

An  interesting  collapse  of  Table  4 occurrs 
for  1=1.  That  is,  if  the  experiment  is  run 
for  only  one  year,  then  Table  4 becomes 
Tab l e 5 . 

When  1=1,  the  only  random  effect  is  due  to 
animal,  A | ( i j k ) , and  the  power  of  the 
various  tests  depends  on  JK(L-I)  the  total 
number  of  animals,  JK].,  minus  the  number  of 
paddocks,  JK.  JK(L-I)  can  be  a number  of 
reasonable  size  and  which  allows  diffe- 
rences to  be  detected  that  are  due  to  both 
stocking  rates  and  the  other  variables  of 
the  treatments. 

What  is  wrong  with  the  design  in  which  the 
same  experiment  is  repeated  over  a number 
of  years,  and  how  can  it  be  corrected?  The 
strong  dependence  on  the  number  of  years  is 
evident  in  (12)  and  Table  4 because  the  EMS 
df  is  always  that  of  a Year  by  Treatment 
Interaction.  This  Is  the  flaw.  It  can  be 
corrected  rather  simply.  Each  year,  ret i I I 
al I paddocks;  reassign  al I JK  combinations 
of  Tj  and  Sk  to  the  JK  paddocks.  Then  each 
repl icated  experiment  will  be  embedded  or 
nested  within  each  year.  This  clearly  has 
relevance  only  to  annual  pastures.  Within 
Table  5 add  a I i ne  for  Yj  with  df  1-1  and 
MSE  = MSA.  Multiply  al I other  df  by  I 
because  the  I inear  additive  model  is 

Y i j k I = B0  + B i F i J k I + Yi  + T j ( i ) 

+ Sk ( j ) + TS  j k ( | ) 

+ A I ( i jk ) • ( 1 3 ) 

With  this  replicated  experiment  one  cannot 
only  obtain  measures  separating  the  T(,  Sk 
and  TS  j k , but  s/he  can  also  test  for  year- 
to-year  variabi I ity  of  the  effects  of 
forage  availability,  F j j k j , by  testing  for 
paral  lei  ism  from  year  to  year. 


CRITIQUE  OF  SOME  PUBLISHED  DESIGNS 


SSF  is  a partition  of  SSYTS  since  F j j k is 
completely  confounded  with  YTS|jk,  except 
that  F | j k is  a continuous  regression  vari- 
able whereas  YTSjjk  is  discrete.  Thus,  the 
degree  of  freedom  for  F j j k is  subtracted 
from  df  for  YTSjjk  and  not  A | ( j j k ) • 


R I ewe  (1961)  reviewed  a number  of  past 
trials,  giving  their  results  and  setting 
forth  a design  in  which  replication  was  not 
used  or  needed.  His  Table  1 gives  14  corre- 
lations between  stocking  rate  and  gain  per 
animal.  Most  of  the  correlations  have 
only  one  df  for  error.  An  r = -.999  has 
a significance  of  only  .028473  with 


17 


df=1.  However,  the  sum  of  -2log(p)  over 
the  14  tests  of  significance  on  the  corre- 
lations is  a chisquare  with  df  = 28  (See 
Sokal  and  Roh I f , 1981,  p.779).  In  this 

case  it  results  in  a chisquare  equal  to 
96.95  and  a level  of  significance  of  1.7  in 
a bi  I I ion.  The  evidence  is  very  strong 
that  average  gain  per  animal  can  be  ex- 
pressed as  a I inear  function  of  stocking 
r ate . 

When  R i ewe  considered  repl ication,  whether 
on  purpose  or  not,  he  treated  a replication 
as  a block  and  uses  the  block  by  treatment 
interaction  as  error.  His  table  3 with  F, 

P and  a added  and  its  replacements  are 
here  presented  as  Table  6a  and  6b. 

The  linear  additive  model  used  is  that  of 
(1)  when  it  should  have  been  that  of  (6) 
without  the  R k ( i j ) term.  Here,  he  has  lost 
two  degrees  of  freedom  for  error,  and  the 
overall  level  of  significance  (p-value)  was 
four  and  one-half  times  as  large  as  it 
should  have  been.  17 

His  claim  that  a linear  relationship  is 
adequate  to  express  gain  per  animal  as  a 


function  of  stocking 
and  leads  to  the  f o 1 

rate  is 

1 ow i ng : 

we 

1 1 supported 

G = A + 

BS 

(Kg  or 

lb. ) /an i ma 1 

- 

(14a) 

= AS  + 

BS2 

(Kg  or 

1 b ) / ( ha 

or 

ac , ) 

(14b) 

and 

smax 

= - . 5A/B , 

( 14c) 

where i n 

A and 

B are 

regress i on 

coef  f 

i - 

cients;  S is  stocking  rate  (animals  per 
area),  and  B is  negative. 

Petersen,  Lucas  and  Mott  (1965)  developed  a 
theory  I inking  stocking  rate  and  per  animal 
and  per  area  performances  but  did  not  ad- 
dress designs.  Their  initial  assumption 
"Amount  and  type  of  forage  available  per 
acre  are  independent  of  stocking  rate"  has 
been  shown  to  be  unrealistic  by  (Conrad, 
Holt  and  Ellis,  1981.  In  fact,  as  stocking 
rate  increases,  forage  availability  de- 
creases . 

Conn  If fe  (1976)  sets  out  to  compare  between 
and  within  herd  variances  by  collecting 
data  from  12  experiments  where  both  of 
these  could  be  estimated.  His  model  for 
error  i s 


+ 

— \ 

< 

II 

— » 

LU 

C | j + H | 

(15) 

with  expected 

mean  squares  of 

EMS ( between ) 

= Va  + r Vh 

(16a) 

EMS ( w I thin) 

= va  + vc 

(16b) 

where  in  A | j 

= an i ma 1 ef  feet 

cl  J 

= competition,  and 

Hi 

= herd  effect. 

He  argues  away  Cjj  by  summing  it  to  zero 
over  the  herd.  This  is  incorrect!  Cj  j is 
inexorably  confounded  with  and  is  a part  of 
the  animal  variability  itself  and  should 
not  be  part  of  the  model  at  all.  Removing 
Vc  in  (16b)  gives  the  correct  EMS . What  is 
more,  the  linear  additive  model  for  the 
data  he  reports  is  that  of  (6)  in  which 
Hj(j)  replaces  Pj(i)  and  ( i j ) replaces 
R k ( i j ) . 1 n every  case  which  he  reports  in 

his  Table  3,  the  F test  should  be  a one- 
tailed,  right-tailed  test.  He  incorrectly 
uses  a two-ta i led  test  in  every  case.  His 
Table  3 is  here  reproduced  as  Table  7 with 
the  correct  value  of  the  F statistic  and  a 
column  added  for  the  probability  of  a 
larger  F,  where 

F = Between  Herd  MS/within  Herd  MS. 

Again,  the  sum  of  -2log(Pr>F)  is  a chi- 
square with  df  equal  to  twice  the  number  of 
independent  tests  of  significance.  in  this 
case  X2  = 33.745;  df  = 24  and  Pr  > X2  = 
0.0893.  This  alone  is  enough  to  fail  to 
reject  the  hypothesis  that  the  between- 
herd  component  of  variance  is  zero.  But 
Coniffee  impeaches  his  1 964-Moor epark  data, 
the  one  very  large  between-herd  estimate  of 
the  vac iance.  If  we  exclude  that  test 
then  X2  = 18.944,  df  = 22,  and  Pr  > X2  = 
0.649.  This  would  lead  me  to  conclude  that 
the  between-herd  variance  is  either  negli- 
gible or  simply  nonexistent. 


CONCLIJS  ION 


Other  studies  could  be  Included,  but  that 
would  be  stretching  the  point.  From  my 
perspective,  empirical  forage  research  has 
advanced  to  the  point  that  it  is  insepa- 
rable from  that  of  mathematical  and  statis- 
tical modeling.  There  is  an  unlimited 
number  of  designs  ava i lable  for  use  in 
conduct inng  grazing  research.  Whether 
replication  of  one  kind  or  another  can/must 
be  used  depends  entirely  upon  the  modeling 
s/he  is  will i ng  to  use  and  defend.  Experi- 
mental materials  are  very  expensive  in  this 
research,  which  goes  almost  without  saying. 
A small  quirk  In  the  experiment  as  actually 
performed,  against  all  good  intentions, 
can  result  in  loss  of  degrees  of  freedom  In 
the  error  mean  square  and  diminish  the 
power  of  the  test  to  the  point  of  voiding 
the  entire  experiment.  Caveat  emptor ! 


18 


Literature  Cl  ted : 

Conniffe,  D.  (1976)  A comparison  of 

between  herd  and  within  herd  variance  in 
grazing  experiments.  Irish  Journal  of 
Agriculture  Research  15:39-46. 

Conrad,  B.E.,  Holt,  E.C.  and  Ellis,  W.C. 
(1981)  Steer  performance  on  costal, 
cal  I ie  and  other  hybrid  bermudagrasses . 
Journal  of An ima 1 Sciences  53:1188-1192. 

Montgomery,  D.C.  (1984)  Desing  and  Analyses 
of  Exper i ments . 2nd  Ed.,  New  York:  John 
Wiley  & Sons . 


Petersen,  R.G.,  Lucas,  H.L.  and  Mott,  G.O. 
(1965)  Relationship  between  rate  of 
stocking  and  per  animal  and  per  acre 
performance  of  pasture.  Agronomy  Journal 
57 : 27-30 . 

Riewe,  M.E.  (1961)  Use  of  the  relationship 
of  stocking  rate  to  gain  of  cattle  in  an 
experimental  design  for  grazing  trials. 
Agronomy  Journal  53:309-313. 

Sokal,  R.R.  and  Roh I f , F.J.  (1981) 

B i ometry , 2nd  Ed.,  San  Francisco: 

W.H.  Freeman  and  Company. 

Steele,  R.G.D.  and  Torrie,  J.H.  (1980) 
Principles  and  Procedures 
of  Statistics,  A Biometrical  Approach, 
2nd  Ed . , New  York : McGraw-H ill. 


Table  1:  Skeletal  ANOVA  for  a two-way  completely  crossed  experiment 
(Randomized  Complete  Block).  Sj,  SjSj,  etc.  indicates  a 
sum  over  all  values  of  I , i and  j,  etc.,  respectively. 

Ve  is  the  error  variance. 


Source 

df 

SS  Contrast 

Expected  Mean  Square 

B 1 ocks 

T r tmnt 

BxT 

1 -1 

J -1 

( 1-1 ) ( J-1  ) 

li.  -Y.. 

— • j -I.. 

Yi  j-Y  j ,-Y.  j+Y.  . 

JS  j B j 2 / ( 1 -1  ) + Ve 

|sjTj2/( J-1 ) + ve 

SjSjBTj  j2/(  ( 1-1  ) ( j-1  ) ) + Ve 

P 1 ots 

1 J -1 

Y i j -I.. 

Error 

0 

Not  Ava i 1 ab  1 e 

Not  Ava i 1 ab  1 e 

Total 

1 J -1 

Yi j - I.. 

Table  2:  Skeletal  ANOVA  of  a two-way  completely  crossed  experi- 
ment with  repl ications  within  every  i j combination. 

Source 

df 

SS  Contrasts 

Expected  Mean  Square 

B 1 ocks 

1 -1 

li . -I. . 

JKS  j B j 2 / ( 1-1 )+Ve 

T r tmnt 

J -1 

Y.j  -Y.. 

IKSjTj2/( J-1 )+Ve 

BxT 

( 1 - 1 ) ( J-  1 ) 

Yi j .-I, . . 

KSiSjBTj  j2/(  ( 1-1  )(J-1  ) )+Ve 

P 1 ots 

1 J -1 

— 1 j • -I... 

Error 

1 J(K  -1  ) 

Yi  jk  -Yj  j . 

ve 

Total 

UK  -1 

Yi jk  -I. . . 

19 


Table  3:  Skeletal  ANOVA  of  a completely  nested 
exper i ment . 


Source 

df 

SS  Contrast 

Expected  Mean  Square 

T r tmnt 

1 -1 

Yj  . ,-Y.  . . 

JKS j T j 2 / ( 1 -1 ) +KVp  + Vr 

Pot 

I -1 

lij-  -Yj-- 

KVp  + Vr 

Root 

1 J ( K- 1 ) 

Yi  jk  "Yj  j • 

vr 

Total 

1 JK-1 

Yj  jk  -I-  • • 

Table  4:  Skeletal  ANOVA  giving  degrees  of  freedom  and 

identifying  appropriate  error  mean  squares  for 
testing  various  components  of  the  linear  additive 
model . L is  the  average  number  of  animals  per 
paddock . 


Source 

df 

Error  Mean  Square 

F i jk 

1 

A 

Yi 

1 -1 

A 

TJ 

J-  1 

YT 

YT  i j 

( 1-1 )( J-1  ) 

A 

sk 

K—  1 

YS 

YSik 

( 1-1 ) (K-1 ) 

A 

TSjk 

(J-1  ) (K-1 ) 

YTS 

YTSj  jk 

( 1 — 1 ) (J  — 1 ) (K  — 

1 ) — 1 A 

A 1 ( i j k ) 

1 JK ( L- 1 ) 

— 

Tota  1 

1 JKL-1 

— 

Table  5:  Skeletal  ANOVA  giving  degrees  of  freedom  and  identifying 
appropriate  error  mean  squares  for  testing  various 
components  of  the  linear  additive  model. 

Source 

df 

Error  Mean  Square 

FJk 

1 

A 

TJ 

J-1 

A 

sk 

K-1 

A 

TSjk 

( J-1  ) (K-1  ) -1 

A 

Al  ( jk) 

JK ( L— 1 ) 

— 

Tota  1 

JKL-1 

20 


Table  6a:  R i ewe ' s Table  with  F/P,  ch I square  and  overall 
significance  added. 


Source 

df 

Mean 

Square 

1 957 

F/P 

Mean 

Square 

1 958 

F/P 

Reps 

2 

264 

- 

1969 

- 

Stock i ng  Rate 

1 

5891 

38 . 5 

7350 

8 . 06 

Error 

2 

1 53 

. 025 

912 

. 105 

X 


2 

4 


11.89 


Pr >X24 


.018 


Table  6b:  Replacement  of  Riewe's  Table  3 with  F/P,  chisquare 
and  overall  significance  added. 


Mean  Mean 

Square  Square 

1957  F/P  1958  F/P 


Stocking  Rate  1 5891 

Reps  4 208 . 5 


28.3  7350 

006  1440.5 


5.10 
. 087 


X24  = 15.12, 


Pr>X24 


. 004 


Table  7 : Conn iffee's  Table  3 with  abbreviated  Experi ment  caption, 
corrected  Fs  and  Pr  > F added. 


Between  Within 

Experiment  Herd  df  Herd  df  F Pr  > F 


1 963  , 

Moor epark 

1 , 770 

2 

704 

76 

2 . 33 

. 10421 

1964 

4 , 200 

2 

514 

76 

8.17 

.00061 108 

1965 

664 

2 

421 

78 

1 . 58 

.21250 

1 966 

58 

2 

638 

78 

. 091 

.91311 

1 967 

271 

2 

734 

79 

. 369 

.69261 

1 968 

1 ,472 

2 

663 

34 

2 . 22 

.12411 

1 969 

82 

2 

422 

80 

. 1 94 

. 82404 

1 970 

1 , 275 

3 

687 

131 

1 . 86 

. 1 3953 

1 967  , 

Grange 

1 45 

3 

345 

29 

. 420 

. 74002 

1 968 

1 50 

3 

323 

30 

. 464 

. 70954 

1 969 

89 

3 

337 

30 

. 264 

. 85077 

1 970 

171 

4 

533 

36 

. 32  1 

. 86204 

21 


FORAGE  BREEDERS  INFORMATION  EXCHANGE  GROUP 
BREEDING  SELF  POLLINATED  FORAGE  LEGUMES1 2 

D.  D.  Baltensperger^ 

"Self-pollinated  forages  are  mostly  annuals.  In 
general,  they  are  grown  less  extensively  and 
have  less  economic  importance  than  the  cross- 
pollinated  species.  For  this  reason  improvement 
in  self-pollinated  forages  has  been  limited  to 
relatively  few  species,  such  as  the  lespedezas, 
vetches,  and  cowpeas." 

POEHLMAN 

Breeding  Field  Crops  1959  (8) 

"The  annual,  self-pollinated  forage  species  are 
limited  in  number  and  are  minor  in  importance 
compared  to  cross  pollinated  species.  For  these 
reasons  they  have  received  less  attention." 

POEHLMAN 

Breeding  Field  Crops  (2nd  Ed.) 

1979  (9) 

With  these  thoughts  by  Dr.  Poehlman  we  might  ask 
why  even  worry  about  breeding  techniques  with 
self-pollinated  forages.  The  answer  I think  may 
lie  in  part  with  rejecting  the  above 
assumption.  Granted  alfalfa,  white  clover  and 
red  clover  have  probably  been  the  most  worked  on 
and  utilized  of  the  forage  legumes  to  this  point 
in  time,  but  crops  such  as  subclover, 
Aeschynomene,  Stylosanthes,  and  leucaena 
probably  have  potential  for  expansion  to  more 
acres  than  the  cross  pollinated  legumes, 
especially  in  the  southeastern  U.S.  In 
Louisiana  and  Florida  the  seed  sales  of  these 
self-pollinated  crops  is  eclipsing  that  of  the 
cross  pollinated  species  and  self  pollinated 
seed  sales  are  not  far  behind  in  Arkansas  and 
Alabama  where  Lespedeza  and  vetches  are  also 
important. 

However,  deciding  that  self-pollinated  forages 
are  worthy  of  a breeding  effort  is  only  the 
first  hurdle  to  overcome  as  a breeder.  It  is 
easy  to  say  that  one  will  use  any  of  the  many 
techniques  developed  by  breeders  of  self- 
pollinated  row-crops.  The  application  of  these 
techniques  is  not  as  easy.  Self-pollinated 
breeding-theory  is  based  on  recombination 
following  crossing  whether  a backcross, 
pedigree,  mass  selection,  single-seed  descent  or 
alternative  technique  is  utilized.  Yet  this  has 
been  the  major  stumbling  block  for  previous 
self-pollinated  breeding  efforts.  There  are  two 
basic  reasons  for  this:  1)  the  botany  including 
taxonomy  of  these  crops  is  poorly  understood;  2) 
the  crops  by  their  very  nature  (forages) 
frequently  set  a small 


1 Florida  Agric.  Exp.  Stn.  Journal  Series  No. 
7904. 

2 Associate  Professor,  Agronomy  Dept., 

University  of  Florida.  IFAS,  Gainesville, 

FL  32611 


percentage  of  flowers  even  when  they  have  not 
been  manipulated  to  emasculate  them  and  cross 
them. 

Subterranean  clover  has  received  the  most  effort 
of  the  self-pollinated  forages  (10)  and  has 
been  separated  into  three  species  based  on 
crossing  barriers  (T.  subteraneum  L. , T. 
yanninicum  Katzn.  and  Morley  an  T. 
brachycal vein urn  Katzn.  and  Morley)  (6).  It 
appears  that  Desmodium  spp.  and  Stylosanthes 
spp.  may  have  many  similar  groupings  (K.  H. 
Quesenberry  and  J.  B.  Brolmann,  personal 
communication)  and  the  entire  classification  of 
Alysicarpus  spp.  has  recently  been  reworked  (4). 
Quesenberry  and  Deren  have  identified 
Aeschynomene  americana  crosses  that  produce 
seed,  but  the  F^  plants  all  die  at  approximately 
the  third  true-  leaf  stage  (K.  H.  Quesenberry 
personal  communication).  Common  alyceclover  is 
a typical  example  of  the  problem  with  poor  seed 
set.  For  every  two  flowers  ; need  only  one 
pod  is  made  under  natural  conditions  (11). 
Attempting  crosses  without  emasculation  resulted 
in  over  70%  of  the  flowers  aborting  and  when 
emasculation  was  attempted  100%  of  the  flowers 
aborted  (11).  Subsequent  work  has  indicated 
that  all  seed  set  were  seifs.  All  currently 
available  alyceclover  and  aeschynomene  seed 
comes  from  new  or  old  collections  of  local 
strains  or  plant  introductions.  Hardy  and 
Quesenberry  have  developed  crossing  techniques 
for  Aeschynomene  (5).  Brolmann  has  developed 
techniques  for  Phaseolus  and  Centrosema  but  has 
not  gone  further  as  adequate  variation  exists  in 
P.I.  material  (2).  The  majority  of  subclover 
cultivars  also  come  from  new  and  old  collections 
of  local  strains  (6).  Kobe,  a cultivar  of 
common  lespedeza,  was  a direct  introduction  from 
Japan.  'Climax’  a variety  of  Korean  lespedeza 
(L.  stipulacea) , originated  as  a selection  out 
of  an  introduction  from  China.  'Rowan' 
originated  from  a single  plant  selection  out  of 
Korean  lespedeza(9) . 

The  difficulty  of  making  crosses  in  these  crops 
may  be  readily  overcome  as  the  botany  of  the 
crops  is  better  understood  (12).  The  inability 
to  readily  make  crosses  has  perhaps  not  been  a 
serious  limit  to  the  breeder  up  to  this  point  as 
so  much  variation  did  exist  in  local  strains  and 
plant  introductions,  but  further  progress  in 
most  is  dependent  upon  recombining  some  of  the 
favorable  traits  now  identified  into  a single 
cultivar.  Selection  of  good  parents  for 
crossing  is  an  important  phase  and  must  not  be 
overlooked  or  P.I.'s  will  be  brought  in  that  are 
better  than  our  improved  cultivars. 

This  brings  us  to  the  next  problem  facing  a 
self-pollinated-forage  breeder.  What  is  the 
genetic  make  up  of  a desirable  self-pollinated- 
forage  cultivar?  Should  it  be  heterogeneous  and 
homozygous,  or  homogeneous  and  homozygous? 
Especially  with  perennial  forages  the  question 
of  the  role  of  heterogeneity  in  disease  and 
insect  resistance  becomes  important.  The 
current  seed  laws  including  the  P.V.P.  are 
certainly  easier  to  invoke  on  homozygous 


22 


homogeneous  cultivars  than  on  alternatives,  but 
that  doesn't  mean  that  they  are  the  best 
perennial  forage  cultivars.  Perhaps  this  can  be 
avoided  by  applying  for  P.V.P.  of  component 
lines  and  releasing  a multiline  cultivar.  This 
issue  is  important  because  it  has  such  a large 
impact  on  how  early  testing  can  and  should  be 
initiated  in  the  program. 

The  Australian  subterranean  clover  breeding 
program  (3,  7)  has  been  directed  toward  a 
homozygous  and  homogeneous  cultivar.  It  has 
been  based  on  a five  stage  testing  program 
initiated  at  homozygosity,  with  initial 
characterization  for  maturity,  hard  seededness, 
formononetin  content  and  other  readily  measured 
characters.  Stage  2 looks  at  similar 
characters,  but  in  alternative  environments. 
Stage  3 evaluates  disease  and  insect  resistance 
and  Stage  4 and  5 are  field  tests  where 
defoliation  effects  and  persistence  characters 
are  measured.  A maximum  of  10  lines  make  it  to 
Stage  5 in  each  cycle.  This  program  has 
resulted  in  the  release  of  several  new  cultivars 
in  Australia. 

We  have  skipped  from  producing  the  FI  to 
homozygosity.  My  interviews  with  current  self- 
pollinated-forage  breeders  including  J.A. 
Mosjidis  (Lespedezas) , Ann  Marie  Thro 
(Aeschynomene , Stylosanthes) , K.  H.  Quesenberry 
(Aeschynomene , Desmodium) , J.  B.  Brolmann 
(Stylosanthes) , Homer  Wells  (Lupine),  G.  R. 

Smith  (Trifolium  hirtum)  and  G.  M.  Prine 
(Cajanus  cajun)  indicate  that  this  area  has 
probably  not  received  adequate  attention  because 
of  the  limited  number  of  crosses  made.  Lupine 
breeding  at  Tifton,  Georgia  has  utilized  a 
pedigree  selection  program  for  gray  leaf  spot 
resistance,  cold  tolerance,  bitterness  and  virus 
resistance  (H.  Wells  personal  communication). 
This  is  similar  to  programs  on  cowpeas,  peanuts 
and  soybeans  which  like  lupine  have  potential 
uses  other  than  forage  and  have  relatively  large 
flowers. 

Quesenberry  has  initiated  a pedigree  program  for 
both  Desmodium  and  Aeschynomene,  but  this 
program  has  not  yet  completed  a breeding  cycle. 

Brolmann  has  utilized  a pedigree  system  to 
evaluate  naturally  occurring  crosses  in 
Stylosanthes  spp.  (personal  communication). 

This  is  based  on  outcrossing  percentages  of 
nearly  10%  at  Fort  Pierce,  FI  compared  with  less 
than  1%  in  Australia  (1).  He  considers  his  end 
product  to  be  a population  that  is  less  than 
homogeneous  and  homozygous,  but  uniform  for 
important  traits. 

It  appears  that  this  is  an  area  ripe  for 
investigation  as  crossing  capabilities  improve. 
Unfortunately,  the  pedigree  system  is  based  on 
the  idea  that  1 plant  produces  1-3  rows  etc. 

Most  forage  legumes  are  grown  as  mixed  swords 
and  the  limited  quantities  of  seed  early  in  the 
program  are  not  conducive  to  such  evaluations. 
Forage  yields  are  also  destructive  as  far  as 


additional  seed  increase  is  concerned. 

Producing  adequate  seed  for  evaluation  under 
grazing  is  an  extremely  time  consuming  and 
expensive  operation.  Evaluating  specific  traits 
such  as  disease  and  insect  resistance  works 
well,  but  produces  many  lines  which  are 
intolerant  of  grazing.  The  pedigree  system 
because  of  its  advantages  in  genetic  studies 
will  continue  to  be  heavily  utilized  by  self- 
pollinated  forage  legume  breeders. 

Single-seed  descent  has  many  appealing 
characteristics  to  the  forage  breeder  including 
1)  rapid  advance  to  homozygosity,  2)  use  of 
greenhouse  or  seed  production  locations  for 
growth  without  testing,  3)  limited  manpower 
requirements  in  steps  prior  to  homozygosity.  It 
however,  requires  modification  for  use  by  most 
forage  legume  breeders  as  most  forage  legumes 
produce  less  than  one  plant  per  seed,  and  some, 
such  as  alyceclover  and  desmodium  may  not  even 
produce  one  plant  for  every  two  seeds.  The 
single  seed  descent  can  easily  be  modified  to 
take  1 pod  (or  a half  dozen  seed). 

Mass  selection  has  certainly  worked  well  in  the 
past  providing  the  local  ecotypes  and  plant 
introductions  that  have  been  so  valuable  in 
selection  programs.  It  does  not,  however,  take 
the  greatest  advantage  of  a small  number  of 
crosses  and  that  is  more  than  likely  what  the 
self-pollinated  forage  legume  breeder  will  have 
to  work  with. 

In  summary  the  major  challenge  of  self- 
pollinated  breeders  is  to  develop  a better 
understanding  of  their  crops.  This  will  lead  to 
improved  crossing  capabilities  and  open  the  door 
for  much  needed  research  on  breeeding 
techniques. 

REFERENCES 

1.  Brolmann,  J.B.  1974.  Progeny  Studies  in 
Stylosanthes  guyanensis  (Aubl.)  SW.  Soil  and 
Crop  Science  Society  of  Florida  Proceedings, 
33:22-24. 

2.  Brolmann,  J.B.  and  A.E.  Kretschmer,  Jr.  1973. 
Agronomical  and  Morphological  Characteristics 
of  Phasey  Bean  (Phaseolus  lathyroides  L.). 
Soil  and  Crop  Science  Society  of  Florida 
Proceedings,  32:50-52. 

3.  Gillespie,  D.J.  and  J.D.  Sandow.  1981. 
Selection  for  Blue  Green  Aphid  Resistance  in 
Subterranean  Clover.  Proc.  XIV  Int.  Grassld. 
Congr.,  14:105-108. 

4.  Gramshaw,  D. , B.C.  Pengelly,  F.W.  Muller, 
W.A.T.  Harding  and  R.J.  Williams.  1987. 
Classification  of  a Collection  of  the  Legume 
Alysicarpus  Using  Morphological  and 
Preliminary  Agronomic  Attributes.  Aust.  J. 
Agric.  Res.,  38:355-372. 

5.  Hardy,  S.R.  and  K.H.  Quesenberry.  1984. 
Artificial  Hybridization  of  Aeschynomene 


23 


americana  L.  (A  tropical  forage  legume). 
Soil  and  Crop  Science  Society  of  Florida 
Proceedings,  43:163-166. 

6.  McGuire,  W.S.  1985.  Subterranean  Clover, 
in : Clover  Science  and  Technology,  pp.515- 
534.  Ed.  N.L.  Taylor.  American  Society  of 
Agronomy,  Crop  Science  Society  of  America. 

616p. 

7.  Nicholas,  D.A.  and  D.J.  Gillespie.  1981. 
Procedure  for  Selecting  Subterranean 
Clover  Cultivars  in  South  Western 
Australia.  Proc.  XIV  Int.  Grassld.  Congr., 
14:135-137. 

8.  Poehlman,  J.M.  1959.  Breeding  Field  Crops. 
Holt,  Reinhart  and  Winston,  Inc.  427p. 

9.  Poehlman,  J.M.  1979.  Breeding  Field  Crops, 
Second  Edition.  AVI  Publishing  Company, 
Inc.  483p. 

10.  Taylor,  N.L.  1987.  Forage  Legumes,  iai: 
Principles  of  Cultivar  Development,  v.2: 
Crop  Species,  pp. 209-248.  Ed.  W.R.  Fehr. 
Macmillan  Publishing  Company.  761p. 

11.  Taylor,  S.G.  1984.  Variation  in  Root-Knot 
Nematode  (Meloidogyne  spp.)  Resistance  and 
Other  Agronomic  Characteristics  in 
Alyceclover  (Alysicarpus  spp.)  Germplasm. 
Master's  Thesis,  University  of  Florida, 
Gainesville.  129p. 

12.  Townsend,  C.E.  1980.  Forage  Legumes,  in: 
Hybridization  of  Crop  Plants,  pp.  367- 
380.  Ed.  W.R.  Fehr  and  H.H.  Hadley. 
American  Society  of  Agronomy,  Crop  Science 
Society  of  America.  765p. 


24 


TANNINS  IN  FORAGE  LEGUMES  AND  IMPLICATIONS  FOR  A 
BREEDING  PROGRAM 

K.  H.  Quesenberryl 


INTRODUCTION 


The  use  of  plant  derived  tannins  for  leather 
making  is  recorded  as  early  as  1500  B.C. 

(Haslam,  1966).  Thus  the  existence  of  tannin 
compounds  has  long  been  known  to  man.  In 
addition  to  this  use  in  the  leather  industry 
tannins  have  other  industrial  uses  including  use 
in  inks,  plastics  and  dyes.  In  fodder  plants 
tannins  have  been  regarded  as  desirable  (for 
possible  protection  against  bird,  insect  and 
disease  attack  and  against  bloat  in  grazing 
animals)  or  as  undesirable  (because  of  their 
adverse  affects  on  animal  acceptance  and/or 
digestibility) . 

VARIABILITY  IN  PLANT  TANNIN  TYPE  AND  CONTENT 

As  cited  by  Halsam  (1966),  the  most  widely 
accepted  classification  of  vegetable  tannins  is 
that  by  Freundenberg  who  divided  them  into 
condensed  tannins  (those  which  do  not  readily 
break  down  under  acid  hydrolysis)  and 
hydrolyzable  tannins  (those  which  have  a 
polyester  structure  readily  hydrolysed  by  acids. 
The  principal  forage  tannins  are  of  the 
condensed  type  (McLeod,  1974).  Sarkar,  et  al. , 
(1976)  further  divided  the  condensed  tannins 
into  proanthocyanidins  (flavan-3-ol  polymers) 
and  leucoanthocyanidins  (flavan-3,4-diol 
polymers)  (Fig.  1). 


Figure  1.  Classification  of  tannins  and  their 

monomer  units.  (After  Sarkar  et  al. , 
(1976) 

TANNINS 


FLAVOLANS  HYDROLYZABLE  TANNINS 

(Condensed  tannins)  (esters  of  sugars 

(flavanol  polymers)  and  gallic  acid) 


PROANTHOCYANIDINS 
(polymers  of 
f lavan-3-ols) 

FLa!aN-3-0LS 
catechin 
epicatechin 
gallocetechin 
epigallocatechin 

^Professor,  Department  of  Agronomy,  University 
of  Florida,  Gainesville,  FL  32611 


Most  researchers  agree  that  the  main  pathway  for 
formation  of  condensed  tannins  is  through 
polymerization  of  leucoanthocyanidins  either 
alone  or  in  conjunction  with  other  flavanoids 
such  as  catechins  (McLeod,  1974).  Tannins  are 
only  a fraction  of  the  polyphenols  present  in 
plants  and  much  of  the  data  on  the  occurrence  of 
tannins  in  plants  is  questionable  because  it  is 
based  on  non-specific  methods  of  identification. 
Further,  several  investigators  (Bate-Smith, 

1973;  Donnelly  and  Anthony,  1973;  and 
Quesenberry  and  Albrecht,  1987)  have  shown  that 
tannin  content  may  vary  within  the  plant  with 
young  leaves  having  higher  content  than  older 
leaves  and  leaves  having  a higher  percentage 
than  stems.  Tannins  are  generally  thought  to  be 
sequestered  in  vacuoles  within  the  plant  cells. 

Although  the  above  variability  in  tannin  content 
among  maturity  stages  and  tissue  types  is  known 
to  exist,  major  differences  among  and  within 
forage  plant  species  have  been  identified. 

Table  1 summarizes  tannin  content  of  various 
forage  legume  genera  as  reported  in  several 
references.  Some  genera  such  as  Glycine , 
Trigonella,  and  Vicia  had  no  species  with 
tannins  while  other  genera  such  as  Astragalus , 
Medicago,  and  Trifolium  had  some  tannin 
containing  species  and  others  without  tannins. 
All  species  examined  of  a third  group  of  genera 
including  Desmodium  and  Onobrychis  contained 
tannins.  Methods  used  to  obtain  the  data 
summarized  in  this  table  varied  among 
researchers  and  only  a limited  number  of  species 
were  sampled  for  some  genera. 


Table  1.  Tannin  content  of  various  forage 
legume  genera. 


Genera 

Tannin + 

Reference 

Astragalus 

V 

Davis,  1973 

Coronilla 

+ 

Marshall  et 

al.  , 

1979 

Desmodium 

+ 

Rotar,  1965 

Doryvnium 

+ 

Marshall  et 

al. , 

1979 

Glycine 

- 

Marshall  et 

al. , 

1979 

Indigofera 

+ 

Marshall  et 

al. , 

1979 

Lotus 

+ 

Foo  et  al. , 

1982 

Lathyrus 

V 

Marshall  et 

al. , 

1979 

Lespedeza 

+ 

Stitt,  1943 

Lupinus 

- 

Marshall  et 

al. , 

1979 

Medicago 

V 

Marshall  et 

al. , 

1979 

Onobrychis 

+ 

Marshall  et 

al.. 

1979 

Trifolium 

V 

Marshall  et 

al. , 

1979 

Trigonella 

- 

Marshall  et 

al.  , 

1979 

Vicia 

— 

Marshall  et 

al. , 

1979 

+V  = species  within  genera  vary  for  presence  or 
absence  of  tannin,  + = all  species  reported  in 
genera  contain  tannin,  - = no  species  reported 
contained  tannin. 

Marshall,  et  al.  (1979)  reported  the  results  of 
screening  a large  number  of  Trifolium  spp.  for 
tannin  content.  Pandey  (1971)  also  reported 
screening  30  species  of  Trifolium  for  tannin 
content.  Table  2 identifies  Trifolium  spp. 


LEUC0ANTH0CY AN IDI NS 
(polymers  of  flavan- 
3,4-diols) 


25 


which  have  been  reported  to  contain  tannins. 
Others  such  as  Davis  (1973)  and  Rotar  (1965) 
have  screened  large  collections  of  Astragalus 
and  Desmodium  spp.,  respectively,  for  tannin 
content. 


Table  2.  Trifolium  spp.  containing  tannins  + 


Species 

Reference 

T.  alpestre^ 

Pandey,  1971 

T.  arvense 

Sarkar  et  al. , 
Pandey,  1971 

1976; 

T„  campestre 

Sarkar  et  al. , 
Marshall  et  al. 

1976; 

, 1979 

T.  dubium 

Sarkar  et  al. , 

1976; 

T.  hohenackeri 

T.  rubensj 

T.  trichocephalumj 

Marshall  et  al. 
Pandey,  1971 
Pandey,  1971 
Pandey,  1971 

, 1979 

+Over  60  other  Trifolium  species  reported  to 
not  contain  tannins. 

^Tannins  only  found  in  a few  plants  and  only 
under  water  stress  conditions. 


METHODS  FOR  QUANTIFICATION  OF  TANNINS 

Vanillin  HC1  Method 

The  vanillin  HC1  method  of  Burns  (1963)  or  some 
modification  of  it  have  been  the  preferred  rapid 
screening  method  for  determining  tannin  content. 
Maxon  and  Rooney  (1972)  modified  this  procedure 
to  include  1%  concentrated  HC1  in  the  methanol 
extract  rather  than  pure  methanol.  Walton  et 
al. , (1983)  showed  that  either  of  the  above 
procedures  could  give  false  positives  with 
sorghum  [ Sorghum  bicolor  (L.)Moench.]  and 
recommended  a chloroform-HCl  modification  to 
remove  the  chlorophyll  from  the  extract.  They 
pointed  out  that  sorghum  forage  extracted  with 
methanol-HCl  with  or  without  vanillin  added 
developed  red  color.  They  suggest  that  this 
color  reaction  is  characteristic  of 
leucoauthocyanidins.  Subtracting  the  color  in 
methanol-HCl  blanks  without  vanillin  could  lead 
to  the  conclusion  that  the  plant  has  no  tannin 
like  substances,  and  use  of  uncorrected  values 
might  lead  to  the  conclusion  that  the  plant  has 
high  levels  of  condensed  proanthocyanidins. 

Chromatographic  and  Other  Techniques 

Sakar  et  al.  (1976)  have  described 
chromatographic  techniques  to  further 
characterize  plant  tannins.  Additional 
characterization  techniques  using  NMR  and  other 
techniques  have  been  used  by  Foo  et  al.  (1982) 
to  study  tannin  polymers.  In  general  these 
techniques  do  not  seem  to  be  adapted  to  the 
large  numbers  and  rapid  time  schedule  needed  by 
breeders  in  a screening  technique. 


Protein  Precipitation  Techniques 

Hagerman  and  Butler  (1978)  have  investigated 
various  protein  precipitation  methods  for 
quantifying  tannin  content  of  plants. 

Quesenberry  and  Albrecht  (1987)  reported  the  use 
of  one  such  technique  to  determine  the  tannin 
levels  in  a group  of  Desmodium  spp.  They  showed 
that  laboratory  run  to  run  variability  of  the 
procedure  was  low.  Immature  leaves  were  higher 
in  tannin  than  the  first  fully  expanded  leaf  or 
mature  leaves.  Mature  stems  and  petioles  had 
the  lowest  tannin  content.  They  reported  tannic 
acid  equivalents  (TAE)  in  D.  heterocarpon 
ranging  from  115  to  51g  kg~^  and  a mean  TAE 
content  of  D.  uncinatum  and  D.  intortum  of  82 
and  75  g kg~^ , respectively.  This  technique  has 
many  of  the  features  of  the  vanillin-HCl 
procedure  and  may  be  more  repeatable.  Further 
work  is  needed  to  correlate  these  results  with 
the  vanillin-HCl  procedure  on  a range  of 
species  and  to  correlate  tannin  levels  measured 
by  all  these  techniques  with  changes  in  animal 
digestive  behavior. 

BREEDING  FOR  ALTERED  TANNIN  LEVELS 

As  indicated  in  the  introduction  tannins  as 
components  of  forage  plants  may  have  both 
beneficial  and  harmful  affects.  Much  of  the 
germplasm  evaluation  research  carried  out  in  the 
genera  Medicago  and  Trifolium  has  been  for  the 
purpose  of  identifying  species  with  tannins. 

The  desired  objective  then  being  to  transfer 
this  characteristic  into  species  of  these  genera 
which  are  of  agronomic  importance,  but  which 
have  potential  for  causing  bloat  in  grazing 
livestock.  Although  species  have  been 
identified  which  contain  tannins,  no  success  has 
been  achieved  in  transferring  this 
characteristic  to  alfalfa  or  any  of  the  clovers 
of  agronomic  importance.  This  research  has 
shown  that  the  bloat  safe  characteristic  of 
sanfoin  (Onobychis  viciifolia  Scop.)  and 
several  Lotus  spp.  can  be  attributed  to  the 
presence  of  condensed  tannins.  Since  the 
biochemical  pathway  for  synthesis  of  tannins  is 
complex,  if  appears  unlikely  that  a single  gene 
trait  will  be  identified  which  is  amenable  to 
genetic  transformation  techniques.  Thus  the 
possibility  of  breeding  for  increased  tannin 
levels  in  temperate  legume  species  which  do  not 
exhibit  naturally  occurring  tannins  may  be 
limited. 

The  most  well  characterized  program  of  selection 
for  decreased  tannin  levels  in  forage  legumes  is 
that  of  sericea  lespedeza  [Lespedeza  cuneata 
(Dum.  de  Caours.)  G.  Don]  carried  out  primarily 
by  Donnelly  and  co-workers  at  Auburn  University. 
Early  research  by  Donnelly  (1954)  showed  a high 
correlation  of  cattle  grazing  preference  with 
fine  stems  and  low  tannin  content.  Bates  and 
Henson  (1955)  studied  the  inheritance  of  tannin 
content  in  sericea  and  estimated  heritability 
values  of  from  34  to  43%.  Using  a formula 
suggested  by  Wright,  they  also  estimated  that 
gene  number  involved  in  tannin  inheritance  was 


26 


19  to  24.  In  a later  study  Cope  (1962) 
estimated  the  heritability  of  tannin  content  to 
be  71%.  He  noted  a correlation  of  from  0.29  to 
0.53  of  tannin  content  with  yield,  but  suggested 
that  this  relationship  was  largely  a 
developmental  phenomenon  related  to  plant 
maturity.  Donnelly  (1959)  studied  season  and 
maturity  effects  on  tannin  content  of  sericea 
and  found  that  tannin  content  increased  as 
temperature  increased  and  precipitation 
decreased.  Tannin  content  also  increased  with 
plant  maturity.  These  reports  suggested  that 
rapid  progress  should  be  possible  in  selecting 
for  decreased  tannin  content  in  sericea,  but 
that  precautions  would  be  required  related  to 
the  seasonal  changes  in  tannin  content  and  to 
avoid  decreased  yield  when  selecting  for 
decreased  tannin. 

A major  concern  in  the  program  of  selection  for 
decreased  tannin  in  sericea  was  determination  of 
the  effects  on  dry  matter  digestibility  (DMD) 
and  subsequent  animal  performance.  Donnelly  and 
Anthony  (1970)  found  a 23%  increase  in  DMD  (56% 
to  69%)  of  low  tannin  lines  compared  to  high 
tannin  lines.  In  further  studies  Donnelly  et 
al.  (1971)  found  that  decreased  tannin  levels 
resulted  in  improved  digestible  dry  matter 
intake  of  grazing  animals.  This  study  was 
complicated  by  the  fact  that  the  low  tannin 
lines  used  had  less  vigor  and  did  not  produce 
sufficient  forage  for  selective  grazing.  This 
result  gave  emphasis  to  the  need  to  select  for 
high  yield  while  selecting  for  low  tannin 
content.  Donnelly  and  Anthony  (1973)  also 
showed  that  breeding  sericea  low  in  tannin 
concentration  not  only  increased  digestibility 
of  dry  matter  and  crude  protein  (CP)  but  that  it 
also  increased  the  CP  of  the  plant.  They 
further  suggested  that  breeding  low-tannin 
sericea  with  a high  percentage  of  leaves  would 
lead  to  a greater  improvement  than  breeding  for 
low  tannin  alone. 

The  culmination  of  this  research  was  the  release 
in  1980  of  AU  Lotan  sericea  (Donnelly,  1981). 
This  variety  was  also  selected  for  resistance  to 
root-knot  nematodes.  Donnelly  points  out  that 
most  low-tannin  plants  in  the  breeding  program 
to  develop  AU  Lotan  were  severely  damaged  by  a 
foliar  disease  caused  by  Rhizoctonia  spp.  and 
thus  an  additional  objective  became  selection 
for  resistance  to  this  pathogen.  This  finding 
further  bears  out  the  fact  that  tannins  may  be 
both  beneficial  and  harmful  aspects  of  plants. 
When  Mosjidis  and  Donnelly  (1986)  compared  nine 
lines  low  in  tannin  to  Serela  (a  high  tannin 
cultivar),  they  found  selection  for  low  tannin 
had  resulted  in  slightly  reduced  stem  length, 
leaf  weight  and  leafiness.  Thus  a larger 
portion  of  the  dry  matter  of  low-tannin  lines 
consisted  of  stems.  These  efforts  of  breeding 
for  decreased  tannin  content  in  sericea 
emphasize  that  any  program  of  selection  for 
altered  tannin  content  must  constantly  monitor 
the  effects  on  related  plant  characteristics  to 
avoid  selection  for  correlated  traits  which  are 
undesirable . 


SUMMARY 

Forage  legume  genera  show  quantitative  and 
qualitative  variability  in  tannin  content. 
Several  related  but  different  polyphenolic 
compounds  often  are  lumped  together  under  the 
general  term  tannins  and  this  may  contribute  to 
different  experimental  findings  concerning  the 
detrimental  or  beneficial  properties  of  tannins. 
The  commonly  used  vanillin-HCl  procedure  is  well 
suited  for  breeding  and  selection  research,  but 
does  not  distinguish  between  different  types  of 
tannins.  Other  techniques  which  do  separate 
tannin  types  may  be  too  slow  for  a breeding 
program.  The  protein  precipitation  technique 
requires  further  research. 

Although  some  tannin  containing  species  have 
been  identified  in  Trifolium  and  Medicago , to 
date  increased  tannin  has  not  been  transferred 
to  cultivated  alfalfa  or  any  of  the  cultivated 
clovers  such  as  red,  white,  crimson,  or  sub. 
Genetic  selection  for  decreased  tannin  was 
successful  in  sericea  lespedeza,  but  low-tannin 
lines  were  usually  lower  in  total  dry  matter 
yield  than  the  best  high-tannin  lines.  Tannin 
content  has  generally  been  shown  to  be  a 
moderately  highly  heritable  trait  which  can  be 
altered  dramatically  by  selection.  Selection 
programs  must  account  for  tannin  differences: 
a)  among  plant  parts,  b)  due  to  maturity 
differences,  c)  related  to  environmental  stress, 
and  d)  among  plants.  Such  selection  research 
must  also  continue  to  work  against  the  positive 
correlation  of  yield  and  tannin  content. 
Additional  research  is  needed  to  establish 
threshold  levels  below  which  tannins  have  little 
effect  on  forage  digestibility.  Research  is 
also  needed  to  correlate  levels  of  "tannins"  or 
isolated  specific  polyphenolic  compounds  as 
measured  by  various  procedures  with  changes  in 
animal  digestive  behavior.  The  prospects  of 
improved  animal  performance  related  to  selection 
for  altered  tannin  levels  in  forage  legumes 
appear  to  be  good. 

REFERENCES 

Bates,  R.  P. , and  P.  R.  Henson.  1955.  Studies 
of  inheritance  in  Lespedeza  cuneata  Don. 
Agron.  J.  47:503-507. 

Bate-Smith,  E.  C. , 1973.  Tannins  of  herbaceous 

leguminosae.  Phytochemistry  12:1809-1812. 
Burns,  R.  E.  1963.  Methods  of  tannin  analysis 
for  forage  crop  evaluation.  Ga.  Exp.  Sta. 
Tech.  Bull.  N.  S.  32. 

Cope,  W.  A.  1962.  Heritability  estimates  and 
correlations  of  yield  and  certain 
morphological  and  chemical  components  of 
forage  quality  in  sericea  lespedeza.  Crop 
Sci.  10-12. 

Davis,  A.  M.  1973.  Protein,  crude  fiber, 
tannin,  and  oxalate  concentrations  of  some 
introduced  Astragalus  species.  Agron.  J. 
65:613-615. 

Donnelly,  E.  D.  1954.  Some  factors  that  affect 
palatability  in  sericea  lespedeza,  L. 
cuneata.  Agron.  J.  46:96-97. 


27 


Donnelly,  E.  D,  1959=  The  effect  of  season, 
plant  maturity,  and  height  on  the  tannin 
content  of  sericea  lespedeza,  L.  cuneata. 
Agron.  J.  51:71-73. 

Donnelly,  E.  D.  1981.  Registration  of  AU  Lotan 
sericea  lespedeza.  Crop  Sci.  21:474. 
Donnelly,  E.  D. , and  W.  B.  Anthony.  1970. 

Effect  of  genotype  and  tannin  on  dry  matter 
digestibility  in  sericea  lespedeza.  Crop 
Sci.  10:200-202. 

Donnelly,  E.  D.,  and  W.  B.  Anthony.  1973. 

Relationship  of  sericea  lespedeza  leaf  and 
stem  tannin  to  forage  quality.  Agron  J. 
65:993-994. 

Donnelly,  E.  D. , W.  B.  Anthony,  and  J.  W. 

Langford.  1971.  Nutritive  relationships 
in  low-  and  high-tannin  sericea  lespedeza 
under  grazing.  Agron.  J.  63:749-751. 

Foo,  L.  Y.,  W.  T.  Jones,  L.  J.  Porter,  and  V.  M. 
Williams.  1981.  Proanthocyanidin  polymers 
of  fodder  legumes.  Photochemistry 
21:933-935. 

Hagerman,  A.  E. , and  L.  G.  Butler.  1980. 

Determination  of  protein  in  tannin-protein 
precipitates.  J.  Agric.  Food  Chem. 
28:944-952. 

Haslam,  E.  1966.  Chemistry  of  vegetable 
tannins.  Academic  Press.  London. 

Marshall,  D.  R.,  P.  Broue  and  F.  Munday.  1979. 
Tannins  in  pasture  legumes.  Aust.  J.  Exp. 
Agric.  Anim.  Husb.  19:192-197 
Maxson,  E.  D. , and  L.  W.  Rooney.  1972.  Two 
methods  of  tannin  analysis  for  Sorghum 
bicolor  (L.)  Moench  grain.  Crop  Sci. 
12:253-254. 

McLeod,  M.  N.  1974.  Plant  tannins-their  role 
in  forage  quality.  Nutrition  Abstracts  & 
Reviews  44:803-815. 

Mosjidis,  J.  A.,  and  E.  D.  Donnelly.  1986. 
Vegetative  characteristics  of  selected 
lines  of  Lespedeza  cuneata  low  in  tannins, 
Agron.  Abst.  p.  74. 

Pandey,  K.  K.  1971.  Prospects  of  breeding  non- 
bloating clovers  for  ruminants. 

Proceedings  of  the  Agronomy  Society  of  New 
Zealand  1:111-120 

Quesenberry,  K.  H. , and  K.  A.  Albrecht.  1987. 
Variability  of  tannin  level  in  Desmodium 
and  other  tropical  legumes.  Abst.  Southern 
Branch  ASA.  p.  3 

Rotar,  P.  P.  1965.  Tannins  and  crude  proteins 
of  tick  clovers  (Desmodium  spp.).  Trop. 
Agriculture,  Trin.  42:333-337. 

Sarkar,  S.  K.,  R.  E.  Howarth,  and  B.  P.  Goplen. 
1976.  Condensed  tannins  in  herbaceous 
legumes.  Crop  Sci.  16:543-546. 

Stitt,  R.  E.  1943.  Variation  in  tannin  content 
of  clonal  and  open  pollinated  lines  of 
perennial  lespedeza.  Agron.  J.  35:944-954 
Walton,  M.  F. , F.  A.  Haskins,  and  H.  J.  Gorz. 
1983.  False  positive  results  in  the 
vanillin-HCI  assay  of  tannins  in  sorghum 
forage.  Crop  Sci.  23:197-200. 


28 


HISTORY  OF  COOL  SEASON  GRASS  BREEDING  IN  THE 
SOUTHEAST 


J.  F.  Pedersen—' 


Plant  breeding  is  a process  of  plant  improvement 
involving  active  human  involvement.  As  such, 
the  history  of  cool  season  grass  breeding  in  the 
southeast  region  is  a history  of  scientists  and 
their  contributions  to  pasture  and  forage  crop 
improvement.  This  review  will  highlight  the 
contributions  of  many  of  these  scientists.  I 
must  confess  that  I worked  under  a considerable 
handicap  while  researching  this  history.  I was 
born  in  1954.  My  first  exposure  to  most  of  the 
individuals  and  crops  discussed  in  this  review 
was  in  the  1980 's.  I am  forced  to  rely  on  the 
literature  and  on  the  memories  of  many  who 
contributed  far  more  to  this  history  than  I.  To 
all  those  who  responded  to  my  telephone  requests 
for  materials  and  information,  many  thanks. 

Relatively  few  cool  season  forage  grass  species 
have  been  researched  and  exploited  successfully 
in  the  southeast  region.  These  are  limited  to 
tall  fescue  (Festuca  arundinacea  Schreb.)  annual 
ryegrass  (Lolium  mult if lorum  Lam.),  orchardgrass 
(Dactylis  glomerata  L.),  timothy  (Phleum 
pratense  L.),  Kentucky  bluegrass  (Poa  pratensis 
L.),  and  Phalaris  aquaticus  L.  Cereal  crops 
utilized  as  winter  forage  contribute 
significantly  to  winter  grazing  in  the  southeast 
region,  but  most  were  developed  for  grain 
production  and  will  not  be  discussed. 

I would  like  to  digress  for  a moment  and  discuss 
breeding  methodology.  Plant  breeding  has  two 
basic  elements,  identifying  superior  gene 
combinations  and  capturing  superior  gene 
combinations.  The  discovery,  increase,  and 
marketing  of  superior  populations  as  cultivars 
certainly  is  a valid  breeding  method,  and  will 
be  a recurring  theme  in  this  review.  Such 
methodology  does  not  always  appear  to  be  "plant 
breeding"  to  the  lay  public,  and  even  to  some 
agronomists.  After  all,  no  hand  crosses  are 
made.  However,  this  methodology  was,  and 
continues  to  be  a valid  breeding  approach  in 
crops  or  regions  where  little  genetic 
improvement  has  been  achieved. 

Although  tall  fescue  was  introduced  to  the 
region  sometime  prior  to  1900,  the  recorded 
history  of  tall  fescue  begins  in  1931  with  the 
discovery  of  'Kentucky  31'  by  E.  N.  Fergus 
(Fergus  and  Buckner,  1972).  It  is  an  adapted 
ecotype  that  probably  makes  up  the  bulk  of  the 
35  million  reported  acres  (Buckner  et  al.,1979) 
grown  in  the  United  States  to  this  day. 


— ^Tobacco  and  Forage  Research  Unit,  USDA-ARS, 
Univ.  of  Kentucky,  Lexington,  KY  40540-0091 


Tall  fescue  was  productive,  persistent,  and 
widely  adapted,  but  cattle  performance  was 
mysteriously  poor  on  the  species.  The  USDA-ARS 
program  headed  by  R.  C.  Buckner  at  the 
University  of  Kentucky  began  addressing  this 
problem  in  the  late  1950 's  by  developing  inbred 
lines  and  evaluating  them  for  palatability  to 
cattle  in  cafeteria-style  comparisons. 

'Kenwell',  the  end  product  of  this  program,  was 
released  in  1965  and  was  shown  to  be 
significantly  more  palatable  than  Kentucky  31 
(Buckner  and  Burrus,  1968),  but  was  not  shown  to 
offer  any  advantage  over  Kentucky  31  in 
subsequent  grazing  trials  (Buckner,  1973). 

Buckner  also  initiated  a program  involving 
crosses  of  tall  fescue  with  annual  ryegrass, 
perennial  ryegrass  (Lolium  per e tie  L.),  and  giant 
fescue  (Festuca  gigantea  (L.)  Vill.)  in  the 
early  1960's  to  improve  the  forage  quality  of 
tall  fescue.  'Kenhy' , released  in  1976,  was  the 
first  cultivar  developed  from  this  endeavor,  and 
does  exhibit  improved  quality  when  graded  by 
laboratory  measures  (Buckner  et  al.,  1977). 
However,  the  animal  problems  were  not  yet 
solved.  'Johnstone',  released  in  1981  (Buckner 
et  al.,  1983)  is  another  cultivar  originating 
from  this  program.  Johnstone  tall  fescue  may 
give  superior  animal  performance,  but  its 
enhanced  quality  was  not  the  primary  cause  of 
the  observed  increase  in  animal  performance. 

The  real  breakthrough  in  tall  fescue  improvement 
was  made  in  the  early  1970's,  and  not  by  plant 
breeders.  The  discovery  of  Acreinonium 
coenophialum  Morgan-Jones  and  Gams  (previously 
identified  as  Epichloe  typhina  (Fr.)  Tul.)  in  a 
tall  fescue  pasture  with  a history  of  poor 
cattle  performance  by  C.  W.  Bacon  (Bacon  et  al., 
1977),  and  the  subsequent  comparison  of  cattle 
performance  on  Aj_  coenophialum-inf ected  and  A._ 
coenophialum-f ree  tall  fescue  pastures  by  C.  S. 
Hoveland  (Hoveland  et  al.,  1980),  identified  the 
fungus  as  the  actual  cause  of  poor  cattle 
performance  on  tall  fescue.  Breeders  now  had 
another  method  for  tall  fescue  improvement: 
elimination  of  this  endophyte.  Breeding 
objectives  for  improvements  in  characters  other 
than  forage  quality  also  assumed  new  importance. 

An  Auburn  University  program  initiated  in  the 
mid  1970's  by  R.  L.  Haaland,  and  continued  in 
the  1980's  by  J.  F.  Pedersen  had  the  objective 
of  increasing  winter  productivity  in  tall 
fescue.  Beginning  with  Mediterranean  plant 
introduction  materials,  this  program  resulted  in 
the  release  of  'AU  Triumph' , which  was  80  % more 
winter  productive  than  Kentucky  31  (Pedersen  et 
al.,  1983).  As  with  Johnstone,  it  was  released 
as  a low  A.  coenophialum  cultivar,  although  a 
low  endophyte  infection  level  was  not  an  initial 
objective  in  the  development  of  either  of  these 
popular  cultivars.  These  two  cultivars  were  the 
first  of  what  has  come  to  be  known  by  some  as 
the  second  generation  tall  fescues,  or  low 
endophyte  tall  fescues. 


29 


Continued  selection  for  winter  productivity  by 
Pedersen  has  resulted  in  lines  with  winter 
productivity  twofold  that  of  AU  Triumph,  but 
which  are  quite  susceptible  to  frost  damage 
(unpublished  data). 

Other  "new"  breeding  objectives  in  tall  fescue 
focus  largely  on  increasing  its  persistence  and 
productivity  in  marginal  areas  of  adaptation. 
This  includes  screening  plant  introduction 
accessions  for  adaptation  to  the  southeast 
region  by  J.  H.  Bouton,  resulting  in  the  release 
of  GaFesl  and  GaFes2  tall  fescue  germplasms 
(Bouton  and  Powell,  1982),  collection  of 
ecotypes  from  south  Georgia  by  Bouton  and  his 
work  investigating  incidence  of  rhizomes  in  tall 
fescue  (D'Urva  et  al.,  1983;  and  Jernstedt  and 
Bouton,  1985).  Collection  and  testing  of 
superior  ecotypes  continues  in  Mississippi  by  C. 
E.  Watson,  while  screening  of  plant  introduction 
accessions  and/or  other  populations  proceeds  in 
south  Alabama  (J.  F.  Pedersen)  and  central 
Florida  (D.  D.  Baltensperger ) . Cultivar  release 
from  one  or  more  of  these  programs  appears 
imminent . 

The  most  recent  challenges  for  the  tall  fescue 
breeder  are  associated  with  possible  losses  in 
performance  due  to  the  elimination  of  A. 
coenophialum.  Insect  resistance  due  to  A. 
coenophialum  infection  has  been  demonstrated  in 
the  laboratory  (Siegel  et  al.,  1987)  and  there 
has  been  one  report  of  increased  disease 
susceptibility  in  A.  coenophialum- free  tall 
fescue  in  the  field  (Bush  and  Burrus,  1988). 
Perhaps  earlier  work  by  C.  D.  Berry  (1973) 
studying  rust  resistance  will  have  renewed 
importance.  A.  coenophialum-f ree  cultivars  are 
perceived  as  having  poorer  seedling  vigor  by 
farmers,  although  this  has  not  been  documented 
by  researchers.  Susceptibility  to  plant 
parasitic  nematodes  has  been  shown  to  be  higher 
in  A.  coenophialum  free  tall  fescue  (Pedersen  et 
al.,  1988).  Indications  of  physiological 
advantages  in  A^_  coenophialum- infected  tall 
fescue  under  drought  conditions  are  beginning 
to  appear  in  the  literature  (Belesky  et  al., 
1987)  Finally,  J.  C.  Read  and  B.  J.  Camp  (1986) 
have  documented  lack  of  stand  survival  in  A. 
coenophialum-f ree  tall  fescue  pastures,  compared 
to  A.  coenophialum-inf ected  tall  fescue  pastures 
in  a Texas  grazing  study  in  a marginal 
environment . 

Certainly  tall  fescue  breeders  in  the  southeast 
region  are  faced  with  more  challenges  now  than 
before  the  A.  coenophialum- tall  fescue 
relationship  was  discovered.  They  also  are 
privilege  to  more  fundamental  information 
regarding  their  crop  than  ever  before.  When  the 
southeast  region  is  considered  as  a whole, 
today's  tall  fescue  breeding  team  includes 
breeders  (Bouton,  Pedersen,  Rice,  Van  Santen, 
Watson,  Wofford),  a cytologist  (Eizenga),  a 
tissue  culturist  (Conger),  and  ready  access  to 
physiologists,  pasture  management  scientists, 
animal  scientists,  pathologists,  etc.  throughout 
the  area.  With  continued  close  cooperation  and 
communication,  our  new  challenges  in  tall  fescue 


improvement  should  be  met . 

Next  to  tall  fescue,  annual  ryegrass  is  probably 
the  most  important  cool  season  species  utilized 
in  the  southeast  region.  The  first  registered 
cultivar  release  was  'Gulf',  from  R.  M. 

Weihing's  program  in  1958  (Weihing,  1963).  Gulf 
was  selected  for  early  maturity  and  rust 
resistance  from  a Uruguay  plant  introduction 
accession.  It  is  still  widely  used  today,  and 
is  considered  the  standard  by  which  to  judge  all 
other  annual  ryegrasses  in  this  region.  H.  W. 
Bennett  and  H.  W.  Johnson  released  another  rust 
resistant  cultivar,  'Magnolia',  in  1965  (Bennett 
and  Johnson,  1968).  It,  however,  has  not  seen 
widespread  use.  The  most  recent  rust  resistent 
annual  ryegrass  release  is  'Florida  80'.  It  was 
selected  from  volunteer  plants  of  several  older 
cultivars  and  germplasms  that  had  reseeded  for 
two  or  more  years  in  pastures  in  Florida  and 
Georgia  (Prine  et  al.,  1986).  The  search  for 
rust  resistance  in  annual  ryegrass  was 
documented  as  early  as  1956  by  E.  C.  Holt  (1956) 
in  Texas  and  H.  D.  Wells  (1956)  in  Georgia,  and 
is  not  over.  Active  programs  headed  by  G.  M. 
Prine  in  Florida,  C.  E.  Watson  in  Mississippi 
and  L.  R.  Nelson  in  Texas  are  continuing  to 
pursue  this  breeding  objective. 

One  other  cultivar,  'Marshall',  released  in 
Mississippi  in  1980  (Arnold,  et  al.,  1981) 
merits  special  mention.  It  is  a cold-hardy 
annual  ryegrass  that  was  "the  result  of  29  years 
of  natural  selection  from  common  ryegrass  as  a 
reseeding  stand  under  grazing  conditions"  in 
north  Mississippi.  It  could  justly  be  called  an 
adapted  ecotype.  It  is  also  a very  successful 
cultivar,  already  widely  utilized  in  the  region. 

The  other  cool  season  grass  species  are  not  as 
broadly  adapted  to  the  entire  region,  and  have 
not  received  as  much  research  emphasis. 
Orchardgrass  improvement  for  the  South  was  the 
topic  of  a 1954  presentation  to  the  Southern 
Pasture  and  Forage  Improvement  Conference 
( SPFCIC)  by  T.  J.  Smith  (1954).  More  recently, 
orchardgrass  breeding  was  continued  in  Virginia 
by  L.  Taylor,  resulting  in  the  release  of 
'Jackson'  and  'Virginia  80'.  R.  C.  Buckner 
collected  "naturalized  strains"  of  orchardgrass 
from  fields  across  Kentucky,  and  selected  a 
broad  based  population  from  these  strains  that 
was  released  as  'Boone'  orchardgrass  (Buckner, 
1963).  McClain,  in  South  Carolina,  released 
'Piedmont'  orchardgrass  in  1978.  It  is  a four 
clone  synthetic  exhibiting  high  yield,  late 
maturity  and  rust  resistance  (McClain,  1986). 

'Clair'  timothy,  and  'Kenblue'  Kentucky 
bluegrass,  are  the  only  registered  cultivars  of 
these  two  species  developed  in  the  southeastern 
region.  Clair  is  an  adapted  ecotype  collected 
in  Indiana  and  released  in  1962  ( Buckner , 1962 ) . 
It  continues  to  be  a popular  variety  as 
evidenced  by  continuing  requests  for  foundation 
class  seed.  Kenblue  Kentucky  bluegrass  is  a 
blend  of  seed  from  old  (circa  1955)  seedfields 
located  in  central  Kentucky  (Buckner,  1968).  A 
Kentucky  bluegrass  population  probably  dating  to 


30 


the  early  1800's  was  collected  in  Virginia  by  L. 
Taylor  and  T.  H.  Taylor  and  released  as 
’Piedmont1  Kentucky  bluegrass,  but  seed  has  not 
been  successfully  increased. 

One  other  cool  season  grass  species,  Phalaris 
aquaticus  L.,  has  seen  breeding  effort  in  the 
southeastern  region.  ’Evergreen* , developed  by 

E.  C.  Holt  in  Texas,  and  'AU  Oasis',  developed 
in  Alabama  in  program  initiated  by  C.  D.  Berry, 
continued  by  R.  L.  Haaland,  and  completed  by  J. 

F.  Pedersen  (Pedersen,  et  al.,  1983)  have  merit 
as  improved  forages,  but  have  not  been  utilized 
by  the  forage  industry  to  date.  Seed  production 
problems  have  plagued  the  AU  Oasis  marketing 
effort,  reminding  us  that  if  our  product  is  to 
be  of  worth  to  farmers,  we  must  be  able  to 
produce  seed. 

A history  of  cool  season  grass  breeding  in  the 
southeastern  region  would  not  be  complete 
without  mention  of  the  concept  of  exclusive 
marketing  of  publicly  developed  cultivars  by 
private  industry.  The  general  argument 
supporting  such  a relationship  centers  on  the 
need  for  expert  seed  production  and  marketing  of 
forage  crops,  combined  with  relatively  low  seed 
volume  and  the  need  for  profit  by  the  seedsman. 
This  relationship  was  discussed  in  depth  at  an 
earlier  SPFCIC  meeting  (Campbell,  1985;  Eisner, 
1985;  Hanna,  1985;  Nelson,  1985;  Pedersen, 

1985),  and  will  not  be  discussed  further  here. 

In  closing,  I must  state  that  although  I have 
confined  my  discussion  to  breeders  of  the 
southeast  region,  public  breeders,  private 
breeders,  and  private  enterprise  outside  this 
region  have  had  considerable  direct  impact  on 
our  forage  history.  This  history  is  in  no  way 
intended  to  detract  from  their  contributions  of 
cultivars,  technology,  seed  production,  and 
marketing  expertise,  or  the  outstanding 
cooperation  in  breeding  efforts  throughout  the 
entire  forage  industry. 


REFERENCES 

ARNOLD,  B.  L.,  C.  E.  Watson,  Jr.,  and  N.  C. 
Edwards,  Jr.  1981.  Registration  of  Marshall 
annual  ryegrass.  Crop  Sci.  21:474-475. 

BACON,  C.  W. , J.  K.  Porter,  J.  D.  Robins,  and  E. 
S.  Luttrell.  1977.  Epichloe  typhina  from  toxic 
tall  fescue  grasses.  Appl.  Env.  Micro. 
34:576-581. 

BELESKY,  D.  P.,  0.  J.  Devine,  J.  E.  Pallas,  Jr., 
and  W.  C.  Stringer.  1987.  Photosynthetic 
activity  of  tall  fescue  as  influenced  by  a 
fungal  endophyte.  Photosynthetica . 21:82-87. 

BENNETT,  H.  W.  and  H.  W.  Johnson.  1968. 
Registration  of  Magnolia  annual  ryegrass.  Crop 
Sci  8:401. 


BERRY,  C.  D.  1973.  Breeding  for  rust 
resistance  in  tall  fescue,  p.  119-120.  Ln  Proc. 
30th.  South.  Pasture  and  Forage  Crop  Improvement 
Conf.  29-31  May  1973.  Lexington,  Kentucky 

BOUTON,  J.  H.  and  J.  D.  Powell.  1982. 
Registration  of  GaFesl  and  GaFes2  tall  fescue 
germplasms.  Crop  Sci.  22:450. 

BUCKNER,  R.  C.  1962.  CLAIR  (Reg.  No.  3).  Crop 
Sci.  2:355. 

BUCKNER,  R.  C.  1963  Registration  of  Boone 
orchardgrass . Crop  Sci.  3:304. 

BUCKNER,  R.  C.  1968.  Kenblue  Kentucky 
bluegrass.  Univ.  of  Kentucky  Agri.  Exp.  Sta . , 
Dept,  of  Agron.  Leaflet  308. 

BUCKNER,  R.  C.  1973.  The  tall  fescue  breeding 
program  at  the  University  of  Kentucky.  p. 

85-88.  _In  Proc.  30th.  South.  Pasture  and  Forage 
Crop  Improvement  Conf.  29-31  May  1973. 

Lexington,  Kentucky 

BUCKNER,  R.  C. , J.  A.  Boling,  P.  B.  Burrus,  II., 
L.  P.  Bush,  and  R.  A.  Hemken.  1983. 

Registration  of  Johnstone  tall  fescue.  Crop 
Sci.  23:399-400. 

BUCKNER,  R.  C.  and  P.  B.  Burrus,  II.  1968. 
Registration  of  Kenwell  tall  fescue.  Crop  Sci. 
8:398. 

BUCKNER,  R.  C.,  P.  B.  Burrus,  II.,  and  L.  P. 
Bush.  1977.  Registration  of  Kenhy  tall 
fescue.  Crop  Sci.  17:672. 

BUCKNER,  R.  C.  and  L.  P.  Bush.  1979.  Preface. 
In  R.  C.  Buckner  and  L.  P.  Bush  (ed.)  Tall 
Fescue.  Agronomy  20:xiii-xiv. 

BUSH,  L.  P.  and  P.  B.  Burrus,  Jr.  1988.  Tall 
fescue  forage  quality  and  agronomic  performance 
as  affected  by  the  endophyte.  J.  Prod.  Agric. 
1:55-60. 

CAMPBELL,  Foy.  1985.  Exclusive  variety 
releases,  p.  45-48.  In  Proc.  41st  South.  Pasture 
and  Forage  Crop  Improvement  Conf.  20-22  May, 
Raleigh,  North  Carolina. 

D'UVA,  P.,  J.  H.  Bouton,  and  R.  H.  Brown. 

1983.  Variability  in  rooted  stem  production 
among  tall  fescue  genotypes.  Crop  Sci. 
23:385-386. 

ELSNER,  J.  Earl.  1985.  The  role  of  state  and 
foundation  seed  associations  in  the  release  of 
improved  lines  from  public  institutions,  p. 
53-54.  In  Proc.  41st  South.  Pasture  and  Forage 
Crop  Improvement  Conf.  20-22  May,  Raleigh, 

North  Carolina. 

FERGUS,  E.  N.  and  R.  C.  Buckner.  1972. 
Registration  of  Kentucky  31  tall  fescue.  Crop 
Sci.  12:714. 


31 


HANNA,  Wayne  W.  1985.  Review  of  ARS  policies 
for  release  of  improved  varieties  and 
germplasm.  p.  51-52.  In  Proc.  41st  South. 
Pasture  and  Forage  Crop  Improvement  Conf . 20-22 

May,  Raleigh,  North  Carolina. 

HOLT,  E.  C.  1956.  Breeding  for  resistance  to 
rust:  ryegrass,  p.  24.  In  Proc.  13th.  South. 
Pasture  and  Forage  Crop  Improvement  Conf.  15-17 
May  1956.  Experiment,  GA. 

HOVELAND,  C.  S.,  R.  L.  Haaland,  C.  C.  King,  Jr., 
W.  B.  Anthony,  E.  M.  Clark,  J.  A.  McGuire,  L.  A. 
Smith,  H.  W.  Grimes,  and  J.  L.  Holliman.  1980. 
Association  of  Epichloe  typhina  fungus  and  steer 
performance  on  tall  fescue  pasture.  Agron.  J. 
72:1064-1065. 

JERNSTEDT,  J.  A.  and  J.  H.  Bouton.  1985. 
Anatomy,  morphology,  and  growth  of  tall  fescue 
rhizomes.  Crop  Sci.  25:539-542. 

MCCLAIN,  E.  F.  1986.  Registration  of 
'Piedmont'  orchardgrass . Crop  Sci.  26:835-836. 

NELSON,  L.  R.  1985.  Breeders  rights  and 
compensation  for  plant  breeders  from  public 
institutions,  p.  49-50.  In  Proc.  41st  South. 
Pasture  and  Forage  Crop  Improvement  Conf.  20-22 
May,  Raleigh,  North  Carolina. 

PEDERSEN,  J.  F.  1985.  Forage  crop  variety 
release  and  agricultural  experiment  station 
policies,  p.  42-44.  In  Proc.  41st  South. 

Pasture  and  Forage  Crop  Improvement  Conf.  20-22 
May,  Raleigh,  North  Carolina. 

PEDERSEN,  J.  F.,  R.  L.  Haaland,  C.  S.  Hoveland, 
C.  D.  Berry,  S.  P.  Schmidt,  and  R.  R.  Harris 
1983.  Registration  of  AU  Triumph  tall  fescue. 
Crop  Sci.  23:182. 

PEDERSEN,  J.  F.,  C.  S.  Hoveland,  R.  L.  Haaland, 
and  C.  D.  Berry.  1983.  Registration  of  AU 
Oasis  phalaris.  Crop  Sci.  23:597. 

PEDERSEN,  J.  F.,  R.  Rodriguez-Kabana , and  R.  A. 
Shelby.  1988.  Ryegrass  cultivars  and  endophyte 
in  tall  fescue  affect  nematodes  in  the  grass  and 
succeeding  soybeans.  Agron.  J.  80:(in  press). 

RINE,  G.  M.,  L.  S.  Dunavin,  Paul  Mislevy,  K.  J. 
McVeigh,  and  R.  L.  Stanley.  1986.  Registration 
of  'Florida  80'  annual  ryegrass.  Crop  Sci. 
26:1083-1084. 

READ,  J.  C.  and  B.  J.  Camp.  1986.  The  effect 
of  the  fungal  endophyte  Acremonium  coenophialum 
in  tall  fescue  on  animal  performance,  toxicity, 
and  stand  maintenance.  Agron.  J.  78:848-850. 

SIEGEL,  M.  R.,  G.  C.  M.  Latch,  and  M.  C. 

Johnson.  1987.  Fungal  endophytes  of  grasses. 
Ann.  Rev.  Phytopathol.  25:293-315. 

SMITH,  T.  J.  1954.  Breeding  orchardgrass  for 
the  South,  p.  13-14.  Tn  Proc.  12th.  South. 
Pasture  and  Forage  Crop  Improvement  Conf.  8-11 
June  1954.  Stillwater,  Oklahoma. 


WEIHING , R.  M.  1963.  Registration  of  Gulf 
annual  ryegrass.  Crop  Sci.  3:366. 

WELLS,  H.  D.  1956.  Breeding  annual  ryegrass 
for  resistance  to  crown  rust.  p.  57.  _In  Proc. 
13th.  South.  Pasture  and  Forage  Crop  Improvement 
Conf.  15-17  May  1956.  Experiment,  GA. 


32 


FORAGE  GERMPLASM  EVALUATION 
M.A,  Hussey  and  D.I.  Bransby  - 


INTRODUCTION 

The  ultimate  goal  of  a pasture  improvement 
program  is  to  provide  "improved"  forage 
cultivars  to  farmers  and  ranchers.  While  the 
final  evaluation  of  a new  species  or  cultivar 
must  be  in  an  environment  similar  to  where  it 
will  be  utilized,  it  is  not  possible  to 
evaluate  all  hybrids  under  every  possible 
combination  of  fertility,  soils,  and 
management.  To  facilitate  the  selection  of 
"superior"  germplasm,  uniform  evaluation 
procedures  are  utilized. 

According  to  Jones  and  Walker  (1983)  "there 
have  been  no  major  new  breakthroughs  in  the 
accepted  theoretical  procedures  for  the 
evaluation  of  pasture  plants."  They 
continued  by  stating  that  "except  under 
circumstances  which  require  special 
strategies,  there  are  no  clear  shortcuts  in 
establishing  the  suitability  of  plants  for 
the  soil -plant-animal  complex  encountered  in 
the  field."  Therefore,  due  to  the 

complexity  of  forage  crop  production  systems, 
a detailed  multi-location,  multi-year,  multi- 
disciplinary evaluation  procedure  is  utilized 
with  a grazing  evaluation  considered 
essential  prior  to  the  final  release  of  a 
cultivar  (Hoveland,  1979;  Mochrie  et  al . , 
1981). 

Specific  methodology  for  the  evaluation  of 
forage  crops  has  been  covered  in  detail  by 
speakers  at  past  meetings  of  the  Southern 
Pasture  and  Forage  Crops  Improvement 
Conference  (SPFCIC)  Breeders  Workgroup 
(Quesenberry  et  al . 1977;  Burton  and  Monson, 
1979;  Coleman  1979;  Haaland,  1979;  Hoveland, 
1979;  Quesenberry,  1980;  Burton,  1982;  Riewe, 
1982;  Lippke,  1983;  Sleper  et  al . , 1983)  as 
well  as  in  several  excellent  reviews  (Mochrie 
et  al.,  1981;  Jones  and  Walker,  1983;  Mott 
and  Moore,  1985).  An  overview  of  such  an 
evaluation  plan  from  Jones  and  Walker  (1983) 
is  presented  in  Table  1.  We  do  not  wish  to 
discuss  forage  evaluation  methodology  in 
detail,  but  rather  point  to  selected 
evaluation  techniques  which  may  facilitate 
the  overall  evaluation  of  forage  germplasm. 


SPACED  PLANT  NURSERY 

The  spaced  plant  nursery  is  utilized  to 
determine  the  overall  adaptation  of  a new 
species  or  hybrid.  The  initial 
identification  of  a superior  genotype  occurs 
in  such  a nursery,  so  techniques  are  required 


1 Assistant  Professor,  Texas  A&M  University, 
and  Professor,  Auburn  University. 


that  allow  the  breeder  to  rapidly  and 
accurately  evaluate  large  numbers  of 
plants. 


FORAGE  YIELD 

Forage  yield,  particularly  when  averaged  over 
several  locations,  or  in  response  to  a 
management  stress  such  as  clipping  is  a 
common  method  for  determining  the  adaptation 
of  a genotype.  In  the  initial  stages  (spaced 
plant  nursery)  of  evaluation,  yield  estimates 
are  generally  made  by  a visual  appraisal  of 
vegetative  vigor.  Yield  estimates  from  small 
plots  generally  are  obtained  by  harvesting 
small  quadrats  or  specially  designed  forage 
plot  harvesters  (Frame,  1981). 

The  use  of  doubl e- sampl i ng  techniques 
represents  one  method  of  predicting  forage 
yield  which  may  reduce  the  time  required  to 
quantify  herbage  mass.  Peterson  (1988) 
compared  the  use  of  plant  height,  the  disk 
meter,  and  the  single  probe  capacitance  meter 
for  estimating  forage  yield.  Results  from 
this  study  indicated  that  all  techniques  were 
effective  in  estimating  forage  yield  in  small 
plots  but  that  plant  height  and  the  disk 
meter  were  influenced  by  season  within  a year 
therefore,  requiring  separate  calibration 
equations  to  be  developed.  Peterson  and 
Hussey  (1987)  have  also  shown  significant 
cultivar  effects  for  plant  height-yield  and 
disk  meter-yield  relationships,  suggesting 
that  caution  may  be  required  when  using 
double-sampling  techniques  for  the  evaluation 
of  genotypes  with  a diverse  range  of  growth 
habits. 


LEAFINESS 

Since  livestock  have  the  ability  to  select  a 
large  percentage  of  leaf  in  their  diets 
(Laredo  and  Minson,  1973),  and  because  leaf 
blades  generally  are  of  higher  nutritive 
value  than  the  stem  component,  increasing  the 
relative  leaf  to  stem  ratio  of  a cultivar 
should  result  in  greater  animal  gains.  To 
date,  few  attempts  have  been  made  to  select 
for  improved  leaf  to  stem  ratio  although, 
dwarf  genotypes  of  pearl  millet  (Burton, 
1982),  have  been  shown  to  result  in  enhanced 
animal  performance  when  compared  to  tall 
cultivars,  even  though  they  produce  less 
forage. 

Most  studies  which  have  attempted  to 
determine  variation  in  the  relative  leaf 
content  of  forage  have  utilized  either  visual 
estimates  of  leafiness  or  tedious  hand 
separations  of  leaf  and  stem.  Holt  (1963)  in 
one  of  the  few  studies  which  compared  visual 
estimates  of  leaf  with  hand  separated 
estimates  reported  that  visual  estimates  of 
leafiness  were  more  closely  related  to  tiller 
density  than  to  leaf  content  of  the  forage. 


33 


making  visual  assessments  of  leaf  content  of 
relatively  little  value. 

Recently,  the  use  of  near  infrared 
reflectance  spectroscopy  (NIRS)  has  been 
investigated  for  its  potential  of  estimating 
leaf  content  of  forage  (Hill  et  al . , 1988). 
They  concluded  that  NIRS  was  an  effective 
tool  for  estimating  the  leaf  content  of 
alfalfa  in  both  plots  and  in  the  diets  of 
livestock.  Studies  with  bermudagrass 
(Peterson,  1988)  have  also  confirmed  the 
potential  of  NIRS  for  predicting  leaf  content 
of  warm-season  grasses. 


FORAGE  QUALITY 

The  ability  to  rapidly  evaluate  forage 
digestibility  i_n  vi  tro  (IVDMD),  is  an 
extremely  important  analytical  tool  in  plant 
breeding.  Most  reviews  on  forage  evaluation 
techniques,  consider  an  estimate  of  forage 
digestibility  an  essential  step  to  the 
evaluation  of  spaced  plant  nurseries. 
Selection  for  greater  IVDMD  has  resulted  in 
the  release  of  several  bermudagrass  cultivars 
with  improved  digestibility  and  enhanced 
animal  performance  (Holt  et  al  . , 1983, 

Eichhorn  et  al . , 1986).  In  the  past  10 
years,  NIRS  has  shown  to  accurately  predict 
forage  IVDMD,  CP,  etc.  when  properly 
calibrated  (Barton  and  Burdick,  1981,  Jones 
et  al . , 1987).  The  major  limitation  to  the 
use  of  NIRS  technology  appears  to  be  the  high 
initial  cost  of  the  instrumentation  and  the 
lack  of  adequate  calibration  sets  for  warm- 
season  forages. 

The  use  of  high  activity  fungal  cellulase 
enzymes  is  another  technique  which  may  be 
utilized  to  estimate  forage  nutritive  value. 
While  commercial  cellulase  enzymes  were 
originally  tried  in  the  early  1960's  as  a 
replacement  for  rumen  fluid,  it  was  not  until 
the  early  1970's  that  enzymes  were  identified 
which  could  adequately  replace  rumen  fluid 
(Jones  and  Hayward,  1973).  Several 
commercial  sources  of  cellulase  have  been 
evaluated  (Gabrielson,  1986).  A comparison 
of  NOVO  celluclast  and  a standard  in  vitro 
assay  indicated  that  both  techniques  were 
equally  effective  for  ranking  warm-season 
perennial  grasses  (Stair  et  al.,  1987). 


SMALL  PLOT  EVALUATIONS 

Small  plots,  evaluated  at  multiple  locations 
are  essential  to  determining  the  adaptation 
of  germplasm.  For  many  species,  small  plot 
evaluations  are  conducted  for  2-3  years  in 
which  multiple  clipping  heights  and/or 
frequencies  are  utilized.  Past  experience 
with  bermudagrass  small  plot  evaluations  has 
indicated  that  a minimum  of  4 years 
(including  the  establishment  year)  are 
required  at  each  location  to  obtain  an 


accurate  reading  of  cultivar  performance 
(Conrad,  Holt,  and  Taliaferro,  personnel 
communication) . 

Multiple  clipping  heights  and/or  frequencies 
are  often  imposed  on  small  plot  evaluations 
to  identify  superior  germplasm.  Such  tests 
may  be  quite  effective  in  determining  the 
relative  persistence  of  diverse  germplasm 
(Jones,  1974),  or  may  contribute  relatively 
little  information,  depending  on  the  growth 
form  of  the  species  under  evaluation.  For 
instance,  multiple  clipping  frequency- 
clipping height  experiments  conducted  with  a 
flail  type  mower,  failed  to  give  differences 
in  relative  persistence  for  bermudagrass 
(Holt,  personnel  communication),  while  tiller 
density  was  significantly  reduced  when 
harvested  to  ground  level  with  electric 
clippers.  These  results  suggest  that  while 
differential  cutting  heights  may  separate 
germplasm,  certain  species  may  require  more 
intensive  defoliation  than  can  be  obtained 
with  standard  forage  harvesting  equipment. 

Many  new  forage  cultivars  fail  because  of 
their  poor  adaptation.  To  prevent  this, 
proper  selection  of  test  sites  is  required, 
lo  properly  select  test  sites,  an 
agroclimatic  approach  similar  to  that 
proposed  by  Nix  (1982)  in  Australia  may  be 
utilized.  Such  a classification  system, 
groups  similar  climatic  zones  based  on 
differences  in  a number  of  climatic 
variables,  particularly  rainfall,  ET, 
temperature,  etc.  In  the  United  States,  many 
evaluation  locations  are  determined  by  the 
presence  of  a research  location.  We  have 
found,  that  cooperative  efforts  involving  the 
USDA-ARS,  USDA-SCS,  seed  companies,  and 
producers  has  greatly  expanded  our  ability  to 
evaluate  germplasm  for  wide  adaptation. 


GRAZING  EVALUATIONS 

The  final  evaluation  of  forage  cultivars  must 
be  conducted  using  the  grazing  animal  if  the 
the  cultivar  is  to  be  utilized  for  pasture. 
While  clipping  can  determine  the  ability  of  a 
plant  to  withstand  defoliation  stress,  it 
does  not  measure  other  effects  associated 
with  the  grazing  animal  (treading,  pulling, 
urine  and  dung,  etc.)  (Watkin  and  Clements, 
1978).  Most  reviews  on  the  evaluation  of 
forage  germplasm  suggest  that  genotypes 
should  be  grazed  as  soon  as  possible  (ie. 
spaced  plants  or  small  plots),  however,  due 
to  logistical  and  financial  constraints, 
grazing  evaluations  are  generally  conducted 
just  prior  to  the  release  of  a cultivar.  In 
certain  species,  it  may  be  possible  to  obtain 
a good  estimate  of  persistence  through  the 
use  of  properly  designed  clipping  trials 
(Jones  and  Walker,  1983),  however,  in  grazing 
sensitive  species  such  as  alfalfa,  grazing 
and  clipping  may  give  entirely  different 
results.  In  crops  such  as  this,  direct 


34 


selection  under  grazing  may  be  the  best 
method  for  developing  germplasm  with 
tolerance  to  herbivory  (Bouton,  personnel 
communication,  Bouton,  1988). 

It  is  strongly  recommended  that  the  final 
evaluations  of  germplasm  should  be  evaluated 
over  a range  of  stocking  rates  or  levels  of 

available  forage.  Such  an  evaluation  will 
provide  useful  animal  data  over  a range  of 
possible  management  alternatives,  as  well  as 
determining  the  persistence  of  the  cultivar. 
A single  set  stocking  rate  comparison  of 
cultivars  may  lead  to  erroneous  conclusions 
concerning  the  potential  of  a cultivar,  since 
"optimal"  stocking  rates  for  maximal  ADG's, 
do  not  always  give  maximal  economic  returns. 


SUMMARY 

The  evaluation  of  forage  germplasm  prior  to 
the  release  of  a cultivar  is  a laborious  and 
time  consuming  process.  Improved  efficiency 
can  be  obtained  by  the  use  of  innovative 
plant  breeding  methods,  such  as  restricted 
recurrent  phenotypic  selection,  or  by  the 
development  of  new  methodology  to  facilitate 
germplasm  evaluation  during  each  cycle  of 
selection.  Because  quantitative  characters 
such  as  forage  IVDMD,  persistence, 
winterhardiness,  etc.  are  extremely  important 
in  determining  the  ultimate  success  of  a 
cultivar,  evaluations  must  be  carried  out 
over  a range  of  environmental  conditions.  As 
has  been  pointed  out  by  Jones  and  Walker 
(1983)  there  are  not  shortcuts  in  breeding 
forage  cultivars,  however,  the  adaptation  of 
techniques  to  facilitate  germplasm  evaluation 
should  improve  the  overall  efficiency  of 
forage  breeding  programs. 


LITERATURE  CITED 

Barton,  F.E.  and  D.  Burdick.  1981. 
Prediction  of  Forage  Quality  with  NIR 
Reflectance  Spectroscopy,  pp.  532-534.  IN: 
Proc.  XIV  Int.  Grassld.  Congress.  Lexington, 
Ky. 

Bouton,  J.H.  and  S.R.  Smith.  1988. 
Germplasm  Evaluation  Methods  for  Grazing 
Tolerance.  IN:  Proceedings  of  10th 

Trifolium  Conference  (Corpus  Christi,  TX.) 

Burton,  G.W.  1982.  Forage  Breeding  and 
Selection.  IN:  Proceeding  38th  SPFCIC.  pp. 
130-133. 

Coleman,  S.W.  1979.  Forage  Quality 
Assessment:  Important  Factors  for  Plant 

Breeders  to  Consider.  IN:  Proceedings  of 
39th  SPFCIC.  pp.  71-84. 

Donald,  C.M.  1978.  Summative  Address:  Two. 
IN:  Plant  Relations  in  Pastures,  pp.  411- 
420.  J.R.  Wilson,  ed.  CSIRO.  Melbourne. 


Eichhorn,  M.M.,  W.M.  Oliver,  W.B.  Hallmark, 
W.A.  Young,  A.V.  Davis,  and  B.D.  Nelson. 
1986.  Registration  of  'Grazer'  Bermudagrass . 
Crop  Sci . 26:835. 

Frame,  J.  1981.  Herbage  Mass.  IN:  Sward 
Measurement  Handbook.  pp.  39-69.  J. 
Hodgeson,  R.D.  Baker,  A.  Davis,  A.S.  Laidlaw, 
and  J.D.  Leaver,  eds.  British  Grassland 
Society,  Hurley. 

Gabrielson,  B.C.  1986.  Evaluation  of 
Marketed  Cellulases  for  Activity  and  Capacity 
to  Degrade  Forages.  Agron.  J.  78:838-842. 

Haal and , R.L.  1979.  Forage  Plant 
Improvement  for  the  Future.  IN:  Proceedings 
of  the  36th  SPFCIC.  pp.  34-36. 

Hill,  N.S.,  J.C.  Peterson,  J.A.  Stuedemann, 
and  F.E.  Barton.  1988.  Prediction  of 
Percentage  Leaf  in  Stratified  Canopies  of 
Alfalfa  with  Near  Infrared  Reflectance 
Spectroscopy.  Crop  Sci.  28:354-357. 

Holt,  E.C.  1963.  Evaluation  of  Leafiness  in 
Blue  Panicgrass,  Panicum  antidotale  Retz. 
Crop  Sci . 3:412-415. 

Holt,  E.C.,  B.E.  Conrad,  W.C.  Ellis,  and  E.C. 
Bashaw.  1983.  Brazos  Bermudagrass.  TAES  L- 
2068. 

Hoveland,  C.S.  1979.  Grazing  Management  and 
Utilization  Research  Prior  to  the  Release  of 
Pasture  Cultivars.  IN:  Proceedings  of  the 
39th  SPFCIC.  pp.  85-88. 

Jones,  D.I.H.  and  M.V.  Hayward.  1973.  The 
Effect  of  Pepsin  Pretreatment  of  Herbage  on 
the  Prediction  of  Dry  Matter  Digestibility. 
J.  Sci.  Food  Agric.  26:711-718. 

Jones,  G.M.,  N.S.  Wade,  T.P.  Baker,  and  E.M. 
Ranch.  1987.  Use  of  NIRS  in  Forage  Testing. 
J.  Diary  Sci.  70:1086-1091. 

Jones,  R.J.  and  B.  Walker.  1983.  Strategies 
for  Evaluating  Forage  Plants.  IN:  Genetic 
Resources  of  Forage  Plants.  pp.  185-201. 
J.G.  Mclvor  and  R.A.  Bray,  eds.  CSIRO, 
Melbourne. 

Laredo,  M.A.  and  D.J.  Minson.  1973.  The 
Voluntary  Intake,  Digestibility,  and 
Retention  Time  by  Sheep  of  Leaf  and  Stem 
Fractions  of  Five  Grasses.  Aust.  J.  Agric. 
Res.  24:875-894. 

Lippke,  H.  1983.  Forage  Attributes  for 
Improved  Animal  Performance.  IN:  Proceedings 
of  the  39th  SPFCIC.  pp.  56-60. 

Mochrie,  R.D.,  J.C.  Burns,  and  D.H.  Timothy. 
1981.  Recommended  Protocols  for  Evaluating 
New  Forages  for  Ruminants,  pp.  553-559.  IN: 
Forage  Evaluation:  Concepts  and  Techniques, 
CSIRO,  Melbourne. 


35 


Mott,  G.O.  and  J.E.  Moore.  1985.  Evaluating 
Forage  Production.  pp.  422-429.  IN: 
Forages:  The  Science  of  Grassland 

Agriculture.  M.E.  Heath,  R.F.  Barnes,  and 
D.S.  Metcalf,  eds.  Iowa  State  University 
Press,  Ames,  IA. 

Peterson,  M.A.  1988.  Comparison  of  Double- 
Sampling  Techniques  for  Estimating  Forage 
Production.  M.S.  Thesis,  Texas  A&M 
University.  84  pp. 

Peterson,  M.A.  and  M.A.  Hussey.  1987.  Use 
of  Double-Sampling  Techniques  to  Estimate 
Herbage  Mass  in  Bermudagrass.  IN:  Forage 
Research  in  Texas.  pp.  69-70.  TAES  CPR- 
4537. 

Quensenberry , K.H.  1980.  Field  and 

Greenhouse  Innovations  for  the  Forage  Plant 
Breeder.  IN:  Proceedings  of  the  37th 

SPFCIC . pp.  81-87. 

Quensenberry,  K.H.,  R.L.  Smith,  S.C.  Schank, 
and  W.R.  Ocumpaugh.  1977.  Tropical  Grass 
Breeding  and  Early  Generation  Testing  with 
Grazing  Animals.  IN:  Proceedings  of  the 
34th  SPFCIC.  pp.  100-103. 

Riewe,  M.E.  1982.  Forage  Evaluation 
Techniques.  IN:  Proceedings  of  the  38th 
SPFCIC.  pp.  116-122. 

Sleper,  D.A.,  F.A.  Martz,  A.G.  Matches,  and 
J.R.  Forwood.  1983.  The  Need  for  Animal 
Trials.  IN:  Proceedings  of  the  39th  SPFCIC. 

pp.  61-66. 

Stair,  D.W.,  M.A.  Hussey,  and  H.  Lippke. 
1987.  Use  of  Pepsin-Cellulase  for  Estimating 
Forage  Nutritive  Value.  IN:  Forage  Research 
in  Texas,  pp.  72-73.  TAES  CPR-4537. 

Watkin,  B.R.  and  R.J.  Clements.  1976.  The 
Effects  of  Grazing  Animals  on  Pastures,  pp. 
273-289.  IN:  Plant  Relations  in  Pastures. 
J.R.  Wilson,  ed.  CSIR0,  Melbourne. 


TABLE  1 Generalized  Plan  for  the 
Evaluation  of  Forage  Germplasm1 


STAGE  I Genotypes  grown  as  spaced  plants 
at  1-2  locations.  Data  is 
collected  to  determine  general 
adaptation.  Lines  are  eliminated 
based  on  forage  production,  IVDMD, 
winterhardiness,  etc.  Duration 
1-2  years. 

STAGE  II  Small  plot  evaluations  of  selected 
lines  from  the  Stage  I evaluation. 
Genotypes  eliminated  based  on 
performance  under  cutting.  Plots 
may  be  subjected  to  grazing. 
Duration  2-3  years. 

STAGE  III  Superior  genotypes  from  Stage  II 
moved  to  larger  plots  or  paddocks. 
Pastures  subjected  to  grazing 
management  at  multiple  stocking 
rates.  Animal  production  measured. 
Lines  eliminated  based  on 
persistence,  animal  performance, 
etc.  Duration  2-5  years. 


Adapted  from  Jones  and  Walker  (1983) 


36 


ECOLOGY  AND  PHYSIOLOGY  INFORMATION  EXCHANGE 
GROUP 

INFLUENCE  OF  THE  FUNGAL  ENDOPHYTE  ON  PHYSIOLOGY 
AND  ECOLOGY  OF  TALL  FESCUE 

C.  S.  Hove 1 and— ^ 

Poor  animal  performance  on  tall  fescue  (Festuca 
arundinacea  Screb . ) is  widespread  and  is 
associated  with  the  fungal  endophyte  Acremonium 
coenophialun  Morgan-Jones  and  Gams  (Stuedemann 
and  Hoveland,  1988).  The  vastly  improved  animal 
performance  on  low-endophyte  tall  fescue  has 
encouraged  release  of  endophyte-free  cultivars 
which  are  being  aggressively  marketed. 

Techniques  have  been  developed  for  destroying 
existing  infested  tall  fescue  sods  and  replant- 
ing with  endophyte-free  seed.  Substantial 
acreages  have  been  replanted  and  this  trend  is 
expected  to  continue.  Surveys  in  Alabama  show 
that  endophyte-free  acreage  has  increased  by 
130,000  acres  from  1984  to  1987  (Ball,  1987). 
However,  concern  has  been  expressed  by  research 
and  extension  workers  as  to  potential  dangers 
in  tolerance  of  endophyte-free  tall  fescue  to 
environmental  stress. 

The  popularity  of  tall  fescue  is  a result  of 
its  wide  adaptation,  ease  of  establishment, 
long  productive  season,  tolerance  to  grazing, 
drought,  poor  drainage,  pests,  and  a wide 
range  in  soil  pH  (Burns  and  Chamblee,  1979). 

Most  endophytic  fungus-grass  associations 
are  mutualistic  (Bacon,  et  al . 1986).  This 
suggests  that  these  fungi  co-evolved  with 
their  grass  host,  are  non-parasitic , and  the 
endophyte-plant  relationship  is  a mutualistic 
symbiosis  (Siegel  et  al.,  1987b;  Bacon  and 
Siegel,  1988).  This  raises  the  question  of 
whether  tall  fescue,  when  free  of  the  endophyte, 
continues  to  have  the  same  productivity  and 
persistence  as  infected  grass  in  stressful 
environments  (Siegel,  et  al.  1987a). 

The  fungus  benefits  from  the  association  by 
receiving  nutrients,  protection,  reproduction, 
and  dissemination  (Bacon  and  Siegel,  1988). 

In  return,  the  plant  may  be  aided  by  modified 
plant  morphology,  enhanced  pest  protection, 
growth  stimulation,  and  greater  tolerance  to 
drought  and  grazing,  resulting  in  better 
competition  with  other  species  in  a pasture. 

This  paper  attempts  to  review  some  of  the 
changes  in  or  benefits  to  the  tall  fescue 
plant  from  the  endophyte  assocation. 

PLANT  RESPONSES  TO  THE  ENDOPHYTE 
Plant  Morphology 

Endophyte-infected  (El)  cloned  plants  had 
thicker  and  narrower  leaf  blades  than  endophyte- 
free  (EF)  plants,  and  flooding,  N rate,  or 
drought  stress  did  not  appreciably  alter  this 


1/Department  of  Agronomy,  University  of  Georgia, 

Athens , GA  30602 


characteristic  ( Arechavale ta , 1987).  The 
benefits  of  this  morphological  change  are 
unknown,  but  it  could  contribute  to  plant 
water  conservation  and  drought  tolerance. 

Hill  et  al . (1987)  reported  that  El  plants  of 
two  clones  had  more  erect  growth  and  crowns 
imbedded  deeper  in  the  soil  than  EF  plants. 

Maturation  and  Seed  Development 

A difference  in  ultrastructural  morphology  of 
mesophyll  tissue  of  leaf  sheaths  was  the 
earlier  occurrence  of  air  spaces  in  El  than 
EF  plants;  this  difference  disappearing  as 
plants  aged  (Arechavaleta , 1987).  This 
suggests  that  the  endophyte  may  accelerate 
plant  maturation  rate.  In  the  field,  cloned 
material  of  El  plants  produced  seedheads  up 
to  2 weeks  earlier  than  EF  plants  (C.  VJ.  Bacon, 
unpublished).  Similar  results  have  been 
observed  in  the  field  on  a number  of  clones 
(N.  S.  Hill,  unpublished).  Thus,  the  endophyte 
may  interact  with  some  developmental  mechanism 
such  as  production  of  a growth  regulator  in  the 
plant  to  alter  growth  processes. 

In  a comparison  of  Ky  31  tall  fescue  El  and  EF 
plants  grown  from  seed,  El  plants  produced  up 
to  twice  the  seed  yield  of  EF  plants  (Clay, 
1987).  The  genetic  diversity  of  this  cultivar 
limits  the  value  of  these  results  but  research 
by  Rice  et  al.  (1987)  in  South  Carolina  with 
20  clones  showed  El  plants  had  greater  seed 
weight,  more  seeds,  and  more  panicles  per  plant 
than  EF  plants.  These  results  suggest  that 
population  shifts  could  occur  over  time  by 
natural  reseeding  of  mixed  populations,  increas 
ing  the  level  of  infestation  in  a pasture. 

Tiller  Development 

Tiller  numbers  were  substantially  greater  on  El 
than  EF  plants  when  grown  at  a high  N rate 
(Belesky  et  al . 1987a:  Clay,  1987;  Hill,  et 

al.  1987).  In  another  study,  similar  results 
were  obtained  with  cloned  plants  at  a high  N 
rate  but  at  low  and  medium  N rates  there  was 
no  difference  In  tillering  of  El  and  Ew  plants 
(Arechavaleta,  1987).  Tiller  development 
is  related  to  N rate  but  responses  to  the 
endophyte  appear  to  differ  among  tall  fescue 
clones . 


Nutrients 

Herbage  production  of  cloned  EF  and  El  plants 
were  similar  at  low  N levels  but  at  higher 
rates  the  El  plants  were  67%  greater  than  EF 
plants  (Arechavaleta,  1987).  More  efficient 
utilization  of  N may  occur  in  El  than  EF 
plants  at  higher  N rates.  Lyons  (1985)  found 
that  El  and  EF  plants  responded  differently  to 
N rates,  and  that  as  N rate  increased,  the 
glutamine  synthetase  activity  of  El  plants  was 
greater.  This  was  interpreted  to  represent 


37 


in  the  blades  of  El  tall  fescue,  a C,,  plant 
that  loses  substantial  amounts  of  NIl^  through 
photorespiration . 

Uptake  of  minerals  was  generally  unaffected  by 
infection  status  of  tall  fescue  except  that  El 
clones  were  slightly  higher  in  K and  lower  in 
B concentration  (Wilkinson,  1987).  One 
interesting  finding  was  that  although  herbage 
growth  was  unaffected  by  infection  status, 
root  growth  was  significantly  greater  in  El 
than  EF  clones  when  grown  in  P deficient  soil. 

Herbage  Yield 

Herbage  yields  of  El  plants  have  been  substan- 
tially higher  than  EF  plants  of  the  same  clone 
( Arechavaleta , 1987,  1987;  Belesky,  1987a). 
However,  this  advantage  is  affected  by  N 
rate,  drought  stress,  and  individual  clones. 
Hill  et  al.  (1987)  found  no  yield  advantage  in 
El  plants  of  five  clones.  In  field  trials 
where  El  and  EF  Ky  31  tall  fescue  from  seed 
have  been  compared,  yield  differences  were 
small  or  non-existent  (Pedersen  et  al.,  1982; 
Siegel  et  al.,  1987).  The  lack  of  yield 
response  in  seed-planted  trials  suggests 
that  the  genetic  plasticity  within  a diverse 
cultivar  such  as  Ky  31  tall  fescue  includes 
individuals  which  would  not  behave  as  a 
single  infected  clone  and  its  noninfected 
ramet.  The  yield  advantage  of  El  over  EF 
plants  of  the  same  clone  may  be  a result  of 
higher  photosynthetic  rate  (Belesky  et  al., 
1987a).  The  endophyte  may  also  alter  growth 
processes  by  producing  a growth  regulator. 
Porter  et  al.  ( 1985)  found  in  vitro  production 
of  auxin  by  one  endophyte,  Balansia  epichloe , 
but  it  is  not  known  if  this  endophytic  fungus 
can  regulate  plant  growth.  In  perennial 
ryegrass  (Lolium  perenne  L.),  gibberillin, 
which  generally  stimulates  cell  expansion,  was 
considered  a likely  possibility  for  increasing 
yield  of  El  plants  (Latch,  et  al.  1985). 

Drought 

Stand  persistence  of  EF  tall  fescue  was  not  a 
problem  in  a 4-year  grazing  trial  on  heavy 
clay  soils  in  central  Alabama  (Hoveland  et  al . 
1983).  This  has  generally  been  the  case 
except  on  coarser  textured  soils  or  under 
greater  drought  stress.  This  is  illustrated 
by  the  results  of  Read  and  Camp  (1986)  in  the 
lower  rainfall  Blacklands  of  Texas  where  El 
tall  fescue  produced  more  herbage  and  had 
better  drought  survival  than  EF  plants  in  a 
3-year  grazing  trial.  Severe  droughts  during 
the  past  several  years  in  the  southeastern 
states  have  resulted  in  stand  losses  of  EF 
tall  fescue,  particularly  when  grazed  closely 
in  summer.  When  harvested  every  3 weeks,  EF 
Ky  31  and  AU  Triumph  tall  fescue  stand  losses 
were  greater  during  two  drought  years  when  cut 
at  a 1-^-inch  as  compared  to  a 3-inch  stubble 
(C.  S.  Hoveland,  unpublished). 


The  mutualistic  relationship  of  endophyte  and 
plant  which  favors  higher  herbage  production 
also  occurred  under  mild  moisture  stress 
(0.05  MPa)  in  a study  with  individual  clones 
(Arechavaleta,  1987).  Under  more  severe 
moisture  stress  (-0.50  MPa)  where  no  signifi- 
cant yield  advantage  occurred,  75%  of  the  EF 
plants  died  and  all  El  plants  survived. 

Belesky  et  al.  (1986b)  reported  that  El 
clones  responded  to  decreased  water  availabil- 
ity by  limiting  growth  while  EF  clones 
continued  to  groxj  for  a time  at  rates  similar 
to  non-limiting  conditions.  White  and  Comeau 
(1987)  reported  that  El  plants  had  lower  CO^ 
exchange  and  transpiration  than  EF  plants 
of  the  same  clone.  In  another  clone,  the 
endophyte  had  the  opposite  effect. 

The  superior  drought  tolerance  of  El  tall 
fescue  plants  would  insure  survival  and 
improve  their  competitive  ability  under 
moisture  stress  in  a pasture.  Drought 
tolerance  may  partially  be  a result  of  leaf 
rolling  which  is  facilitated  by  narrower  and 
thicker  leaves  of  El  plants.  Increased  leaf 
rolling  was  observed  for  El  plants  under 
drought  stress  (Arechavaleta,  1987).  Roll- 
ing is  an  adaptive  mechanism  that  reduces 
effective  leaf  area  and  hence  the  amount  of 
heat  that  strikes  that  area  (Parsons,  1982). 
The  difference  in  stomatal  resistance  in  El 
plants  relative  to  EF  plants  reported  by 
Belesky  et  al.  (1987)  is  also  indicative  of 
a drought  tolerance  mechanism. 

Grazing  tolerance 

It  is  often  suggested  that  a beneficial 
effect  of  the  endophyte  on  tall  fescue 
persistence  is  reduced  palatability  and  thus 
less  plant  stress  from  overgrazing  (Bacon 
and  Siegel,  1988).  The  lower  grazing  pressure 
on  El  as  compared  to  EF  pastures  may  be  a 
result  of  lower  intake  and  animals  spending 
more  time  in  the  shade  because  of  intolerance 
to  heat  when  afflicted  with  fescue  toxicosis 
(Stuedemann  and  Hoveland,  1988).  Results  of 
a grazing  preference  trial  with  El  and  EF 
space  plants  in  South  Carolina  furnished  no 
evidence  of  a significant  preference  for  EF 
plants  (Chrestman  et  al.,  1987). 

Plant  Pathogens 

Various  endophytic  fungi  infecting  Festuca 
species  have  been  shown  to  cause  in  vitro 
inhibition  of  several  grass  pathogens 
(White  and  Cole,  1985,  1986).  However,  this 
reviewer  is  not  aware  of  any  published  or 
unpublished  evidence  of  improved  disease 
resistance  in  El  tall  fescue  grown  in  the 
field . 


Insects 

There  is  strong  evidence  that  EF  tall  fescue 
is  more  susceptible  than  El  grass  to  attack 


38 


by  several  insect  species.  In  a review  by 
Bacon  and  Siegel  (1988),  endophytes  were 
involved  in  tall  fescue  resistance  to  attack 
by  Argentine  stem  weevil,  fall  armyworm, 
house  cricket,  oat  birdcherry  aphid,  greenbug 
aphid,  and  milkweed  bug.  Jessup  tall  fescue 
seeded  into  bahiagrass  (Paspalum  notatum 
Fltigge)  sod  in  south  Georgia  was  severely 
damaged  by  crickets  on  EF  plots  while  El 
plots  were  unaffected  (J.  H.  Bouton,  personal 
communication) . 


Nematodes 

Nematodes  were  shown  to  adversely  affect 
persistence  and  productivity  of  tall  fescue 
in  sandy  Coastal  Plain  soils  (Hoveland  et 
al . , 1975).  Plant  parasitic  nematodes  reduced 
forage  and  root  growth  of  large-rooted  much 
more  than  small-rooted  genotypes  (Elkins  et 
al.,  1S79).  However,  since  the  endophyte 
infection  status  of  plants  in  this  study 
was  not  known,  one  can  not  be  sure  that  root 
size  was  the  determining  factor.  In  a recent 
Alabama  study  by  Pedersen  et  al . (1988),  soil 
and  root  nematode  populations  were  much  lower 
on  El  than  EF  plants.  In  an  Arkansas  field 
study,  EF  tall  fescue  was  more  severely 
drought  stressed  than  El  grass  as  indicated 
by  higher  canopy  temperature  (West,  1987). 
Nematode  populations  were  substantially 
higher  in  EF  than  El  tall  fescue.  The 
severe  stand  losses  reported  by  Joost  (1987) 
in  Louisiana  on  EF  as  compared  to  good  per- 
sistence on  El  tall  fescue  of  GA  5 and  Ky  31 
cultivars  may  be  related  to  nematode  resis- 
tance. If  these  results  are  confirmed,  it 
may  require  that  El  tall  fescue  be  used  if 
this  grass  is  to  be  grown  on  sandy  Coastal 
Plain  soils. 


CONCLUSIONS 

The  benefits  to  enhanced  animal  performance 
with  EF  tall  fescue  has  encouraged  destruc- 
tion of  El  pastures  and  replanting  with  EF 
seed.  However,  increasing  evidence  indicates 
that  the  endophyte  may  have  beneficial  effects 
on  the  host  plant  in  respect  to  plant  morpho- 
logy, nutrient  responses,  herbage  yield, 
drought  tolerance,  and  insect  and  nematode 
tolerance.  These  factors  may  contribute  to 
plant  persistence,  competition  with  weeds 
and  warm  season  perennial  grasses,  and 
tolerance  to  abusive  grazing  practices. 

Genetic  variation  among  tall  fescue  clones  in 
relation  to  the  endophyte  indicate  that  the 
relationship  is  complex  and  deserves  much 
study.  However,  it  is  certain  that  the 
tall  fescue-endophyte  is  one  of  mutualistic 
symbiosis . 

Since  the  endophyte  benefits  the  tall  fescue 
plant  in  a number  of  ways,  the  question  must 
be  asked  about  the  wisdom  of  replanting 
infested  pastures  with  EF  seed.  In  the 


extreme  lower  part  of  the  tall  fescue  belt 
where  environmental  stresses  are  greatest  and 
warm  season  perennial  grasses  are  an  important 
component  in  swards,  the  dilution  effect 
in  animal  diets  may  allow  El  grass  to  be  used 
successfully.  Likewise,  where  legumes  are 
maintained  in  tall  fescue,  El  grass  may  be 
satisfactory  for  certain  classes  of  livestock. 
However,  where  tall  fescue  is  the  sole  com- 
ponent in  a pasture  and  high  rates  of  N are 
applied  (such  as  in  broiler-producing 
areas)  it  would  appear  best  to  replant  with  EF 
seed.  Even  if  stands  persist  only  5 or  6 
years,  the  enhanced  animal  performance  is  well 
worth  the  cost  of  replanting  thinning  stands. 
Present  indications  are  that  EF  tall  fescue  is 
less  tolerant  of  abuse  in  stressful  environ- 
ments. This  suggests  that  grazing  pressure 
on  EF  tall  fescue  should  be  reduced  in  summer, 
especially  under  drought  conditions. 

LITERATURE  CITED 

Arechavaleta , M.  1987.  Effect  of  endophyte 
presence  on  tall  fescue  response  to  drought, 
flooding,  and  nitrogen  fertilization.  H.S. 
thesis.  Univ.  of  Georgia,  Athens,  GA . 

Bacon,  C.  W.  and  M.  R.  Siegel.  1988. 

Endophyte  parasitism  of  tall  fescue.  J.  Prod. 
Agr ic . 1 :45-55 . 

Bacon,  C.  W.,  P.  C.  Lyons,  J.  K.  Porter,  and 
J.  D.  Robbins.  1986.  Ergot  toxicity  from 
endophyte-infected  grasses:  a review.  Agron. 
J.  106-116. 

Ball,  D.  1987.  Extension  report.  Proc.  Tall 
Fescue  Endophyte  Meeting.  Memphis,  TN 

Belesky,  D.  P.,  0.  J.  Devine,  J.  E.  Pallas, 
Jr.,  and  W.  C.  Stringer.  1987a.  Photosynthe- 
tic activity  of  tall  fescue  as  influenced  by  a 
fungal  endophyte.  Photosynthetica  21:82-87. 

Belesky,  D.  P.,  W.  C.  Stringer,  and  M.  A. 
Thornton.  1987b.  Endophyte  effects  upon  tall 
fescue  growth  under  water  deficit  conditions. 
Agron.  Abst.  p.  87. 

Burns,  J.  C.  and  D.  S.  Chamblee.  1979. 
Adaptation.  In  R.  C.  Buckner  and  L.  P.  Bush 
(ed.)  Tall  fescue.  Agronomy  29:9-30. 

Chrestman,  R.  E.,  W.  C.  Stringer,  D.  L.  Cross, 
and  J.  S.  Rice.  1987.  Selection  of  beef 
cattle  between  endophyte-infected  and  non- 
infected  tall  fescue.  Agron.  A,bst.  p.  107. 

Clay,  K.  1987.  Effects  of  fungal  endophytes 
on  the  seed  and  seedling  biology  of  Lol ium 
perenne  and  Festuca  arundinacea . Oecologia 
73:358-367. 

Elkins,  C.  B.,  R.  L.  Haaland , R.  Rodriguez- 
Itabana,  and  C.  S.  Hoveland.  1979.  Plant 
parasitic  nematode  effects  on  water  use  and 
nutrient  uptake  of  a small-  and  large-rooted 
tall  fescue  genotype.  Agron.  J.  71:497-500. 


39 


Hill,  N.  S.,  W.  C.  Stringer,  and  J.  C. 
Petersen.  1987.  Morphology,  yield,  and 
carbohydrate  reserves  of  tall  fescue  acces- 
sions as  affected  by  endophyte  infection  and 
harvest  management.  Agron . Abst.  p.  143. 

Hoveland,  C.  S.,  R.  L.  Haaland,  C.  C.  King, 
Jr.,  J.  W.  Odom,  S.  P.  Schmidt,  E.  M.  Clark, 

J.  A.  McGuire,  L.  A.  Smith,  H.  W.  Grimes, 
and  J.  L.  Holliman.  1983.  Steer  performance 
and  association  of  Acremonium  coenophialum 
fungal  endophyte  on  tall  fescue  pasture. 

Agron.  J.  75:821-824. 

Hoveland,  C.  S.,  R.  Roar  iguez-ICabana , and 
C.  D.  Berry.  1975.  Phalaris  and  tall  fescue 
forage  production  as  affected  by  nematodes 
in  the  field.  Agron.  J.  67:714-717. 

Joost,  R.  E.  1987.  Louisiana  report.  Proc. 
Tall  Fescue  Endophyte  Meeting.  Memphis,  TN . 

Latch,  G.  C.  M.,  W.  F.  Hunt,  and  D.  R. 
Musgrave.  1985.  Endophytic  fungi  affect 
growth  of  perennial  ryegrass.  N.  Z.  J.  Agric. 
Res.  28:165-168. 

Lyons,  P.  C.  1985.  Infection  and  in  vivo 
ergot  alkaloid  synthesis  by  the  tall  fescue 
endophyte  and  effects  of  the  fungus  on  host 
nitrogen  metabolism.  PhD  dissertation. 
University  of  Georgia,  Athens,  GA. 

Parsons,  L.  R.  1982.  Plant  responses  to 
water  stress.  In  M.  N.  Christiansen  and  C.  F. 
Lewis  (ed.)  Breeding  plants  for  less  favorable 
environments.  John  Wiley  & Sons.  New  York. 

Pedersen,  J.  F.,  C.  S.  Hoveland,  and  R.  L. 
Haaland.  1982.  Performance  of  tall  fescue 
varieties  in  Alabama.  Auburn  Univ.  Agric. 

E::p.  Stn.  Cir.  262  . 

Pedersen,  J.  F.,  R.  Rodr iguez-Kabana , and 
R.  A.  Shelby.  1988.  Ryegrass  cultivars  and 
endophyte  in  tall  fescue  affect  nematodes  in 
the  grass  and  succeeding  soybeans.  Agron.  J. 
(Accepted  for  publication). 

Porter,  J.  K.,  C.  W.  Bacon,  H.  G.  Cutler, 

R.  F.  Arendale,  and  J.  D.  Robbins.  1985. 

In  vitro  auxin  production  by  Balansia 
epichloe . Phytochemistry  24:1429-1431. 

Read,  J.  C.  and  B.  J.  Camp.  1986.  The 
effect  of  the  fungal  endophyte  Acremonium 
coenophialum  in  tall  fescue  on  animal 
performance,  toxicity,  and  stand  maintenance. 
Agron.  J.  78:848-850. 

Rice,  J.  S.,  B.  W.  Pinkerton,  W.  C.  Stringer, 
and  D.  J.  Undersander.  1987.  South  Carolina 
report.  Proc.  Tall  Fescue  Endophyte  Meeting. 
Memphis.  TN . 

Siegel,  M.  R.,  L.  P.  Bush,  and  D.  D.  Dahlman. 
1987a.  What  am  I losing  by  removing  endophyte 
from  tall  fescue?  pp . 41-44  in  Proc.  43rd 
Southern  Pasture  and  Forage  Crop  Improvement 
Conf.  April  2-22.  Clenson,  SC. 


Siegel,  M.  R.,  G.  C.  M.  Latch,  and  M.  C. 
Johnson.  1987b.  Fungal  endophytes  of  grasses 
Annu.  Rev.  Phytopathol.  25:293-315. 

Stuedemann,  J.  A.  and  C.  S.  Hoveland.  1988. 
Fescue  endophyte:  history  and  impact  on 
animal  agriculture.  J.  Prod.  Agric.  1:39-44. 

West,  C.  P.  1987.  Arkansas  report.  Proc. 
Tall  Fescue  Endophyte  Meeting.  Memphis,  TN . 

Wilkinson,  S.  R.  1987.  Georgia  report. 

Proc.  Tall  Fescue  endophyte  Meeting.  Memphis, 
TN. 

White,  J.  F.,  Jr.,  and  G.  T.  Cole.  1985. 
Endophyte-host  association  in  forage  grasses. 

III.  In  vitro  inhibition  of  fungi  by 
Acremonium  coenophialum.  Mycologia  77:487-489 

White,  J.  R.,  Jr.,  and  G.  T.  Cole.  1986. 
Endophyte-host  association  in  forage  grasses. 

IV.  The  endophyte  of  Festuca  ver suta . 

Mycologia  78:102-107. 

White,  R.  H.  and  M.  Comeau.  1987.  Tall 
fescue  leaf  gas  exchange  and  water  relations 
as  influenced  by  endophytic  fungi.  Agron. 
Abst.  p . 140. 


THE  ECOLOGY  AND  PHYSIOLOGY  OF  COOL-SEASON 
FORAGES  UNDER  INTENSIVE  ROTATIONAL  GRAZING 
SYSTEMS 

C.T.  Dougherty—^ 

Development  of  simulation  models  of  crops  and 
pastures  is  a prequisite  for  the  construction  of 
models  of  grazing  systems.  The  current  rate  of 
expansion  of  our  knowledge  base  in  the 
physiology  of  crop  yield  has  been  quite  slow, 
perhaps  as  a consequence  of  a reductionist 
approach  in  the  agricultural  and  biological 
sciences,  and  this  has  hindered  the  development 
of  simulation  models.  Short-comings  of  models 
based  on  physiology  have  not  restricted  the 
march  of  computer-based  technology  into 
agriculture  for  expert  systems  models  have 
replaced  many  of  the  complex  ones  and  have  also 
generated  many  new  applications  because  of  their 
versatility,  low  cost,  and  because  most  can 
function  in  personal  computers.  Models  of 
grazing  systems  are  complex  because  of  the 
involvement  of  grazing  animals,  the  number  of 
plant  species  the  complexities  at  the  interfaces 
of  soil,  plant,  animal  and  environment.  One  of 
the  areas  that  we  are  not  particularly 
well-versed  is  in  the  physiology  of  species  and 
communities  in  grasslands  as  modified  by 
grazing.  Recent  emphasis  at  the  plant-animal 
interface  has  helped  in  the  definition  of  the 
problem.  Plants  grown  in  communities  (eg 
pastures)  are  different  and  react  differently 
from  plants  that  are  grazed  by  herbivores. 
Parsons  and  Johnson  (1986)  also  point  out 
differences  in  physiological  responses  of  grazed 
and  cut  swards . 

A symposium  in  Australia  in  1976  addressed  many 
aspects  of  plant  relations  in  pastures  (Wilson 
1978)  while  others  in  United  Kingdom  were 
concerned  with  sward  composition  and 
productivity  (Charles  and  Haggar,  1979)  and 
plant  physiology  and  herbage  production  (Wright, 
1981).  More  recent  views  were  expressed  in  a 
volume  of  grazing  published  by  British  Grassland 
Society  (Frame,  1986)  and  the  proceedings  of 
session  dedicated  to  the  plant-animal  interface 
in  grazing  lands  at  the  15th  International 
Grassland  Congress  in  Japan  in  1985  (Horn  et 
al.,  1987). 

The  following  discussion  is  limited  to 
physiological  aspects  of  vegetative  tillers  of 
cool-season  grasses. 


—Department  of  Agronomy,  University  of 
Kentucky,  Lexington,  KY  40540-0091 


ENERGY  RELATIONS 

Defoliation  by  grazing  impacts  the  energy 
balance  of  pastures  with  often  negative  effects 
on  photosynthesis  and  growth  largely  because  of 
reduction  in  photosynthetic  surface  area.  Brown 
(1987)  reviewed  this  area  at  this  conference 
last  year,  however,  light  relations  are  critical 
to  other  reactions  of  swards  to  grazing  that 
will  be  discussed  subsequently. 

The  amount  of  energy  absorbed  by  a pasture  is 
represented  by  the  following  equation: 

Ia  = I - IQe  (1) 

Where  I3  = absorbed  energy 
I = incident  energy 
K = extinction  coefficient 
L = leaf  area  index 

Defoliation  of  tillers  by  grazing  drastically 
reduces  leaf  area  index  and,  consequently,  the 
amount  of  energy  absorbed.  Defoliation  may  also 
decrease  the  extinction  coefficient  through 
charges  in  canopy  architecture.  Since  animals 
remove  the  younger  upper  canopy  leaves  that  have 
higher  photosynthetic  efficiency,  the  efficiency 
of  conversion  of  absorbed  light  into  herbage 
mass  may  also  be  reduced.  In  swards  at  ceiling 
yields,  where  crop  growth  rates  are  zero, 
grazing  may  restore  accumulation  of  herbage 
mass.  This  situation  may  occur  in 
rotationally-grazed  pastures  with  long  regrowth 
periods  and  in  pastures  under  summer  and  fall 
stockpile  managements. 

Reductions  in  the  amount  of  energy  absorbed  by 
the  sward  due  to  grazing  reduce  the 
transpiration  component  of  evapotranspiration 
(latent  energy  flux)  and  increase  the  sensible 
heat  and  soil  heat  fluxes,  modifying  the  thermal 
environment  of  the  soil,  sward  and  grazing 
animal . 


TILLER  DYNAMICS 

The  responses  of  vegetative  cool-season  grasses 
to  grazing  and  their  recovery  during  rest 
periods  is  a function  of  populations  and  size  of 
tillers  (Simon  and  Leraaire,  1987).  Grasses  that 
produce  and  persist  under  grazing  have  the 
ability  to  adapt  in  terms  of  both  size  and 
population.  Tillers  of  ryegrass  and  fescue  can 
exist  under  close  grazing  or  mowing  as  in  turf 
yet  can  respond  to  long  recovery  phases  and 
support  high  herbage  mass.  Grasses  such  as 
Kentucky  bluegrass  (Poa  pratensis  L.)  lack  such 
adaptability  and  produce  lower  yields. 


TILLER  MORPHOLOGY 

Phytomer  structure  of  cool-season  grasses  has 
been  reduced  during  evolution  in  a similar 
manner  to  the  reduction  of  floral  parts.  True 


41 


leaf  lamina  are  absent  in  vegetative  tillers  and 
their  functions  have  been  taken  over  by  petioles 
which  have  evolved  into  blades  and  sheaths 
(Langer,  1979).  Fully  developed  leaf  blades  are 
dispersed  in  the  canopy  to  facilitate 
photosynthesis  while  their  sheaths  are  arranged 
concentrically,  with  the  oldest  on  the  outside. 
This  structure  supports  the  leaf  display  in 
place  of  a true  stem  (which  does  not  develop). 
The  concentrically-arranged  leaf  sheaths  also 
shelter  the  growing  point,  tiller  and  leaf 
primordia  and  expanding  young  leaves.  This 
sheath  structure  is  called  a pseudostem  since  it 
serves  many  stem  functions.  Many  of  the 
responses  of  grass  swards  under  grazing  are  a 
consequence  of  its  petiole  morphology.  The 
extinction  coefficient,  high  optimun  leaf  area 
indices,  capacity  to  accumulate  high  herbage 
masses  and,  perhaps,  inferior  quality  (compared 
with  legumes)  are  a few  examples. 


TILLER  MORPHOLOGY  AND  THE  GRAZING  HORIZON 

Pseudostems  are  likely  to  be  of  more 
significance  in  grazing  of  species  that  have  the 
ability  to  form  large  tillers  such  as  tall 
fescue  (Festuca  arundinacea  Schreb.)  and 
ryegrass  (Lolium  spp.)  for  they  may  determine 
the  lower  level  of  the  grazing  horizon 
(Barthram,  1980),  the  unavailable  herbage  mass, 
the  level  of  utilization,  height-intake 
functions  and  the  rate  of  recovery.  The 
pseudostera  stumps  remaining  after  close  grazing 
protect  the  next  series  of  expanding  leaves  that 
form  the  new  canopy.  In  tall  fescue  the  height 
of  the  pseudostem  is  about  8cm  in  summer 
pastures  and  after  grazing  by  cattle,  some  blade 
tissue  (1-2 cm  in  length)  remains  and  probably 
contributes  to  regrowth. 

Unless  grazing  at  low  allowances  or  at  low 
levels  of  availability  cattle  will  seldom  eat 
grass  tillers  below  the  plane  represented  by  the 
tops  of  the  pseudostems.  The  pseudostem 
represents  a physical  barrier  to  the  limited 
ingestive  mechanics  of  grazing  cattle.  Sheep, 
however,  graze  white  clover  (Trifolium  repens 
L.)  between  rejected  grass  pseudostems 
(L'Huillier  et  al.,  1986). 

The  readily-grazed  horizon  of  cool-season 
vegetative  grasses  is  at  a variable  height  above 
the  soil  surface.  Tiller  populations  fluctuate 
through  out  the  year  in  response  to  variable 
rates  of  production  and  mortality  (Langer, 

1979),  consequently,  the  average  age  and  size  of 
tillers  is  continuously  changing.  Since  width 
and  length  of  blade  and  sheath  of  unstressed 
tillers  increase  progressively  with  age  the 
height  of  pseudostem  the  grazing  horizon  may  be 
elevated  during  the  periods  of  growth  when  not 
grazed.  The  pseudostem  height  is  also  modified 
in  summer  tillers  since  they  are  seldom 
vertical.  This  feature  and  its  effect  on  leaf 
angle,  increases  extinction  coefficient  and  has 
been  associated  with  summer  slump  of  growth  of 
cool-season  grasses. 


REGULATION  OF  TILLER  PRODUCTION 

Production  and  growth  of  tillers  during  the 
recovery  phase  gradually  slow  as  herbage 
accumulates.  Tillering  in  grazed  stands 
responds  to  favorable  nutrition,  water  supplies, 
temperature  and  energy  often  through  mechanisms 
based  on  the  availability  of  carbohydrate. 
Recently  Argentina  researchers  have  demonstrated 
the  importance  of  phytochrome  in  regulating 
tillering  (Casals  et  al.,  1987).  In  sparse 
swards  phytochrome,  located  near  tiller  bases, 
senses  high  proportions  of  red  light  relative  to 
far-red  light  and  stimulates  the  growth  and 
development  of  tiller  primordia  located  in  leaf 
axils.  In  dense  canopies  decreases  in  red 
relative  to  infra-red  wavelengths  are  detected 
by  phytochrome  which  then  stops  the  initiation 
of  new  tillers.  Phytochrome  is  also  involved  in 
the  etiolation  response  observed  in  the  blade 
and  sheath  of  grasses  grown  under  shade.  As  a 
consequence  of  etiolation  pseudostem  height  may 
also  increase. 


REGULATION  OF  SIZE  AND  POPULATIONS  OF  TILLERS 

When  the  tillering  ceases  in  a sward  recovering 
from  grazing,  possibly  because  of  the 
phytochrome  response,  it  enters  another  critical 
phase  of  growth  and  development.  The  increasing 
size  of  tillers  results  in  intensifying 
competition  between  tillers,  primarily  for  light 
energy.  The  sward  enters  a period  of 
self-thinning  when  weaker  tillers  die.  In 
self-thinning  populations  the  -3/2  thinning  law 
is  in  force  0^3(s  and  Harper,  1974): 
w = cp 

where  w = mean  weight  of  individual  tillers 
c = maximum  weight  of  tiller 
p = tiller  population 

This  law  is  of  significance  in  recovering  swards 
since  it  predicts  that  tillers  must  either  die 
if  their  weight  (or  area)  increases  or  become 
smaller  if  they  all  survive.  In  practice  both 
fates  befall  tillers  for  weaker  tillers  die  and, 
in  survivors,  accelerated  leaf  senesence  and 
leaf  death  reduces  tiller  size.  When  these 
events  occur  in  recovering  swards,  quality 
declines  and  intake  and  productivity  of 
livestock  suffers. 


ADAPTABILITY 

Co-evolution  of  grassland  species  and  grazing 
herbivores  has  been  responsible  for  many  of  the 
features  of  both  present  day  grassland  species 
and  domestic  livestock  (Stebbins,  1981). 
Characteristics  of  grasses,  such  as  the 
protected  growing  point,  are  widely  recognised 
adaptations  to  selection  pressures  during 
evolution.  Some  grasses  have  options  that  are 
expressed  during  stresses  such  as  those  that 
occur  during  grazing.  Axillary  buds  give 
vegetative  grasses  a major  advantage  during 
grazing.  Options  of  the  axillary  bud  include, 


42 


not  growing,  growing  as  a tiller,  or  growing  as 
a rhizome  or  as  a stolon.  Grasses  that  survive 
under  intensive  grazing  may  have  all  these 
options  open. 

Perennial  ryegrass  is  an  example  of  one  of  the 
modern  grasses  best  suited  to  grazing  under  lax 
or  intensive,  rotational  or  continuous  grazing. 
Stolons  or  rhizomes  are  seldom  mentioned  in 
descriptions  of  perennial  ryegrass  but  Korte  and 
Harris  (1987)  found  2420  stolons/ni  and  6985 
tillers/nr  in  grazed  pastures  in  summer  with 
79%  of  tillers  attached  to  stolons.  Stolons 
formed  in  grazed  swards  after  tiller  bases  were 
covered  by  earthworm  casting  or  were  pushed  into 
soil  by  animal  treading.  Korte  and  Harris 
(1987)  found  40,  450,  and  2000  stolons/m  in 
spring,  during  culm  elongation,  and  during 
summer,  respectively,  in  another  study. 

Rhizomes  are  also  found  in  tall  fescue  (Porter, 
1958)  and  are  common  under  intensive  turf 
management  (A.J.  Powell,  pers.  comm.). 


CONCLUSIONS 

Grasses  suited  to  intensive  rotational  grazing 
may  have  considerable  versatility  in  expression 
of  morphological  features.  Grasses  adapted  to 
grazing  have  the  capacity  to  produce  tillers  of 
a wide  range  in  sizes,  they  have  the  ability  to 
increase  or  decrease  tiller  populations 
according  to  management  and  environmental 
stresses.  Adaptability  of  grasses  to  grazing 
may  also  depend  on  the  versatility  of  axillary 
buds  to  produce  tillers,  rhizones  or  stolons 
under  grazing.  Several  of  the  responses  of 
swards  to  grazing  are  related  to  changes  in  the 
light  environment.  The  grass  sward  under 
grazing  is  a complex  and  dynamic  ecosystem  that 
is,  and  will  continue  to  be,  difficult  to 
describe  and  quantify  in  terms  of  physiology  and 
ecology. 


REFERENCES 

Barthram,  G.T.  1980.  Sward  structure  and  the 
depth  of  the  grazed  horizon.  Grass  and  Forage 
Science,  36:130-131. 

Brown,  H.R.  1987.  Photosynthesis  and  growth  of 
pastures  as  influenced  by  defoliation,  pp. 
34-35.  Proceedings  of  the  43rd  Southern  Pasture 
and  Forage  Crop  Improvement  Conference.  ARS. 
USDA. 

Casals,  J.J.,  R.A.  Sanchez  and  V.A.  Deregibus. 
1987.  Tillering  responses  of  Lolium  multif lorum 
plants  to  change  of  red/far-red  ratio  typical  of 
sparse  canopies.  Journal  of  Experimental 
Botany,  38:1432-1439. 


Charles,  A.M. , and  R.J.  Haggar,  1979.  Changes 
in  sward  composition  and  productivity. 

Occasional  Symposium.  No.  10.  British 
Grassland  Society,  Hurley,  England. 

Frame,  J.  1986.  Grazing.  Occasional  Symposium. 
No.  19.  British  Grassland  Society,  Hurley, 
England . 

Grant,  S.A.,  J.  King  and  G.T.  Barthram.  1987. 
The  role  of  sward  adaptations  in  buffering 
herbage-production  responses  to  grazing 
management,  pp. 21-32  In  F.P.  Horn,  J.  Hodgson, 
J.J.  Mott  and  R.W.  Brougham  (ed.). 

Grazing- lands  Research  at  the  Plant-animal 
Interface.  Winrock  International,  Morrilton, 
Arkansas . 

Horn,  F.P.,  J.  Hodgson,  J.J.  Mott  and  R.W. 
Brougham.  1987.  Grazing-lands  Research  at  the 
Plant-animal  Interface.  Winrock  International, 
Morrilton,  Arkansas. 

Kays,  S.,  and  J.L.  Harper,  1974.  The  regulation 
of  plant  and  tiller  density  in  a grass  sward. 
Journal  of  Ecology,  62:97-103. 

Korte,  C.J.,  and  W.  Harris.  1987.  Stolon 
development  in  grazed  Grasslands  Nui  perennial 
ryegrass.  New  Zealand  Journal  of  Agricultural 
Research,  30:139-148. 

Langer,  R.H.M.  1979.  How  Grasses  Grow.  Studies 
in  Biology  No.  34.  Edward  Arnold,  London. 

L'Huillier,  P.J.,  D.P.  Poppi  and  T.J.  Fraser. 
1986.  Influence  of  structure  and  composition  of 
ryegrass  and  prairie  grass-white  clover  swards 
on  the  grazed  horizon  and  diet  harvested  by 
sheep.  Grass  and  Forage  Science,  41:259-267. 

Parsons,  A.J.,  and  I.R.  Johnson.  1986. 
Physiology  of  grass  growth  under  grazing. 
pp.3-13.  In  J.  Frame  ( ed . ) Grazing.  Occasional 
Symposium  No.  19.  British  Grassland  Society, 
Hurley,  England. 

Porter,  H.L.  1958.  Rhizones  in  tall  fescue. 
Agronomy  Journal,  50:493-494. 

Simon,  J.C.,  and  G.  Lemaire.  1987.  Tillering 
and  leaf  area  index  in  grasses  in  the  vegetative 
phase.  Grass  and  Forage  Science,  42:373-380. 

Stebbins,  G.L.  1981.  Coevolution  of  grasses  and 
herbivores.  Annals  of  the  Missouri  Botanical 
Garden,  68:75-86. 

Wilson,  J.R.  1978.  Plant  Relations  in  Pastures. 
CSIRO,  Melbourne. 

Wright,  C.E.  1981.  Plant  Physiology  and  Forage 
Production.  Occasional  Symposium.  No.  13. 
British  Grassland  Society,  Hurley,  England. 


43 


ECOLOGY  AND  PHYSIOLOGY  OF  WARM  SEASON  FORAGES  IN 
INTENSIVE  ROTATIONAL  GRAZING  SYSTEMS 

F.  M.  Rouquette,  Jr.^ 

As  early  as  1598,  Archibald  Napier  of  Scotland 
suggested  a system  of  rotational  grazing  along 
with  the  use  of  common  salt  as  a fertilizer.  In 
addition,  he  received  a patent  from  James  VI  for 
this  idea  and  was  to  be  paid  10  shillings  per 
acre  for  its  use.  In  1755,  James  Anderson 
advised  farmers  to  divide  pasture  land  into  15  to 
20  divisions  and  to  allow  animals  to  graze  one 
division  at  a time.  By  1788,  Marshall  suggested 
that  farmers  should  divide  their  pastures  into 
three  parts  with  fattening  cattle  or  dairy  cows 
given  the  first  bite  of  each  division  and  then  to 
be  followed  by  replacement  or  dry  stock.  Each 
division  would  then  be  rested.  And,  in  1800, 

John  Thompson  recommended  rotational  grazing 
because  it  would  increase  grass  yield.  He  con- 
cluded that  too  heavy  grazing  pressures  caused 
reduced  forage  yields  and  caused  animals  to 
consume  too  little  forage.  Although  the 
professional  research  community  has  given  renewed 
attention  to  rotational  systems  of  grazing,  many 
of  the  concepts  and  philosophy  surrounding  this 
method  of  grazing  are  200-400  years  old.  One  of 
the  common  denominators  to  the  historically  used 
rotational  grazing  techniques  is  that  most  of 
these  systems  were  used  with  temperate  forages 
such  as  species  of  Lolium,  Trifolium,  etc.  Thus, 
some  of  the  factors  which  have  encouraged  the  use 
of  rotational  grazing  of  the  warm-season  forages 
include  historical  observations,  clipped  plot 
data,  grazing  data  from  temperate  species,  and 
grazing  data  from  arid  or  semi-arid  environments. 
There  is  a general  lack  of  data  which  investigates 
the  ecology  and  physiology  of  warm  season  grasses 
used  in  intensive  rotational  grazing  systems  in 
the  humid  southeastern  U.S. 


CLIPPED  PLOT  STUDIES 

Warm  season  perennial  grasses  of  the  southeastern 
U.S.,  and  particularly  the  bermudagrasses,  have 
been  typically  exposed  to  clipping  studies  which 
sought  to  determine  the  influence  of  defoliation 
frequency  and  severity  on  forage  dry  matter,  and 
nutritive  value  attributes.  Holt  and  Lancaster 
(1968)  reported  data  from  a 5-year  study  which 
indicated  that  dry  matter  yields  of  Coastal 
bermudagrass  were  greater  with  a short  stubble 
height  (2"),  infrequent  harvest  (14-16"  stubble) 
and  nitrogen  fertilization  (240  lbs/ac) . The 
height  of  cutting  was  less  important  in  bermuda- 
grass production  than  either  frequency  of 
defoliation  or  level  of  nitrogen  fertilization. 

It  was  concluded  that  Coastal  bermudagrass  will 
tolerate  a wide  range  of  height  and  frequency  of 
defoliation  regimens.  Clapp  et  al.  (1965) 
obtained  maximum  yields  from  Coastal  bermudagrass 
by  harvesting  at  a .75  inch  stubble  each  time  a 
2-inch  height  was  obtained.  This  treatment 


■*Texas  ASM  University  Agricultural  Research  and 
Extension  Center,  Overton  75684 


included  23  harvest  dates  during  the  season  and 
produced  the  highest  initial  yield  each  spring 
which  indicated  an  adequate  root  reserve  or  tiller 
development.  Prine  and  Burton  (1956)  and  Burton 
et  al . (1963)  reported  increases  in  bermudagrass 
yield  with  less  frequent  cutting  with  maximum 
production  occurring  at  the  6-week  harvest 
frequency.  Quality,  however,  declined  with  each 
delay  in  harvest.  Holt  and  Conrad  (1984) 
evaluated  several  bermudagrass  selections  at 
harvest  frequencies  at  3,  6,  and  9 weeks  during 
a 3-year  period.  Dry  matter  production  ranged 
from  about  5 tons/acre  for  Callie  to  nearly  7.5 
tons/acre  for  Coastal.  They  also  reported  no 
additional  increases  in  yield  after  a 6-week 
interval . 

Rouquette  and  Florence  (1981)  evaluated  the  effect 
of  1,  2,  and  4-week  harvest  intervals  on  dry 
matter  production,  vigor,  and  density  of  several 
bermudagrasses  for  a 3-year  period.  Bermudagrass 
yield  at  the  4-week  interval  was  nearly  twice 
that  of  yields  from  the  1 and  2 week  harvest 
intervals.  Although  dry  matter  production,  vigor 
and  density  rating  varied  among  bermudagrasses, 
harvest  frequency  did  not  affect  vigor  within  a 
bermudagrass  selection.  Density  ratings  were 
generally  higher  on  the  more  frequently  harvested 
plots.  Grass  vigor  and  its  ability  to  survive 
under  grazing  conditions  is  related  to  type 
(bunchgrass  vs  sodgrass)  and  its  resistance  to 
grazing  under  a given  set  of  environmental 
conditions.  Briske  (1986)  modified  the  approach 
taken  by  Levitt  (1980)  and  concluded  that  grazing 
resistance  was  dependent  upon  both  an  avoidance 
and  tolerance  mechanism.  In  this  context,  a 
tolerance  mechanism  encouraged  regrowth  following 
defoliation  and  was  dependent  upon  both  morpho- 
logical and  physiological  parameters.  An 
avoidance  mechanism,  on  the  other  hand, 
essentially  reduced  the  probability  of  defoliation 
and  was  dependent  upon  morphological  parameters 
as  well  as  biochemical  compounds. 

Leaf  to  stem  ratios  of  bermudagrass  varies  with 
the  season,  but  generally,  leaf  production  peaks 
at  20-24  days  in  early  summer  and  at  28-32  days 
in  late  summer  (Peterson,  1988) . Stem  production 
continues  to  accelerate  with  age  which  drastically 
alters  the  leaf:stem  ratio  as  well  as  nutritive 
value.  Most  of  these  clipping  studies  have 
attempted  to  critically  evaluate  defoliation 
regimens  as  they  may  impact  hay  production  and 
related  management  techniques.  Clipping  studies 
have  always  provided  certain  baseline  data  from 
which  grazing  studies  have  been  initiated. 

However,  many  have  made  direct  translations  that 
clipped  plot  data  simulates  grazing  data. 
Unfortunately,  clipped  plot  data  simulates  mowing 
or  haying  conditions  much  more  proportionally 
than  grazing  conditions.  However,  many  of  the 
previously  documented  rotational  grazing  schemes 
included  graze-rest  periods  on  multiples  of  7-day 
increments . 

GRAZING  STUDIES 

Most  of  the  ecological  and  plant  physiological 
measurements  of  rotationally  grazed  pastures  have 
been  conducted  using  temperate  grasses  and/or 


44 


clovers.  A few  examples  of  similar  research  with 
tropical  or  warm-season  grasses  will  be  used  for 
illustrative  purposes  but  do  not  represent  a 
complete  literature  review.  Bransby  (1983)  used 
Coastcross  II  bermudagrass  in  a 4-year  study  to 
evaluate  the  influence  of  variable  stocking  rate 
under  both  continuous  and  rotationally  grazed 
systems.  The  rotational  scheme  at  each  level  of 
forage  availability  consisted  of  6 paddocks  with 
residence  time  in  each  paddock  of  about  one  week 
and  recovery  time  of  about  5 weeks.  As  stocking 
rate  was  increased,  the  rotational  grazing 
method  had  an  increasing  advantage  over  contin- 
uous grazing  as  measured  by  available  forage. 

In  other  words,  at  the  same  stocking  rate,  there 
was  more  forage  available  for  consumption  in  the 
rotationally  grazed  system  as  compared  to  the 
continuously  grazed  pasture.  At  equivalent 
levels  of  forage  availability,  average  daily 
gain  (ADG)  was  greater  for  animals  on  the 
continuously  grazed  areas  compared  to  the 
rotationally  grazed  paddocks.  It  was  concluded 
that  the  lower  ADG  from  animals  assigned  to  the 
rotational  scheme  results  from  forced  consumption 
of  5-week  old  forage  which  was  low  in  quality, 
and  a possible  behavioral  problem  associated 
with  paddock  size  and  movement  schedules. 

Conrad  (1982)  grazed  both  Callie  and  Coastal 
bermudagrass  rotationally  and  continuously  at 
each  of  four  set  stocking  rates.  A 7-day  graze, 
21-day  rest  scheme  was  used  in  the  rotational 
paddocks.  Although  steers  grazing  Callie  had  a 
12%  advantage  in  ADG,  there  were  no  differences 
between  grazing  method  within  bermudagrass  hybrid 
as  measured  by  ADG  over  the  2-year  study.  The 
greatest  difference  between  hybrids  occurred  at 
the  lightest  stocking  rates  (2.7  and  3.2  hd/ac) 
where  animal  selectivity  was  optimum.  In  an 
extension  of  this  study,  Kanyama-Phiri  and 
Conrad  (1986)  evaluated  the  influence  of 
grazing  method  (continuous  vs  rotational)  and 
stocking  rate  on  grazing  behavior,  animal 
performance,  and  sward  response.  Forage-on- 
offer,  bite  rate,  and  bite  size  decreased  from 
day  1 to  day  7 of  the  residence  time  in  each 
paddock,  but  remained  relatively  constant 
across  all  stocking  rate  pastures.  Percent  leaf 
of  the  sward  under  continuous  grazing  was  nearly 
identical  across  all  four  stocking  rates  at  51% 
leaf.  Under  the  rotationally  grazed  scheme, 
percent  leaf  in  the  sward  on  day  1,  4,  and  7 of 
the  residence  period  was  72,  52,  and  50%, 
respectively,  across  all  stocking  rates.  The 
mean  percent  leaf  during  the  7-day  period  was 
57,  54,  59,  and  62%,  respectively,  for  high, 
moderately  high,  moderately  low,  and  low  stocking 
rates.  Percent  leaf  of  animal  extrusa  as 
measured  esophageally  was  about  82%  across  all 
stocking  rates  on  the  continuously  grazed 
pastures  and  about  86%  on  the  rotationally 
grazed  paddocks.  Roth  et  al.  (1985)  used 
stocker  cattle  to  graze  Coastal  bermudagrass  to 
four  levels  of  availability.  Percent  green  leaf 
in  the  sward  ranged  from  about  40%  on  the  low 
stocked  pastures  (2500  lbs  body  weight  per  acre) 
to  about  50%  on  the  high  stocked  pastures  (8000 
lbs  body  weight  per  acre) . However,  percent 
leaf  of  the  animals'  diet  was  approximately  80% 


and  was  not  affected  by  level  of  forage  avail- 
ability or  stocking  rate.  In  the  Kanyama-Phiri 
study,  percent  leaf  of  animal  extrusa  was  94,  84, 
and  80%  respectively,  on  days  1,  4,  and  7 of  the 
residence  time.  The  benefits  derived  from  the 
94%  leaf  in  the  diet  of  the  animals  on  day  1 was 
offset  by  the  lack  of  forage  availability  on 
day  7.  Thus,  this  data  may  suggest  that  if 
rotational  grazing  is  to  increase  ADG  and/or  gain 
per  acre  over  continuously  grazed  pastures,  then 
either:  (a)  residence  time  may  need  to  be 
dictated  by  forage  availability  rather  than  a 
set  number  of  days,  and/or;  (b)  more  than  one  set 
of  grazers  may  be  necessary  to  avoid  the  complete 
utilization  of  mature  bermudagrass  forage  which 
is  high  in  percent  stems  and  low  in  nutritive 
value.  Further  quantification  of  the  bermudagrass 
sward  by  Kanyama-Phiri  (1988)  showed  that  both 
percent  protein  and  in  vitro  digestible  dry 
matter  (IVDDM)  were  lower  on  the  continuously 
grazed  pastures  as  compared  to  those  grazed 
rotationally.  In  each  grazing  system,  the  sward 
was  partitioned  into  an  upper,  mid-  and  lower 
portion.  The  lower  portion  of  the  sward  was 
35-40%  lower  in  protein  and  16-18%  lower  in 
IVDDM  as  compared  to  the  upper  portion  of  the 
sward. 

Under  more  arid  conditions,  Heitschmidt  et  al. 
(1987)  used  a 16-paddock  rotational  grazing 
scheme  and  a continuously  grazed  range  area  to 
evaluate  the  effects  of  grazing  system  on 
vegetation  responses  and  cow-calf  production. 
Residence  time  in  each  paddock  ranged  from  1 to 
4 days  with  a rest  period  of  30  to  65  days.  The 
total  herbaceous  standing  crop  was  greater  in 
the  continuously  grazed  areas  because  of  a 
greater  amount  of  senesced  forage,  however, 
method  of  grazing  did  not  affect  herbage  growth 
dynamics.  Forage  quality  was  generally  higher 
in  the  rotationally  grazed  paddocks.  It  was 
concluded  that  stocking  rate  and  not  the  method 
of  grazing  was  responsible  for  differences  that 
occurred  among  treatments.  And,  further,  that 
the  rotational  grazing  scheme  influences 
carrying  capacity  (10-15%  improvement)  and 
overall  range  condition  by  enhancing  grazing 
distribution  rather  than  increasing  forage 
production.  Stuth  et  al . (1987)  simulated  a 
16-paddock,  one  herd,  short  duration  grazing 
scheme  under  four  stocking  rates  to  quantify 
various  plant-animal  interactions  on  a little 
bluestem-brownseed  paspalum  savanna.  Stocking 
rate  altered  the  composition  of  species  grazed, 
affected  the  amount  of  live  tissue  that  escaped 
grazing,  and  reduced  the  total  amount  of  lamina 
during  the  succeeding  grazing  period.  Increased 
stocking  rates  negatively  affected  the  number  of 
active  meristems  and  associated  root  systems. 

Under  a moderate  rate  of  stocking,  previous 
defoliation  and  number  of  live  leaves  influenced 
the  grazing  of  a little  bluestem  tiller,  and 
previous  grazing  history  and  amount  of  live 
lamina  influenced  the  selection  of  a brownseed 
paspalum  tiller.  Selection  of  little  bluestem 
was  influenced  by  abundance  and  morphology  of 
the  plant  with  abundance  becoming  more  important 
as  stocking  rate  increased.  Brownseed  paspalum, 
on  the  other  hand,  was  selected  primarily 


45 


because  of  morphology,  and  selection  was  not 
greatly  influenced  by  abundance  or  stocking  rate. 
As  stocking  rates  were  increased,  intake  of 
protein  was  maintained  by  an  increase  in 
selection  of  dicots.  Season  of  the  year  was  the 
most  important  factor  affecting  grazing  behavior. 


SUMMARY 

There  is  a definite  void  of  basic  physiological 
data  from  rotationally  vs  continuously  grazed 
warm-season  grasses.  Most  of  the  grazing 
studies  have  reported  definite  quality  advantages 
and  some  dry  matter  production  advantages  for 
rotationally  grazed  warm-season  grasses.  From 
the  standpoint  of  animal  performance,  however, 
little  if  any  advantages  in  ADG  have  been 
reported.  Quality  of  the  lower  portion  of  the 
warm-season  grass  sward  is  usually  dramatically 
lower  than  the  quality  of  the  upper  portion  of 
the  sward.  This  same  trend  is  not  the  case  with 
perennial  ryegrass  or  other  temperate  grasses. 
Thus,  the  forced  consumption  of  the  low  quality 
stems  in  the  lower  strata  of  the  warm-season 
grass  sward  has  probably  served  to  offset  the 
beneficial  effects  of  a rest-graze  scheme. 
Certainly,  the  research  efforts  with  perennial 
ryegrass  (Parsons  et  al.,  1988)  serve  to 
illustrate  the  need  for  similar  research  with 
tropical  forages.  In  the  most  elementary 
scenario,  as  long  as  stocking  rate  allows  for 
selective  grazing,  all  grazing  could  be 
classified  as  rotational  grazing.  However,  under 
most  commonly  accepted  definitions,  rotational 
grazing  requires  several  paddocks  (more  than  one) 
with  some  graze-rest  period.  Unfortunately,  the 
decision  to  graze  warm-season  grasses 
rotationally  is  usually  based  on  testamonial- 
type  information  rather  than  scientific  fact. 

It  is  generally  understood  that  a multi-paddock 
divided  farm  has  more  flexibility  than  a single 
paddock.  And,  this  is  especially  true  when 
optimum  forage  utilization  is  to  be  attained 
either  from  grazing  and/or  mechanical  harvesting. 
Certainly  the  dramatic  quality  differences 
within  the  sward  of  the  warm-season  grass,  and 
the  stubble  height  or  forage  availability  of  the 
residence  paddock  are  primary  factors  to 
consider  before  entering  into  the  economic  and 
time  demands  of  a rotational  grazing  scheme. 

The  primary  objective  of  the  grazier  of  warm- 
season  grasses  is  to  make  optimum  biological 
and  economic  utilization  of  forage  that  is 
produced,  and  at  the  same  time,  to  prevent  the 
deterioration  of  the  soil-plant  resource. 

Literature  Cited 

Bransby,  D.  I.  1983.  Herbage  availability  as  a 
stress  factor  on  grazed  Coastcross  II 
bermudagrass . S.  Afr.  J.  Anim.  Sci. 

13:3-5. 

Briske,  D.  D.  1986.  Plant  response  to  defol- 
iation: Morphological  considerations  and 
allocation  priorities,  pp.  425-427  In 
P.  J.  Joss,  P.  W.  Lynch  and  0.  B.  Williams 
(eds.)  Rangelands:  A Resource  Under  Seige. 


Australian  Acad,  of  Science,  Canberra, 

Burton,  G.  W.  and  R.  H.  Hart.  1963.  Effect  of 
utilizing  frequency  on  yield,  in  vitro 
digestibility,  and  protein,  fiber,  and 
carotene  content  of  Coastal  bermudagrass. 
Agron.  J.  55:500-502. 

Clapp,  J.  G.,  Jr.,  D.  S.  Chamblee,  and  H.  D. 

Gross.  1965.  Interrelationships  between 
defoliation  systems,  morphological  character- 
istics, and  growth  of  Coastal  bermudagrass. 
Crop  Sci.  5:468-471. 

Conrad,  B.  E.  1982.  Rotational  vs  continuous 
grazing  bermudagrass  types.  Forage  Res.  in 
Texas  CPR  4024:23-24. 

Heitschmidt,  R.  K. , S.  L.  Dowhower,  and  J.  W. 

Walker.  1987.  Some  effects  of  a rotational 
grazing  treatment  on  quantity  and  quality  of 
available  forage  and  amount  of  ground  litter. 
J.  Range  Mgt.  40:318-321. 

Holt,  E.  C.  and  B.  E.  Conrad.  1984.  Response  of 
experimental  bermudagrass  hybrids  and 
cultivars  to  defoliation  frequency.  Forage 
Res.  in  Texas  CPR  4253:118-128. 

Holt,  E.  C.  and  J.  A.  Lancaster.  1968.  Yield 
and  stand  survival  of  'Coastal'  bermudagrass 
as  influenced  by  management  practices.  Agron. 
J.  60:7-11. 

Kanyama-Phiri , G.  1988.  Personal  communication. 

Kanyama-Phiri,  G.  and  B.  E.  Conrad.  1986. 

Effects  of  stocking  rate  on  forage-on-offer, 
bite  rate,  bite  size,  bite  quality,  and 
animal  weight  gain  on  warm  season  pastures. 
Forage  Res.  in  Texas  CPR  4499:41-43. 

Levitt,  J.  1980.  Responses  of  plants  to 

environmental  stresses.  I.  Chilling,  freez- 
ing and  high  temperature  stresses.  Academic 
Press,  New  York. 

Parsons,  A.  J.,  I.  R.  Johnson,  and  A.  Harvey. 

1988.  Use  of  a model  to  optimize  the  inter- 
action between  frequency  and  severity  of 
intermittent  defoliation  and  to  provide  a 
fundamental  comparison  of  the  continuous  and 
intermittent  defoliation  of  grass.  Grass  and 
Forage  Science  43:49-59. 

Parsons,  A.  J.,  I.  R.  Johnson,  and  J.H.H.  Williams. 
1988.  Leaf  age  structure  and  canopy  photo- 
synthesis in  rotationally  and  continuously 
grazed  swards.  Grass  and  Forage  Science 
43:1-4. 

Parsons,  A.  J.  and  P.  D.  Penning.  1988.  The 

effect  of  the  duration  of  regrowth  on  photo- 
synthesis, leaf  death  and  the  average  rate  of 
growth  in  a rotationally  grazed  sward.  Grass 
and  Forage  Science  43:15-27. 

Peterson,  Martha  A.  1988.  Comparison  of  double- 
sampling techniques  for  estimating  forage 
production.  M.S.  Thesis.  Texas  A$M  Univ.  83  p. 


46 


Prine,  G.  M,  and  G.  W.  Burton,  1956.  The  effect 
of  nitrogen  rate  and  clipping  frequency  upon 
the  yield,  protein  content  and  certain 
morphological  characteristics  of  Coastal 
bermudagrass  (Cynodon  dactylon  (L.)  Pers.). 
Agron.  J.  48:296-301. 

Roth,  L.  D.,  F.  M.  Rouquette,  Jr.,  and  W,  C. 

Ellis.  1985.  Diet  selection  and  nutritive 
value  of  Coastal  bermudagrass  as  influenced 
by  grazing  pressure.  Forage  Res.  in  Texas 
CPR  4347:7-9. 

Rouquette,  F.  M. , Jr.  and  M.  J.  Florence.  1981. 
Dry  matter  production  and  vigor  of  bermuda- 
grass selections.  Forage  Res.  in  Texas 
Dept.  Tech.  Report  81-12:112-120. 

Stuth,  J.  W.,  J.  R.  Brown,  P.  D.  Olson,  M.  R. 

Araujo,  and  H.  D.  Aljoe.  1987.  Effects  of 
stocking  rate  on  critical  plant-animal 
interactions  in  a rotationally  grazed 
Schizachyrium-Paspalum  savanna,  p.  115-139. 
In  F.  P.  Horn,  J.  Hodgson,  J.  J.  Mott,  and 
R.  W.  Brougham  (eds.)  Grazing-lands  research 
at  the  plant-animal  interface.  Winrock 
Int'l. 


CARBOFURAN  FOR  FORAGE  ESTABLISHMENT:  AN 
UPDATE 

D.  D.  Wolf1 

Seldom  can  the  "bottom  line"  be  stated  in 
the  beginning.  This  situation  may  be 
appropriate  as  far  as  possibilities  for 
future  approval  of  carbofuran  that  can  be 
used  for  establishment  of  forages  other 
than  alfalfa.  The  new  label  for  car- 
bofuran includes  the  entire  United  States 
as  far  as  alfalfa  establishment  is 
concerned.  Either  broadcast  spray  or 
granules  can  be  used  with  the  restriction 
that  the  material  is  incorporated.  No- 
till plantings  where  granular  carbofuran 
is  applied  in  the  row  with  the  seed  has 
been  interpreted  to  be  adequate  incor- 
poration. Regardless  of  the  benefits 
that  can  be  proven  by  research  for  other 
forage  there  will  probably  never  be 
approval  for  additional  usage.  Everyone 
will  agree  that  the  hazards  with  new 
crops  are  no  greater  than  with  other 
crops  with  current  approval;  however,  due 
to  the  extremely  high  cost  to  do  the 
investigative  research  and  the  low 
potential  return,  companies  are  not 
willing  to  invest  in  procedures  for 
expanding  registration.  State  and 
federal  government  agencies  prohibit 
written  recommendations  that  are  not 
specifically  stated  as  approved  in 
registration. 

Some  possibilities  for  approval  of 
carbofuran  use  on  new  crops  may  exist. 
Since  warm-season  grasses  will  not  be 
used  for  grazing  for  one  year,  a non-food 
crop  designation  could  be  used.  This  may 
free  us  of  establishment  of  tolerances 
and  need  for  toxicology  information.  If 
considered  as  a non-food  crop  then  a 
special  needs  (24C)  or  an  experimental 
use  (Sec.  18)  registration  may  be 
possible.  Also  there  is  a provision  for 
a "third-party"  label.  This  means  that  a 
group  such  as  a state  Forage  and  Grass- 
land Council  could  request  24C  registra- 
tion. At  present  there  is  a serious 
effort  to  find  some  way  to  gain  legal 
registration  for  carbofuran  use  in  forage 
seedings  where  beneficial.  Such  food 
crops  as  peanuts,  sweet  corn,  and 
potatoes  have  registered  usage  even 
though  they  are  consumed  by  humans  in  the 
year  of  planting  yet  no  possibility 
exists  for  a crop  to  be  consumed  by 
livestock  in  the  year  after  planting. 

The  mode  of  action  of  carbofuran  where 
beneficial  to  initial  growth  and  subse- 
quent productivity  of  the  plant  has  not 


department  of  Agronomy,  Virginia 
Polytechnic  Institute  and  State  Univer- 
sity, Blacksburg,  VA  24061 


been  confirmed.  Often  some  insects  both 
soil  borne  and  above  ground  can  be 
reduced  by  the  use  at  planting,  however, 
when  other  insecticides  are  used  to 
control  the  insects  there  is  an  unex- 
plained additional  benefit  from  car- 
bofuran. Often  this  benefit  appears  to 
be  in  the  nature  of  a plant  growth 
stimulus  that  can  occur  early  in  germina- 
tion, as  is  seen  by  increased  seedling 
population,  and  can  carry  into  yield 
productivity  of  the  mature  plant. 


ALFALFA. 

Benefits  from  carbofuran  in  alfalfa  es- 
tablishment are  not  consistent.  Recom- 
mendations vary  among  states.  Some 
states  make  no  recommendation  for  usage 
but  indicate  that  there  can  be  some 
benefits  if  used.  Other  states  make 
recommendations  for  usage  in  both 
conventional  and  no-till  alfalfa  plant- 
ings. Benefits  are  most  consistent  when 
no-till  plantings  are  made  into  a field 
where  the  previous  cover  has  been  some 
type  of  perennial  vegetation  as  compared 
with  a previous  crop  of  small  grain  or 
millet.  Dramatic  benefits  from  car- 
bofuran can  occur  where  infestation  of 
seed  corn  maggot  exists.  There  is  some 
indication  that  benefits  have  been  most 
prevalent  where  alfalfa  is  planted  into 
cool  moist  soil.  These  data  indicated 
that  carbofuran  may  be  acting  as  a 
fungicide  in  the  control  of  diseases  in 
the  seedling  stage. 


COOL-SEASON  GRASSES. 

Establishment  of  endophyte-free  tall 
fescue  has  sometimes  been  less  successful 
with  weaker  stands  occurring  as  compared 
with  endophyte-infected  tall  fescue.  A 
large  planting  made  in  southeast  Virginia 
had  adequate  stands  of  endophyte  infected 
tall  fescue,  however,  in  a very  similar 
planting  of  the  experiment  endophyte-free 
fescue  was  a failure.  Where  endophyte- 
free  tall  fescue  failed  there  were  places 
where  the  planting  equipment  passed 
through  cow  paths.  The  clear  areas 
without  vegetation  had  adequate  stands 
which  lead  to  the  suspicion  that  either 
disease  or  insects  was  contributing  to 
the  failure.  A benefit  of  carbofuran  for 
tall  fescue  plant  population,  plant 
height,  and  stand  rating  was  observed  in 
unpublished  data  at  the  University  of 
Maryland  (Table  1).  The  level  of 
endophyte  was  not  established  but 
probably  the  seed  was  infected. 
Unpublished  data  from  the  University  of 
North  Carolina  indicate  an  advantage  in 
seedling  population  resulting  from  the 
use  of  carbofuran  at  each  of  three 


48 


planting  dates  (Table  2).  The  influence 
on  subsequent  productivity  is  not 
available  from  these  experiments  but  as 
with  many  plantings  where  early  stimula- 
tion of  seedling  growth  has  occurred, 
compensatory  plant  development  soon  masks 
any  early  advantage  in  the  seedling 
stages  of  development.  Recommendation 
can  not  be  made  due  to  the  label  restric- 
tion for  the  use  of  carboufran. 

Table  1.  Plant  population,  plant 
height,  and  stand  of  tall  fescue 
and  switchgrass  as  influenced  by 
carbofuran.  Unpublished  data  from 
Morris  Decker  Maryland,  1981. 

Data  taken  2 months  after  seeding, 
average  of  3 planting  dates. 


Species 

Carbo 
fur  an 

Pop. 

Height 

Stand 

Switchgrass 

No 

9.2 

11.4 

1.3 

Yes 

10.2 

14.4 

2.6 

Tall  fescue 

No 

14.4 

7.6 

1.2 

Yes 

17.8 

9.4 

2.2 

Statististics  not 

available . 

-'-Population 

= no. 

/ft2. 

height 

= inche 

Stand  - rating  with  1 being  low  and  5 
high. 

Table  2.  Plant  population  of  Forager 
tall  fescue  planted  at  3 dates  as 
influenced  by  carbofuran  drilled  in 
the  row  with  the  seed  at  1.7  kg  ha--*-. 
(Unpublished  data  from  Doug  Chamblee, 
University  of  North  Carolina). 


Planting  Carbofuran 


date 

No 

Yes 

Dif . 

- No . 

ft-2  - 

1 Sept . 

3.0 

7.0 

4.0 

1 Oct. 

6.0 

14.5 

8.5 

21  Oct. 

12.0 

14.0 

2.0 

Avg 

7.0 

11.8 

00 

WEEDS. 

Phenomenal  advantages  of  carbofuran  in 
the  growth  and  development  of  some  weeds 
such  as  crabgrass,  fall  panicgrass, 
Johnsongrass , and  spiney  pigweed  have 
been  observed.  These  responses  have  not 
been  documented  in  the  literature  but 
observed  by  Extension  Specialist  and 
research  workers  under  field  situations. 
The  stimulus  of  weeds  could  lead  to 
serious  problems  if  competition  is 
greater  for  the  planted  crop  where 
carbofuran  is  applied  than  without 


application.  Furthermore,  established 
stands  where  carbofuran  is  used  as  an 
insecticide  could  be  counter  productive 
in  stimulating  the  competition  of 
existing  weeds.  There  is  some  evidence 
to  indicate  that  warm-season  C-4  plants 
are  the  most  likely  to  be  stimulated  by 
the  presence  of  carbofuran. 


TALL-GROWING  PERENNIAL  WARM-SEASON 
GRASSES. 

The  dramatic  benefit  of  carbofuran  when 
used  with  warm-season  grasses  such  as 
switchgrass  and  Caucasian  bluestem  has 
been  observed  and  documented  by  many 
research  and  Extension  workers  (Table 
2,3).  The  use  of  either  broadcast  spray 
or  granular  formulation  with  the  seed  is 
effective  in  increasing  needling  popula- 
tion and  vigor  of  the  seedlings.  Lorsban 
insecticide  did  not  increase  plant 
population  and  had  only  moderate  in- 
fluence on  vigor.  Therefore,  increased 
population  from  carbofuran  could  be 
interpreted  to  be  due  to  non-insecticidal 
benefits  while  vigor  of  the  seedling 
could  be  considered  as  about  66%  of  the 
increase  that  could  be  attributed  to 
carbofuran  (Table  3).  Some  extreme  weed 
competition  has  occurred  from  warm-season 
annual  grasses  as  mentioned  above  which 
may  have  been  in  part  due  to  the  stimulus 
of  carboufran. 

Table  3.  Switchgrass  (Pathfinder  seed- 
ling population  and  vigor  as  influe- 
nced by  Furadan  and  Lorsban  placed 
in  row  as  granules  at  0,1,  and  2 
lb. /acre  and  Furadan  broadcast  at  2 
lb. /acre.  Data  are  averages  of  no-till 
plantings  on  2 and  15  May  1987  using 


4.2  lb. 

PLS/acre 

• 

Insect- 

Place- 

Rate 

( lb. /acre) 

icide 

ment 

0 

1 

2 

Furadan 

In  row 
Brdcast 

Pop.  ( 
17.7b* 
17.7b 

No. /sq. 
25.8a 

ft.  ) 

25.0a 

26.2a 

Lorsban 

In  row 

17.7b 

17.8b 

17.2b 

Furadan 

In  row 
Brdcast 

Vigor  rating 

2.7c  9.3a  9.2a 

2.7c  - 9.0a 

Lorsban 

In  row 

2.7c 

4.8b 

5.4b 

*Means  followed  by  similar  letters  do  not 
differ  significantly  at  the  0.05  level. 


49 


Cave-in-rock  switchgrass  was  no-till 
planted  on  22  April,  15  May,  and  15  June 
in  a 1987  study.  Seedling  population  was 
more  than  doubled  with  the  application  of 
Furadan  as  compared  with  the  check  at  all 
seeding  dates.  Seedling  weights  from 
carbofuran  treated  plots  were  about  3 
times  greater  than  seedlings  without 
carbofuran  (Table  4). 


Table  4.  Switchgrass  seedling  popula- 


tion  and  weight  (at  5th  leaf  stage)  as 
influenced  by  carobufran  and  planting 
date  in  1987. 

Planting 

Carbofuran( lb/a) 

dates 

0 1 

22  April 

15  May 

15  June 
Avg. 

pop.  (no. /ft'2) 

5.0  12.5 

6.8  13.3 

6.6  12.6 

6.1  12.8 

22  April 

15  May 

15  June 

mg/seedling 

24  86 

19  56 

66  166 

Excerpts  from  a thesis  recently  completed 
by  Jim  McKenna  at  Virginia  Polytechnic 
Institute  and  State  University  will  be 
presented  in  order  to  document  the 
responses  of  switchgrass  and  Caucasian 
bluestem  to  carbofuran  at  seeding. 

Plantings  of  two  tall-growing,  perennial, 
warm-season  grasses  were  made  in  Black- 
sburg, VA  on  a Groseclose  loam  soil 
(clayey,  mixed,  mesic  Typic  Hapludult). 
Establishment  of  switchgrass  and  Cauca- 
sian bluestem  using  no-till  procedures 
was  evaluated  with  treatments  of  car- 
bofuran. Seedling  growth  rate  and  leaf 
appearance  rate  were  recorded  prior  to 
the  sixth-leaf  stage  of  development. 
Seedling  weights,  populations,  and 
heights  were  measured  at  the  sixth-leaf 
stage  of  development.  Leaf  elongation 
rates  were  measured  for  leaves  7,  8,  and 


9.  Yields  of  forage  and  percentage 
perennial  warm-season  grass  in  the 
harvested  herbage  were  determined  in  the 
year  of  planting  and  the  year  after 
planting.  Our  data  indicate  that  1.1  kg 
carbof uran/ha , placed  in  the  row  with  the 
seed  at  the  time  of  no-till  planting 
enabled  seedlings  to  develop  faster, 
elongate  more  rapidly,  and  provide  more 
and  heavier  seedlings  than  without 
carbofuran.  Carbofuran  at  the  time  of 
planting  increased  yields  in  with  both 
species  (Table  5).  Carbofuran  at  1.1 
kg/ha  was  considered  the  best  recommenda- 
tion for  establishment  of  switchgrass  and 
Caucasian  bluestem. 

Table  5.  Influence  of  carbofuran  applied 
in  the  row  with  the  seed  of  switchgrass 
and  Caucasian  bluestem  on  seedling  leaf 
elongation  rate  (LER),  seedling 
population,  seeding  weight,  and  herbage 
yield  from  Ph.D.  thesis  by  Jim  McKenna, 
VPI  & SU,  Blacksburg,  VA. 


Species-*- 

Carbo- 

furan 

Seedling 

LER  Pop. 

2 

Wt. 

Yield 
first  cut 

bwitch- 

grass 

0 

26 

201 

132 

0.73 

1.1 

43 

361 

338 

1.56 

Dif . 

* * 

* * 

** 

* * 

C.  Blue- 

stem 

0 

51 

207 

50 

0.93 

1.1 

70 

612 

74 

2.39 

Dif. 

* * 

* * 

* * 

* * 

■*-Switchgrass  data  are  averages  of  1985 

and  1986  plantings.  Caucasian  bluestem 
data  are  from  a 1985  planting. 


2LER  = mm/d.  Pop.  = No./m2,  wt  = mg/pl, 
yield  = Mg/ha. 

**Indicates  a significant  difference 
between  carbofuran  rates  at  the  0.01 
level  or  less. 


50 


FORAGE  UTILIZATION  INFORMATION  EXCHANGE  GROUP 

USE  OF  NIRS  TO  PREDICT  BOTANICAL  COMPOSITION  OF 
FORAGE  MIXTURES 

N.  S.  Hill,  J.  A.  Stuedemann,  and  G.  0.  Ware—^ 
INTRODUCTION 

By  now  all  of  us  should  be  aware  of  the  benefits 
of  rapid  and  accurate  analysis  of  feedstuffs 
using  near  infrared  reflectance  spectroscopy 
(NIRS).  For  those  of  you  who  have  not  been 
associated  with  NIRS  analysis  and  wish  to 
become  more  familiar  with  the  technology,  the 
USDA-ARS  Agricultural  Handbook  No.  643,  "Near 
Infrared  Reflectance  Spectroscopy  (NIRS): 
Analysis  of  Forage  Quality",  is  an  excellent 
beginning  manual  to  review.  The  first  printing 
is  no  longer  available  but  it  is  being  revised 
and  the  second  printing  should  be  available 
soon . 

The  majority  of  the  NIRS  research  analyzing 
forages  for  feed  have  been  with  conserved 
forages  for  feed  formulation  (Martin  and  Linn, 
1985;  Norris  et  al . , 1976;  Shenk  et  al . , 1979). 
The  homogeneous  and  static  nature  of  conserved 
forages  made  the  feedbunk  an  attractive  arena 
in  which  animal  nutrition  and  forage  quality 
scientists  could  begin  evaluating  NIRS  technol- 
ogy and  its  applicability  to  the  plant-animal 
interface.  For  example,  monthly  testing  of 
forage  inventories,  and  the  resulting  feed 
formulations,  have  increased  protein  and  fat 
content  from  four  Minnesota  dairy  farms  without 
adversely  affecting  milk  production.  In  that 
same  study,  farmers  realized  about  $2,000  in 
savings  on  feed  cost  by  reduced  protein 
supplementation  (Walter  et  al.,  1987). 

THE  PASTURE-ANIMAL  INTERFACE 

Matching  animal  needs  with  forage  quality  in 
the  pasture  environment  is  a more  crude  estim- 
ation because  scientists  don't  have  the  luxury 
of  working  with  forages  that  are  homogeneously 
mixed  (as  in  silages)  or  whose  quality  changes 
little  with  time.  Brown  et  al.  (1987)  described 
the  pasture  plant-animal  interface  as  being 
dynamic,  whose  changes  are  a function  of  animal 
defoliation  of  the  pasture  canopy,  and  the 
changing  pasture  canopy  effects  on  ingestive 
behavior  of  the  animal.  The  changes  which  occur 
in  the  pasture  during  defoliation  are  most 
evident  in  grass-legume  mixtures  of  rotational 
grazing  systems.  Animals  that  rotationally 
grazed  stargrass-aeschynomene  pastures  preferen- 
tially grazed  leaves  over  stems  and  aeschynomene 
over  stargrass  (Brown  et  al.,  1987).  In  a 
similar  experiment  where  animals  rotationally 
grazed  alfalfa-orchardgrass  pastures,  it  was 
found  that  pasture  nutrition  was  high  at  the 
onset  of  grazing  but  considerably  lower  at  the 


_1 / Assistant  Professor  of  Agronomy,  USDA-ARS 
Research  Animal  Scientist,  Professor  of  Forest 
Resources  and  Experiment  Station  Statistician, 
University  of  Georgia,  Athens,  GA. 


end  of  the  grazing  cycle  (Blazer  et  al . , 1986, 
1977).  The  relationships  between  selectivity, 
digestibility,  intake,  and  grazing  pressure 
from  rotationally  grazed  pastures  are  summarized 
in  Table  1.  From  this  it  becomes  evident  that 
under  low  grazing  pressures,  when  selectivity, 
intake,  and  digestibility  of  the  grazed  forage 
is  high,  animal  performance  is  expected  to  be 
high.  Conversely,  when  grazing  pressure  is 
high,  selectivity,  intake,  and  digestibility 
are  low  and  animal  performance  is  expected 
to  be  lower.  In  practice,  there  is  a lag  in 
animal  performance  when  animals  are  moved  from 
low  quality  pasture  to  high  quality  pasture  in 
a rotational  grazing  system.  This  was  demon- 
strated with  milk  cows  which  rotationally 
grazed  pasture  cells  for  eight  days.  After 
rotating  to  fresh  pasture,  a 2-3  day  lag 
occurred  until  milk  production  peaked  after 
which  production  steadily  declined  (Blazer  et 
al.,  1986).  Therefore,  performance  of  the 
pasture  animal  is  a function  of  available 
herbage,  the  ability  of  the  animal  to  obtain 
a quality  diet  by  selecting  more  nutritious 
plant  species  and  plant  components  of  the 
species  in  the  pasture,  and  the  residual  effect 
of  lower  quality  digesta  in  the  rumen  after 
changing  pastures.  The  problem  facing  pasture 
and  animal  scientists  is  how  to  minimize 
nutritional  fluctuations  through  pasture  and 
animal  management  and  match  available  nutrition 
of  the  pasture  with  nutritional  demands  by  the 
amimal  without  sacrificing  complete  utilization 
of  the  pasture  herbage.  Experiments  designed 
to  study  the  plant-animal  interface  in  the 
pasture  require  extensive  sampling,  chemical, 
and  botanical  analysis  to  accurately  determine 
the  sequence  of  events  which  result  in  animal 
performance.  Therefore,  analysis  by  NIRS  is 
an  attractive  alternative  laboratory  analysis 
because  of  the  expeditious  manner  in  which  it 
analyzes  large  volumes  of  samples. 

NIRS  CONCEPTS 

Near  infrared  analysis  of  samples  is  based  upon 
absorption,  re-radiation,  and  reflection  of  near 
infrared  light  by  the  functional  groups  of 
molecules  which  are  components  of  chemical 
constituents  in  the  forage  (Norris,  1985). 

These  may  be  peptide,  carboxyl,  acetyl, 
hydroxyl,  aldehyde,  or  ketone  groups,  to  name  a 
few.  Therefore,  NIRS  methodology  is  best  suited 
for  chemical  analysis  of  pasture,  esophageal, 
and  fecal  samples.  However,  a series  of 
epxeriments  reported  in  the  literature  suggests 
NIRS  can  also  be  used  to  estimate  components 
of  herbage  and  esophageal  samples  which  do 
not  have  specific  and  unique  chemical  constitu- 
ents . 

Coleman  et  al . (1985)  were  the  first  to  test 
the  hypothesis  that  species  composition  of  hay 
could  be  predicted  using  NIRS  methodology. 

Using  artificial  mixtures  of  cool  and  warm 
season  grass  species  and  a dicot  weed,  they 
demonstrated  that  NIRS  could  accurately  and 
precisely  predict  components  of  the  mixtures. 


51 


Eq.  [1] 


Ifi  a second  experiment  where  samples  of  pure 
hay  were  taken  from  feed  troughs  to  create 
artificial  mixtures  for  NIRS  calibration  and 
prediction,  precision  (and  in  one  species 
accuracy)  was  reduced.  This  suggests  that 
extreme  care  must  be  taken  to  properly  sample 
the  units  to  be  estimated  by  NIRS. 

A second  experiment  which  demonstrated  the 
usefulness  of  NIRS  analysis  in  predicting 
botanical  composition  was  that  of  Petersen  et 
al . (1987)2  They  found  that  pasture  samples 
from  0.1  m quadrats  harvested  in  May  and 
August  could  be  analyzed  for  tall  fescue  and 
white  clover  when  hand  separations  of  pasture 
samples  were  used  to  calibrate  the  NIRS 
instrument.  Their  study  suggests  that  proper 
selection  of  a proportion  of  the  samples  is 
necessary  to  calibrate  NIRS  for  accurate 
prediction  of  botanical  composition  in  pasture 
samples . 

Thirdly,  it  has  been  demonstrated  that  leaf 
components  in  alfalfa  samples  can  be  accurately 
and  precisely  predicted  by  NIRS  (Hill  et  al.  , 
1987).  When  saliva  was  added  to  the  samples 
in  varying  concentratons , there  was  little 
effect  on  the  ability  of  a NIRS  calibration  to 
predict  leaf  in  samples  without  added  saliva  and 
vice  versa.  In  fact,  the  same  math  treatments 
of  the  spectral  data  were  selected  for  calibra- 
tion equations  developed  for  samples  with  or 
without  added  saliva,  and  of  a possible  eight 
wavelength  matches  between  the  two  equations, 
seven  were  identical  (Table  2).  This  suggests 
that  not  only  can  NIRS  be  used  to  predict 
botanical  composition  and  plant  components  in 
pasture  samples,  but  that  the  same  measurements 
can  be  made  in  esophageal  samples  from  a 
calibration  equation  generated  from  pasture 
samples . 

ERRORS  ASSOCIATED  WITH  NIRS  ANALYSIS 

As  with  any  laboratory  analytical  procedure, 

NIRS  has  error  associated  with  its  analysis. 

When  predicting  chemical  or  botanical  constit- 
uents of  the  sample,  two  sources  of  error  can 
be  measured.  The  first,  the  bias,  is  an 
estimation  of  how  accurately  NIRS  is  predicting 
the  laboratory  value  of  the  sample.  Therefore, 
it  can  be  used  as  an  evaluation  of  whether  NIRS 
systematically  over-  or  under-predicts  the 
laboratory  value  of  the  samples.  The  second, 
the  standard  error  of  prediction  (SEP)  is  an 
estimation  of  the  precision  with  which  NIRS 
predicts  the  laboratory  value  of  the  sample. 

The  SEP,  although  called  a "standard  error", 
is  the  square  root  of  the  mean  square  error, 
or  standard  deviation,  between  the  NIRS  analysis 
and  laboratory  values  (Martin  and  Linn,  1985). 
Therefore,  the  SEP  can  be  used  to  calculate  a 
boundary  on  the  error  of  the  estimation  using 
Eq.  [1]  (Mendenhall,  1987). 


where:  t = t value  based  upon  NIRS  sample  size 

used  to  determine  the  SEP 

0 = SEP 

n = pasture  sample  size 

B = boundary 

For  example,  if  NIRS  is  calibrated  for  a 
pasture  species  component  using  51  samples, 
and  has  an  SEP  of  4.50%,  the  boundary  of  a 
single  observation  (n  = 1)  is: 

2.008(4.50%)  = 9.04%  Eq.  [2] 

1 

Having  a boundary  of  9.04%  may  not  be  accepta- 
ble, particularly  when  it  may  be  as  high  as 
the  mean.  So  the  logical  question  is,  "How 
does  error  associated  with  NIRS  analysis  affect 
the  estimation  of  my  pasture  components?".  The 
answer  is  that  the  boundary  of  the  NIRS 
analysis  is  a funcation  of  the  number  of 
samples,  or  repeated  observations,  you  obtain 
to  estimate  your  pasture.  Using  Eq.  [1]  one 
can  calculate  the  boundary  of  the  NIRS  mean 
value.  To  demonstrate  the  concept,  consider 
the  following  experiment. 

In  1987,  a study  was  initiated  to  determine 
diet  selection,  digestibility,  and  intake  of 
steers  rotationally  grazing  alfalfa,  tall 
fescue,  and  two  alfalfa-tall  fescue  mixtures. 
Stratified  pasture  samples  were  obtained  to 
provide  a three  dimensional  analysis  of 
botanical  composition  and  the  plant  components 
(alfalfa  leaf,  stem  and  tall  fescue  leaf). 
Because  of  the  labor  required  for  separation 
and  chemical  analysis,  NIRS  methodology  was 
chosen  to  analyze  the  samples. 

2 

Each  day  of  the  experiment  8-0. 5m  samples 
were  clipped  from  each  pasture  and  a portion 
(about  20%  of  the  total  sample  number)  were 
hand  separated  into  alfalfa  leaf,  alfalfa  stem, 
and  tall  fescue  fractions.  The  fractions  were 
weighed  to  obtain  the  percentage  of  each,  then 
remixed,  ground  through  a cyclone-type  mill, 
and  scanned  by  NIRS.  Two-thirds  of  the  hand 
separated  samples  were  used  to  calibrate  NIRS 
and  one-third  used  to  validate  the  calibration. 
The  NIRS  validation  statistics  are  provided  in 
Table  3.  By  knowing  the  number  of  samples  in 
the  validation  set  and  the  SEP  for  each 
component,  the  boundary  of  the  NIRS  analysis 
can  be  determined  if  one  were  to  sample  1,  5, 

8,  or  10  samples  per  pasture  using  Eq.  [1]. 

From  the  calculations  (Table  3),  it  becomes 
evident  that  the  boundary  of  the  NIRS  estimate 
is  inversely  related  to  the  number  of  samples 
obtained  to  estimate  the  pasture. 

Because  biases  of  the  NIRS  predictions  were 
small,  the  mean  of  the  NIRS  predictions  would 
be  expected  to  lie  within  the  NIRS  boundary  of 
the  mean  of  the  hand  separated  values  for  the 
!,  5,  8 or  10  pasture  samples.  Comparisons  of 
NIRS  predicted  and  laboratory  means  for  the 


52 


pasture  components  among  samples  were  similar 
(Table  4).  In  addition,  standard  deviations 
of  the  pasture  samples  for  both  NIRS  and 
laboratory  values  were  similar  suggesting  the 
variability  between  pasture  samples  had  a 
greater  effect  on  variability  among  NIRS 
predictions  than  error  associated  with  the 
spectral  analysis. 


CONCLUSIONS 

NIRS  can  be  used  to  predict  chemical  or 
botanical  composition,  and  plant  components  of 
pasture  and  esophageal  samples.  Because  the 
accuracy  and  precision  of  the  prediction  are 
dependent  upon  how  the  calibration  subset  is 
selected  from  the  entire  sample  population, 

NIRS  will  not  totally  replace  laboratory 
analysis  of  the  samples.  Therefore,  NIRS  is 
best  suited  for  experiments  where  large  sample 
numbers  are  generated. 

The  accuracy  of  the  prediction  is  never 
guaranteed  to  be  below  the  boundary  for  NIRS 
predicted  means  of  the  pasture  samples.  It 
is  possible  that  the  error  estimate  will  exceed 
the  bound,  but  depending  upon  the  probability 
designated  when  determining  the  "t"  statistic, 
it  is  unlikely. 

REFERENCES 

Blazer,  R.  E.  and  Colleagues.  1986.  Forage- 
animal  management  systems.  Virginia  Agricul- 
tural Exp.  Sta.  Bulletin  86-7.  VPI  and  SU , 
Blacksburg,  VA. 

Blazer,  R.  E.,  W.  C.  Stringer,  E.  B.  Rayburn, 

J.  P.  Fontenot,  R.  C.  Hammes , Jr.,  and  H.  T. 
Bryant.  1977.  Increasing  digestibility 
and  intake  through  management  of  grazing 
systems.  pp . 301-345.  In  J.  A.  Stuedemann 
(ed.)  Forage-fed  beef:  Production  and  market- 
ing alternatives  in  the  South.  Southern 
Cooperative  Series  Bulletin  220. 

Brown,  W.  F.,  J.  E.  Moore,  and  P.  Mislevy. 

1987.  Managing  the  plant-animal  interface  in 
tropical  grass-legume  pasture,  pp.  45-48.  In 
Proc . 43rd  Southern  Pasture  and  Forage  Crop 
Improvement  Conference.  April  20-22,  1987. 
Clemson,  SC. 

Coleman,  S.  W. , F.  E.  Barton,  II,  and  R.  D. 
Meyer.  1985.  The  use  of  near  infrared 
reflectance  spectroscopy  to  predict  species 
composition  of  forage  mixtures.  Crop  Sci. 
25:834-837. 

Hill,  N.  S.,  J.  C.  Petersen,  J.  A.  Stuedemann, 
and  F.  E.  Barton,  II.  1988.  Prediction  of 
percentage  leaf  in  stratified  canopies  of 
alfalfa  with  near  infrared  reflectance  spec- 
troscopy. Crop  Sci.  28:354-358. 


Martin,  N.  P.,  and  J.  G.  Linn.  1985. 

Extension  applications  in  NIRS  technology 
transfer.  pp.  48-53.  Ik}.  Mart  in . G.  C.,  J.  S. 
Shenk,  and  F.  E.  Barton,  II  (ed.)  Near 
infrared  reflectance  spectroscopy  (NIRS): 
Analysis  of  forage  quality.  USDA-ARS 
Agricultural  Handbook  643.  Washington,  DC. 

Norris,  K.  H.  1985.  Definition  of  NIRS 
analysis.  pp . 6.  Ln  G.  C.  Marten,  J.  S. 

Shenk,  and  F.  E.  Barton,  II  (ed.)  Near 
infrared  reflectance  spectroscopy  (NIRS): 
Analysis  of  forage  quality.  USDA-ARS  Agric. 
Handbook  643.  Washington,  DC. 

Norris,  K.  H.,  R.  F.  Barnes,  J.  E.  Moore, 
and  J.  S.  Shenk.  1976.  Predicting  forage 
quality  by  near  infrared  reflectance  spec- 
troscopy. J.  Animal  Sci.  43:889-897. 

Petersen,  J.  C.,  F.  E.  Barton,  II,  W.  R. 
Windham,  and  C.  S.  Hoveland.  1987. 

Botanical  compositoin  definition  of  tall 
fescue-white  clover  mixtures  by  near  infrared 
reflectance  spectroscopy.  Crop  Sci. 

27:  1077-1080. 

Shenk,  J.  S.,  M.  0.  Westerhaus,  M.  R.  Hoover. 
1979.  Analysis  of  forages  by  infrared 
reflectance.  Dairy  Sci.  62:807-812. 

Snedecor , G.  W.,  and  W.  G.  Cochran.  1967. 
Design  and  analysis  of  sampling.  Chapter  17. 
Statistical  Methods.  Iowa  State  Univ.  Press. 
Ames , Iowa . 

Walter,  R.  C.,  J.  G.  Linn,  R.  L.  Ellingboe,  and 
N.  P.  Martin.  1987.  Preliminary  results  from 
monthly  forage  testing  and  ration  formulation 
on  four  daily  forms.  p.  24-33.  Jin  Proc.  XII 
Annual  Minnesota  Forage  Day.  Feb.  25,  1987. 
Rochester,  Minnesota. 


Table  1 . The  effect  of  grazing  pressure  on 
animal  selectivity  and  dietary  digestibility 
and  intake.  (From  Blazer  et  al . , 1977). 

Grazing  pressure 

Item 

L 

M H 

Selectivity 

H 

M L 

Digestibility 

H 

M L 

Intake 

H 

M L 

L,  M,  H = Low,  medium,  high,  respectively. 


53 


Table  2.  Calibration  statistics  and  wavelength 
selection  by  NIRS  to  predict  percent  leaf  in 


alfalfa  with 

saliva  treatments. 

Calibration  Statistics 

Treatment 

DN  DC  SL1  SL2 

R2  SEC 

No  Saliva 
Saliva 

1 20  20  2 

1 20  20  2 

0.97  4.69 

0.98  4.53 

Wavelengths  Selected,  nm 

No  Saliva 

2378,2258,2238,2218 
1718,1  138,1  1 18 

, 1998,1758, 

Saliva 

2318,2258,2238,22  18. 
1718,1138 

,1998,  1758, 

Table  3.  Boundaries  of  NIRS  prediction  means  for  pasture  components 
when  different  numbers  of  samples  (n)  are  used  to  estimate  the 
pasture  means. 


Component 

N 

X NIRS 

SEP 

Bias 

Boundary  when  n 

1 5 8 

10 

Fescue 

25 

37.0 

3.4 

-0.9 

- % Units  - 

7.0  3.1 

2.5 

2.2 

Alfalfa  leaves 

25 

39.9 

4 .4 

0.  1 

9.1  4.1 

3.2 

2.9 

Alfalfa  stems 

25 

24.  1 

5.5 

-0.2 

11.3  5.1 

4.0 

3.6 

Numbers  of  samples  in  NIRS  prediction  file 

t = 2.06  for  calculations. 

.06,24 


Table  4.  Hand  separated  (LAB)  vs.  NIRS  predicted  (NIRS)  components 
of  alfalfa/tall  fescue  pasture  mixtures  from  stratified  samples 
harvested  on  July  13,  1987. 


Component 

Method 

Strata 

(cm) 

0-10 

10 

-20 

20- 

■30 

30- 

-40 

X 

(SD) 

X 

(SD) 

X 

(SD) 

X 

(SD) 

% Units 

Fescue 

LAB 

6.0 

(1.8) 

8.5 

(6.4) 

0.7 

(1.1) 

0.0 

(0.0) 

NIRS 

7.7 

(4.4) 

7.6 

(4.7) 

-0.  1 

( 1.3) 

3.8 

(1.2) 

Alf  Leaf 

LAB 

34.8 

( 13.3) 

54.9 

(6.7) 

77.7 

(4.0) 

81.3 

(2.6) 

NIRS 

34.  1 

( 16.0) 

56.  1 

(6.0) 

78.3 

(2.1) 

80.2 

(1.7) 

Alf  Stem 

LAB 

59.2 

(11.9) 

36.7 

(4.2) 

22.0 

(3.9) 

18.5 

(2.4) 

NIRS 

58.6 

( 13.8) 

36.2 

(3.4) 

22.6 

(4.6) 

14.9 

(0.2) 

n + 

4 

4 

7 

3 

+ SD  = Standard  deviation  of  the  mean. 


n = Number  of  subsamples  analyzed. 


54 


FORAGE  INTAKE  AS  INFLUENCED  BY  SWARD 
CHARACTERISTICS 

T.D.A.  Forbes ^ 


INTRODUCTION 

The  efficient  production  of  animal 
products  from  grazed  forages  relies  largely 
on  the  producers  ability  to  efficiently 
manage  his  resources . The  most  important 
resource  the  producer  has  is  the  forage  that 
his  animals  graze,  and  his  main  objective 
should  be  to  maximize  forage  intake  by  his 
animals.  Efficient  grazing  management 

depends  to  a large  extent  on  the  producers 
understanding  of  the  influence  of  sward 
composition  and  structure  on  forage  intake. 
With  such  an  understanding  grazing  management 
strategies  can  be  devised  that  optimize 
productivity,  reduce  inputs  into  the  system, 
and  which  reduce  the  element  of  risk. 

FORAGE  INTAKE 

Conventionally,  forage  intake  is 
determined  over  relatively  short  periods, 
usually  seven  days,  and  the  results  are  then 
reported  on  a daily  basis.  For  many  purposes 
this  may  well  be  satisfactory.  However,  it 
may  be  more  realistic  to  consider  daily 
forage  intake  as  the  end  result  of  a series 
of  interrelated  processes,  beginning  with  the 
individual  bite  and  including  individual 
meals.  The  interpretation  of  the  interrela- 
tionships between  sward  characteristics  and 
animal  behavior,  both  social  and  ingestive 
are  the  basis  for  plant-animal  interface 
studies,  and  are  essential  to  the  understan- 
ding of  how  both  swards  and  animals  may  be 
manipulated  in  order  to  achieve  production 
objectives.  The  plant-animal  interface,  as 
defined  by  Moore  and  Sollenberger  ( 1986 ) , has 
been  much  discussed  (Hodgson,  1982;  Minson, 
19  83  ; Forbes  et  al.  1985;  Moore  and 
Sollenberger  1986),  but  has  been  studied  in 
depth  to  a much  lesser  extent  relative  to  its 
importance  to  animal  production.  Much  of  the 
work  that  has  been  done  has  been  conducted  in 
the  United  Kingdom  (Jamieson  and  Hodgson, 
1979a&b;  Hodgson  1981;  Forbes  and  Hodgson, 
1985;  Penning,  1986)  or  in  Australia  (Allden 
and  Whittaker,  1970;  Stobbs  1973a&b;  Chacon 
and  Stobbs,  1976;  Chacon  et  al.,  1978). 
Relatively  little  work  has  been  done  in  the 
United  States  considering  the  range  of 
forages  and  environments  encountered  (Forbes 
and  Colonan,  1987  and  unpublished;  Moore  et 
al.  1987;  Stuth  et  al.  1987). 

Intake  in  the  grazing  animal  is  a very 
much  more  complex  process  than  metabolism 
barn  studies  would  lead  us  to  believe.  Local 
climatic  and  inter-animal  social  factors  can 
have  large  effects  on  intake,  in  addition  to 


■'■Department  of  Agricultural  Engineering, 
University  of  Kentucky,  Lexington,  KY  40546. 


the  effects  of  diet  quality  and  quantity. 
Phillips  (quoted  by  Butris  and  Phillips, 
1987 ) found  that  DM  intake  of  dairy  cattle 
was  related  to  the  amount  of  rainfall  by  the 
expression  y (kg  DM/day)  = 13.2  -0.48  nan  rain 
day.  The  influence  of  both  heat  and  cold 
stress  on  intake  in  grazing  animals  is  well 
known  ( NRC  1981).  However,  the  most 
important  determinants  of  forage  intake  are 
sward  based,  and  include  mass  or  allowance, 
height,  structure,  botanical  composition,  as 
well  as  the  quality  aspects  including 
digestibility,  protein  and  fibre  content.  If 
it  is  accepted  that  animals  are  hedyphagic 
(McClymont,  1967)  then  it  can  be  argued  that 
most  limits  to  forage  intake  are  imposed  by 
sward  characteristics.  Intake  may  be  limited 
by  three  mechanisms:  (1)  Metabolic,  (2) 
Distention,  and  (3)  Behavioral  (Moore  and 
Sollenberger,  1986),  acting  independently  or 
together.  Metabolic  control  of  intake  is 
seldom  encountered  during  grazing  in 
productive  animals  and  will  not  be  considered 
further.  Distention  of  the  rumen  or  the 
intestinal  tract  (Waldo,  1986)  is  a likely 
form  of  intake  control,  particularly  on 
abundant,  poor  quality  forages  such  as  are 
found  in  tropical  and  subtropical  regions. 
In  this  case  rates  of  digestion  and  passage 
are  insufficient  to  allow  the  animal  to  take 
advantage  of  the  time  available  to  graze. 
The  third  control  mechanism,  that  of 
ingestive  behavior  works  through  the 
influence  of  sward  conditions  on  rate  of 
intake  and  time  spent  grazing. 

The  Grazing  Cycle. 

Figure  1 shows  a stylized  and  much 
simplified  interpretation  of  the  cycle  of 
events  from  the  start  of  one  grazing  period 
to  the  start  of  the  next.  Six  phases  and/or 
components  are  identified  within  each  cycle. 
The  accumulated  quantity  of  forage  eaten  at 
each  meal  equals  daily  forage  intake,  and  the 
number  of  cycles  that  can  occur  in  any  one 
day  is  influenced  by  both  sward  quantity  and 
quality . 

The  grazing  cycle  starts  with  site 
selection.  Grazing  animals  may  have  a great 
deal  of  choice  or  little  or  no  choice  as  to 
where  they  begin  grazing,  depending  on  the 
type  of  management  system  they  are  in.  Range 
animals  will  have  the  greatest  choice  while 
strip-grazed  dairy  cattle  will  have  the 
least.  Ignoring  the  extensive  range  system 
and  the  intensive  dairy  cow,  most  animals 
grazing  sown  swards  have  a limited  range  of 
choices.  Should  the  animal  graze  a totally 
ungrazed  patch  that  allows  rapid  satisfaction 
of  its  appetite  but  possibly  at  the  expense 
of  quality,  or  should  it  chose  previously 
grazed  patches  which  may  be  higher  in  quality 
but  which  provide  small  bite  sizes  and  which 
may  be  depleted  before  the  animal  is  full? 
Once  the  decision  has  been  made,  the  animal 
starts  grazing.  At  this  point  the  eating 
drive  is  dominant,  but  that  is  not  to  say 
that  the  animal  is  particularly  hungry.  It 
only  implies  that  other  drives  are  currently 


55 


set  at  lower  levels.  The  intensity  with 
which  it  grazes  in  terms  of  rate  of  intake  is 
a function  of  both  the  eating  drive  and  the 
sward  structure.  Rate  of  intake  is  the 
product  of  individual  bite  size  and  rate  of 
biting,  both  of  which  are  influenced  by 
changes  in  sward  structure. 

The  size  of  the  individual  bite  is 
determined  by  the  area  encompassed  by  the 
mouth  parts,  including  the  tongue  in  cattle, 
the  volume  of  the  buccal  cavity  and  the 
structure  of  the  sward.  Changes  in  the  size 
of  the  mouth  parts  occur  too  slowly  to 
influence  the  size  of  individual  bites, 
except  in  the  long  term.  However,  the  change 
from  milk  to  adult  teeth  may  have  a 
considerable  affect  on  potential  bite  size. 
Young  animals  appear  to  be  more  sensitive  to 
changing  sward  conditions  than  adult  animals 
(Hodgson  and  Jamieson,  1981). 

Changes  in  sward  structure  are  likely  to 
be  the  greatest  influence  on  reductions  or 
increases  in  bite  size.  Sward  structure  may 
vary  on  both  the  large  and  the  small  scale, 
changing  from  patch  to  patch  or  from  plant  to 
plant.  While  the  latter  may  result  in  bite 
size  changing  from  bite  to  bite,  it  may  be  of 
little  consequence  relative  to  the  much 
larger  increases  or  reductions  in  bite  size 
that  may  occur  as  animals  move  from  one  patch 
to  the  next.  Due  to  the  animals  ability  to 
compensate  for  reductions  in  short  term  bite 
size  changes  in  sward  structure  over  a period 
of  days  or  weeks  will  have  the  greatest 
effect  on  forage  intake.  Individual  bite 

size  generally  correlates  better  with  sward 
height  than  mass  over  a range  of  sward 
types.  It  is  likely,  however,  that 

individual  bite  size  is  more  responsive  to 
changes  in  density  of  leaf  in  the  surface 
horizon  or  the  leaf:sten  ratio.  Hodgson 
(1981)  reported  that  bite  size  was  more 
sensitive  to  variation  in  sward  height  than 
sward  surface  density.  In  contrast,  T.H. 
Stobbs  and  co-workers,  (Stobbs,  1973  a & b; 
Stobbs,  1975,  Chacon  and  Stobbs,  1976, 
Chacon,  Stobbs  and  Dale,  1978)  found  that  on 
sub- tropical  grass  swards  bite  size  was  more 
sensitive  to  variations  in  sward  surface 
density  and  leaf:  stem  ratio  than  sward 
height.  Studies  on  warm-season  grass  swards 
in  Oklahoma  (Forbes  and  Colonan,  1987;  and 
unpublished  data)  indicated  that  bite  size 
was  most  sensitive  to  variations  in  leaf: stem 
ratio.  On  vegetative  swards  or  swards 
without  a tall  diffuse  flower-head  horizon 
bite  size  appears  to  be  positively  related  to 
sward  mass,  height  and  surface  density 
(Hodgson,  1981;  Forbes,  1988).  However,  with 
the  appearance  of  an  upper  flower  horizon 
bite  size  declines  relative  to  sward  height 
and  surface  density  (Chacon  and  Stobbs,  1976; 
Chacon  et  al.,  1978;  Forbes,  1988).  Bite 
size  will  also  decline  as  swards  are  grazed 
down  and  the  sward  surface  approaches  the 
height  of  the  psuedo-stem  horizon  (Barthram 
and  Grant,  1984;  Forbes  and  Hodgson,  1985). 

Rate  of  biting  is  usually  negatively 
related  to  increases  in  sward  height  (Allden 


and  Whittaker,  1970;  Hodgson,  1981) 
reflecting  either  the  increased  time  required 
to  masticate  large  bites  or  the  difficulty  in 
prehending  preferred  components  within  a low 
density  canopy  depending  on  the  particular 
sward  structure.  As  grazing  continues  bite 
size  or  rate  of  biting  or  both  may  decline, 
leading  to  a decline  in  rate  of  intake.  On 
uniform  swards  the  decline  in  rate  of  intake 
may  be  very  abrupt  occurring  at  the  end  of 
the  grazing  period.  Under  other 
circumstances  it  may  be  more  gradually. 
Grazing  stops  when  the  eating  drive  is  less 
than  any  one  other  drive,  such  as  the  need  to 
drink,  rest,  seek  shade  or  because  the  rumen 
is  full . At  this  point  the  influence  of 
sward  structure  ends  and  sward  quality 
begins.  Also  at  this  point  the  animal  will 
have  eaten  some  proportion  of  its  daily 
requirements  for  maintenance  and  production. 


RUMINATION 

After  some  indeterminate  time  interval 
rumination  occurs.  A detailed  discussion  of 
the  mechanics  of  rumination  will  not  be  given 
here.  The  subject  has  been  reviewed  recently 
by  Deirment  et  al.,  (1987),  Ellis  et  al., 
(1987),  Martz  and  Belyea  (1986),  and  Pond  et 
al.,  (1987).  Rumination  serves  to  breakdown 
forage  particles  exposing  internal  structures 
to  microbial  attack  (Pond  et  al.,  1987),  and 
the  time  spent  ruminating  increases  with 
increases  in  forage  maturity  and  fibre 
content.  However,  high  quality  forages 
appear  to  stimulate  rumen  motility  and  flow 
per  contraction,  perhaps  explaining  the  high 
passage  rates  seen  in  such  forages  (Martz  and 
Balyea,  1986).  As  emptying  of  rumen  contents 
continues  and  the  animal  satisfies  other 
drives,  the  eating  drive  once  more  becomes 
dominant , and  the  animal  once  more  has  to 
make  a site  selection  decision.  Sward 
conditions  have,  however,  changed  since  the 
start  of  the  previous  grazing  period  and  the 
animals  choice  of  grazing  site  may  reflect 
this.  The  rate  of  intake  of  the  new  grazing 
period  will  reflect  the  change  in  sward 
structure.  The  length  of  the  grazing  session 
will  reflect  the  rapidity  of  the  reduction  in 
eating  drive  which,  in  part,  is  dependent  on 
the  amount  of  material  that  has  passed  from 
the  rumen  since  the  last  grazing  period. 
There  is  evidence,  however,  that  sheep,  at 
least,  do  not  eat  to  maximum  rumen  fill  at 
each  meal  (Thomson  et  al.,  1985).  These 
authors  found  that  while  legume  forage 
resulted  in  lower  levels  of  fill  than 
grasses,  maximum  rumen  fill  was  only  found  at 
the  end  of  the  afternoon  grazing  period. 
Forbes  (unpublished)  observed  that  cattle 
grazing  winter  wheat  in  the  early  spring  had 
up  to  14  distinct  grazing  periods  in  24h 
compared  with  4 periods  per  24h  when  grazing 
warm-season  grass  in  the  summer  months.  This 
difference  is  most  probably  a consequence  of 
differences  in  the  quality  of  the  forage 
influencing  rate  of  passage,  though  the 
influence  of  cold  stress  on  rumen  emptying 


56 


may  also  be  important  (Kennedy,  1985). 

In  conclusion,  sward  characteristics 
impact  on  forage  intake  in  two  ways. 
Firstly,  sward  structure  regulates  the  rate 
of  intake.  At  low  rates  of  intake,  other 
drives,  such  as  thirst,  may  override  the 
eating  drive  before  satiety  has  been 
reached.  Secondly,  sward  quality  determines 
the  amount  of  time  required  for  rumination, 
the  rate  of  digestion  and  the  supply  of 
nutrients.  Diets  that  contain  high  levels  of 
indigestible  fibre  reduce  the  rate  of  passage 
of  material  from  the  rumen,  and  the  rumen 
remains  full  for  longer  and  is  filled  in  a 
shorter  period  of  time  and  subsequent  grazing 
periods.  Ultimately,  the  animal  has  to 
increase  rumination  time  and  the  length  of 
time  between  meals.  Since,  however,  the  time 
available  for  grazing  is  finite,  the  scope 
for  increases  in  the  number  and  duration  of 
meals  is  limited  and  intake  is  depressed. 

The  producer,  then,  should  aim  to  provide 
his  animals  with  a sward  that  allows  maximum 
rates  of  intake  to  be  achieved.  Tall,  stemmy 
swards  not  only  limit  the  rate  of  intake  but 
also  may  be  of  such  low  quality  that  rumen 
fill  limits  the  amount  that  can  be  eaten  in 
any  one  grazing  period.  Carry-over  effects 
then  limit  the  number  of  grazing  periods 
possible  in  any  24  hour  period  and  con- 
sequently depress  intake. 

REFERENCES 


Allden,  W.G.  and  I.A.McD.  Whittaker.  1970. 

The  determinants  of  herbage  intake  by 
grazing  sheep:  the  interrelationship  of 
factors  influencing  herbage  intake  and 
availability.  Australian  J.  Agr.  Res. 
21:  755-766. 

Barthram,  G.T.  and  S.A.  Grant.  1984. 

Defoliation  of  ryegrass-dominated  swards 
by  sheep.  Grass  For.  Sci.  39:211. 

Butris,  G.Y.  and  C.J.C.  Phillips.  1987.  The 
effect  of  herbage  surface  water  and  the 
provision  of  supplementary  forage  on  the 
intake  and  feeding  behaviour  of  cattle. 
Grass  For.  Sci.  42:259. 

Chacon,  E.A.  and  T.H.  Stobbs.  1976. 

Influence  of  progressive  defoliation  of  a 
grass  sward  on  the  eating  behaviour  of 
cattle.  Australian  J.  Agr.  Res. 
27:709-727. 

Chacon,  E.A. , T.H.  Stobbs  and  M.B.  Dale. 

1978.  Influence  of  sward  characteristics 
on  grazing  behaviour  and  growth  of 
here  ford  steers  grazing  tropical  grass 
pastures.  Australian  J.  Agr.  Res.  29:89. 

Demxnent,  M.W. , E.A.  Laca  and  G.B.  Greenwood. 
1987.  Intake  in  grazing  ruminants:  a 
conceptual  framework.  Proc.  Symp.  Feed 
Intake  by  Beef  Cattle.  Oklahoma  State 
Univ.  MP-121 . pp.  208-225. 


Ellis,  W.C.,  J.H.  Matis,  C.  Lascano,  M. 

Mahloogi,  and  K.R.  Pond.  1987.  Size 
reduction,  fermentation  and  passage  of 
forage  particles  and  forage  intake  by 
cattle.  Proc.  Symp.  Feed  Intake  by  Beef 
Cattle.  Oklahoma  State  Univ.  MP-121. 
pp. 81-95. 

Forbes,  T.D.A.  1988.  Researching  the 

plant-animal  interface.  The 
investigation  of  ingestive  behavior  in 
grazing  animals.  J.  Anim.  Sci.  (In 
press ) . 

Forbes,  T.D.A.  and  J.  Hodgson.  1985. 

Comparative  studies  of  the  influence  of 
sward  conditions  on  the  ingestive 
behaviour  of  cows  and  sheep.  Grass  For. 
Sci.  40:69. 

Forbes,  T.D.A.  and  S.W.  Coleman.  1987. 

Herbage  intake  and  ingestive  behavior  of 
grazing  cattle  as  influenced  by  variation 
in  sward  characteristics.  Proc.  Special 
Session  Grazingland  Research  at  the 
Plant-Animal  Interface.  F.P.  Horn,  J. 
Hodgson,  J.J.  Mott  and  R.W.  Brougham 
(Eds.)  Winrock  International,  Morrilton, 
Arkansas,  pp.  141-152. 

Forbes,  T.D.A.,  E.M.  Smith,  R.B.  Razor,  C.T. 
Dougherty,  V.G.  Allen,  L.L.  Erlinger, 
J.E.  Moore  and  F.M.  Rouguette,  Jr. 
1985.  The  plant-animal  interface.  In: 
V.H.  Watson  and  C.M.  Wells,  Jr.  (Eds). 
Simulation  of  Forage  and  Beef  Production 
in  the  Southern  Region.  Southern  Coop. 
Series  Bull.  308.  pp.  95-116. 

Hodgson,  J.  1981.  Variations  in  the  surface 
characteristics  of  the  sward  and  the 
short-term  rate  of  herbage  intake  by 
calves  and  lambs.  Grass  For.  Sci. 
36:49. 

Hodgson,  J.  1982.  Influence  of  sward 

characteristics  on  diet  selection  and 
herbage  intake  by  the  grazing  animal . 
In:  J.B.  Hacker  (Ed.)  Nutritional 

Limits  to  Animal  Production  from 
Pasture.  Commonwealth  Agr.  Bur.  Famham 
Royale.  pp.  153-166. 

Hodgson,  J.  and  W.S.  Jamieson.  1981. 

Variations  in  herbage  mass  and 
digestibility,  and  the  grazing  behaviour 
and  herbage  intake  of  adult  cattle  and 
weaned  calves.  Grass  For.  Sci.  36:39. 

Jamieson,  W.S.  and  J.  Hodgson.  1979a.  The 
effect  of  daily  herbage  allowance  and 
sward  characteristics  upon  the  ingestive 
behaviour  and  herbage  intake  of  calves 
under  strip  grazing  management.  Grass 
For.  Sci.  34:261. 

Jamieson,  W.S.  and  J.  Hodgson.  1979b.  The 
effects  of  variation  in  sward 
characteristics  upon  the  ingestive 


57 


behaviour  and  herbage  intake  of  calves 
and  lambs  under  continuous  stocking 
management.  Grass  For.  Sci.  34:273. 

Kennedy,  P.M.  1985.  Influences  of  cold 

exposure  on  digestion  of  organic  matter, 
rates  of  passage  of  digesta  in  the 
gastrointestinal  tract  and  feeding  and 
rumination  behaviour  in  sheep  given  four 
forage  diets  in  the  chopped,  or  ground 
and  pelleted  form.  British  J.  Nutr. 
53:159. 

Martz,  F . A.  and  R.L.  Belyea.  1986.  Role  of 
particle  size  and  forage  quality  in 
digestion  and  passage  by  cattle  and 
sheep.  J.  Dairy  Sci.  69:1996. 

McClymont,  G.L.  1967.  Selectivity  and 

intake  in  the  grazing  ruminant.  In:  C.F. 
Code  (Ed.)  Handbook  of  Physiology, 
Section  6 : Alimentary  Tract,  Vol . I. 
American  Physiological  Soc.  Washington, 
D.C.  pp.  129-137. 

Minson,  D.J.  1983.  Forage  quality: 

Assessing  the  plant-animal  complex.  Proc. 
XIV  International  Grassland  Congress, 
Lexington,  KY.  pp.  23-27. 

Moore,  J.E.  and  L.E.  Sollenberger . 1986. 

Canopy  structure  effects  on  ingestive 
behavior.  In:  Proc.  42nd  Southern 

Pasture  and  forage  Crop  Improvement 
Conference,  Athens,  GA.  USDA/ARS.  pp. 
53-57. 

Moore,  J.E.,  L.E.  Sollenberger,  G.E. 

Morrantes  and  P.T.  Bede.  Canopy 
structure  of  Aesch^nomene  americana  - 
Hemarthria  altissima  pastures  and 
ingestive  behavior  of  cattle.  Proc. 
Special  Session  Grazingland  Research  at 
the  Plant -Animal  Interface.  F.P.  Horn, 
J.  Hodgson,  J.J.  Mott  and  R.W.  Brougham 
(Eds.)  Winrock  International,  Morrilton, 
Arkansas,  pp.  93-114. 

N.R.C.  1981.  Effect  of  environment  on 

nutrient  requirements  of  domestic 
animals.  National  Academy  of  Sciences  - 
National  Research  Council.  Washington, 
D.C. 

Pond,  K.R.,  J-M.  Luginbuhl  and  J.C.  Bums. 

1987.  Salivation,  mastication  and 
rumination  - limits  to  intake  by  beef 
cattle.  Proc.  Symp.  Feed  Intake  by  Beef 
Cattle.  Oklahoma  State  Univ.  MP-121. 
pp.  160-172. 

Stobbs,  T.H.  1973a.  The  effect  of  plant 

structure  on  the  intake  of  tropical 
pastures.  1.  Variation  in  the  bite  size 
of  cattle.  Australian  J.  Agr.  Res. 
24:809. 

Stobbs,  T.H.  1973b.  The  effect  of  plant 

structure  on  the  intake  of  tropical 


pastures.  2.  Differences  in  sward 
structure,  nutritive  value  and  bite  size 
of  animals  grazing  Setaria  anceps  and 
Chloris  guyana  at  various  stages  of 
growth.  Australian  J.  Agr.  Res.  24:821. 

Stobbs,  T.H.  1975.  The  effect  of  plant 

structure  on  the  intake  of  tropical 
pastures.  3.  Influence  of  fertilizer 
nitrogen  on  the  size  of  bite  harvested  by 
Jersey  cows  grazing  Setaria  anceps  cv. 
kazungula  swards.  Australian  J.  Agr. 
Res.  26:997. 

Stuth,  J.W.,  J.R.  Brown,  P.D.  Olson,  M.R. 

Aranjo  and  H.D.  Aljoe.  Effects  of 
stocking  rate  on  critical  plant -animal 
interactions  in  a rotationally  grazed 
Schizachyrium  - Paspalum  savanna . Proc . 
Special  Session  Grazingland  Research  at 
the  Plant-animal  Interface.  F.P.  Horn, 
J.  Hodgson,  J.J.  Mott  and  R.W.  Brougham 
(Eds.)  Winrock  International,  Morrilton, 
Arkansas,  pp.  115-140. 

Thomson,  B.C.,  G.J.  Cruickshank,  D.P.  Poppi 
and  A.R.  Sykes.  1985.  Diurnal  patterns 
of  rumen  fill  in  grazing  sheep.  New 
Zealand  Soc.  Anim.  Prod.  45:117. 

Waldo,  D.R.  1986.  Effect  of  forage  quality 
on  intake  and  forage-concentrate 
interactions.  J.  Dairy  Sci.  69:617. 


IHE  GRAZING  Oyor.Tt 


SITE 


Figure  1 . 

The  grazing  cycle:  major  components 
in  the  cycle  from  the  start  of  one  grazing 
period  to  the  start  of  the  next  grazing 
period. 


58 


BIOLOGICAL,  PRACTICAL  AND  STATISTICAL 
CONSIDERATIONS  ASSOCIATED  WITH  MEASURING 
FORAGE  AVAILABILITY  IN  GRAZING  TRIALS 

David  I.  Bransby  and  G.  Peter  Clarke1 

Introduction 

Under  grazing  conditions,  production  per 
animal  depends  on  the  quantity  of  forage 
consumed  (intake),  but  estimates  of 
intake  by  grazing  animals  are  often 
difficult  to  obtain.  However,  intake  and 
production  per  animal  are  strongly 
influenced  by  forage  availability  which 
can  be  modified  by  adjusting  the  number 
of  animals  per  unit  area.  In  a sense, 
therefore,  forage  availability  can  be 
regarded  as  an  index  of  intake  by  grazing 
animals.  As  such,  it  is  an  extremely 
important  variable  to  measure  in  grazing 
trials.  Our  objective  is  to  discuss  some 
biological,  practical  and  statistical 
considerations  associated  with  measuring 
forage  availability  in  grazing 
experiments . 

BIOLOGICAL  AND  PRACTICAL  CONSIDERATIONS 

The  purpose  of  forage  availability 
measurements . 

Forage  availability  measurements  may  be 
used  in  many  ways  to  explain  results  from 
grazing  studies.  Only  three  situations 
are  considered  here.  First,  forage 
availability  should  be  measured  in  single 
availability  put-and-take  trials  in  order 
to  facilitate  equalization  of  pasture 
conditions  between  treatments  and 
replications,  and  within  treatments  and 
replications  over  time,  by  adjusting 
animal  numbers.  Early  reports  on 
put-and-take  studies  often  did  not 
include  forage  availability  measurements 
(probably  because  forage  availability  was 
only  visually  rated)  while  more  recent 
studies  have  usually  quoted  only  an 
average  or  a range  in  availability  for  an 
entire  experiment.  Ideally,  reports  on 
put-and-take  studies  should  include  an 
analysis  of  forage  availabiltiy  data  to 
indicate  how  successful  the  procedure 
was  in  maintaining  a constant  pasture 
condition  across  the  experiment.  Read 
and  Camp  (4)  for  example,  showed  that 
they  were  not  successful  in  maintaining 
equal  availability  across  treatments  in  a 
put-and-take  trial  which  compared  animal 
production  from  tall  fescue  ( Festuca 


Department  of  Agronomy  and  Soils,  Auburn 
University,  AL  36849  and  Department  of 
Statistics  and  Biometry,  University  of 
Natal,  South  Africa,  respectively. 


arundi naceae)  with  and  without  the  fungal 
endophyte,  Acremonium  coenophialum.  In 
such  cases  it  may  be  appropriate  to 
perform  an  analysis  of  covariance  using 
forage  availability  as  a covariate. 

A second  situation  in  which  forage 
availability  measurements  are  essential 
is  where  forage  availability  is  varied 
experimentally.  This  facilitates  the 
development  of  three  important 
relationships  (production  per  animal  vs. 
stocking  rate,  production  per  animal  vs. 
available  forage,  and  available  forage 
vs.  stocking  rate)  and  meaningful 
interpretation  of  results  (1). 

Finally  by  measuring  forage  availability 
before  and  after  grazing  a subdivision  in 
a rotational ly  grazed  treatment,  an 
estimate  of  forage  consumed  by  grazing 
animals  can  be  made. 

Expression  of  forage  availability 

Quantity  of  forage  can  be  expressed  in 
several  ways  (eg.  kg  forage  per  unit 
area,  kg  forage  per  unit  weight  of 
animals,  kg  forage  per  unit  weight  of 
animals  per  day  or  forage  height). 
However,  it  is  likely  that  forage 
availability  affects  intake  and 
production  of  grazing  animals  through  its 
effect  on  the  ease  of  prehension  of  that 
forage.  Consequently,  the  form  in  which 
forage  availability  is  expressed  should 
preferably  reflect  the  ease  of  forage 
prehension  by  animals.  In  this  regard, 
kg  of  forage  per  unit  weight  of  animals 
may  not  be  appropriate.  For  example,  a 
one-  and  a five-ha  field  may  each  contain 
2000  kg  of  available  forage  and  5 
animals:  the  weight  of  forage  per 
animal  is  the  same  but  ease  of 
prehension,  intake  and  production  per 
animal  are  likely  to  be  different. 
Consequently,  we  recommend  that  forage 
availability  be  expressed  either  as  kg  of 
forage  per  unit  area,  or  in  terms  of  some 
height  measurement  when  used  to  explain 
responses  in  animal  production.  Weight 
of  forage  per  animal  per  day  would  likely 
have  most  application  under  rotational 
grazing.  However,  in  order  to  estimate 
intake  of  forage  by  measuring 
availability  before  and  after  grazing  a 
subdivision  under  rotational  grazing,  it 
is  clearly  necessary  to  express  forage 
quantity  as  weight  of  forage  per  unit 
area . 

Methods  for  measuring  forage 
avai 1 abi 1 i ty . 

From  a practical  point  of  view,  methods 
for  measuring  forage  availability  should 


59 


preferably  be  (1)  objective,  (2) 
non-destructive,  (3)  quick  and 
inexpensive,  (4)  repeatable  and 
(5)  simple  in  procedure  and 
instrumentation.  Generally,  these 
methods  can  be  divided  into  three 
categories:  direct  measurement  by 
clipping  quadrants  or  cutting  mower 
strips,  height  measurements,  and  double 
sampling  procedures  in  which  some 
easy-to-measure  forage  attribute  is 
related  to  forage  yield.  The  main 
disadvantage  of  direct  clipping  is  that 
it  is  labor-intensive  and  therefore 
expensive.  It  is  also  destructive. 

Height  measurements  may  be  made  with  a 
measuring  stick,  but  these  estimates  are 
subjective  and  may  be  low  in 
repeatabi 1 i ty  from  one  worker  to  another. 
Alternatively,  height  can  be  measured 
with  measuring  devices  such  as  a disk 
meter  or  rising  plate  meter  in  which  case 
repeatability  should  be  high. 

Many  double  sampling  procedures  for 
measuring  forage  availability  have  been 
used,  including  visual  estimation, 
various  kinds  of  electronic  meters  and 
disk  or  rising  plate  meters.  Visual 
estimation  is  subjective  and  may  be  low 
in  repeatability,  while  electronic  meters 
require  careful  handling  and  can  be 
expensive,  fragile  and  sensitive  to 
moisture  in  the  soil.  The  disk  meter 
(2,3)  or  rising  plate  meter  method  has 
provided  good  results  in  a wide  range  of 
conditions,  but  in  certain  situations 
(such  as  very  tall  pasture)  it  may  be 
inappropriate.  In  general,  however,  it 
ranks  well  in  terms  of  requirements  1 
through  5 listed  above. 

Use  of  the  disk  meter  involves  two 
stages,  as  for  any  other  double  sampling 
procedure.  The  first  stage,  or 
calibration,  requires  the  collection  of  a 
paired  disk  meter  reading  and  weight  of 
forage  beneath  the  disk  data  set.  The 
function  of  this  calibration  is  simply  to 
facilitate  conversion  of  disk  meter 
reading  to  kg  of  forage/ha.  Therefore, 
these  samples  do  not  need  to  be  randomly 
located.  In  fact,  it  may  be  better  to 
deliberately  select  samples  for 
calibration,  in  order  to  ensure  that  very 
low  and  very  high  values  are  included 
from  the  population  to  be  sampled  in  the 
second  stage,  ie.  the  calibration  should 
be  representative.  When  developing  the 
regression  equation,  the  line  should  not 
be  constrained  to  pass  through  the 
origin.  Extrapolation  of  the  line  below 
the  lowest  value  in  the  calibration  may 
result  in  intersection  of  the  horizontal 
axis.  This  is  entirely  reasonable, 
because  even  when  no  forage  is  present  a 


reading  above  zero  can  be  obtained  from 
the  disk  meter  due  to  uneven  ground 
surface.  On  the  other  hand, 
extrapolation  may  result  in  intersection 
of  the  vertical  axis.  This  may  occur 
because  (a)  the  true  relationship  between 
forage  weight  and  disk  height  is  not 
linear,  yet  it  is  being  approximated  by  a 
straight  line,  and  (b)  all  available 
forage  above  ground  level  is  often  not 
harvested. 

Several  factors  are  known  to  affect  the 
calibration  equation  for  the  disk  meter 
(2,3).  These  include  different  species, 
different  seasons,  reproductive  vs. 
vegetative  growth  phases  and  grazed  vs. 
ungrazed  pasture.  Separate  calibrations 
should  therefore  be  developed  and  tested 
for  difference  in  each  of  these 
situations.  Furthermore,  the  disk  meter 
should  not  be  used  when  forage  is  wet 
(from  dew  or  rain)  or  under  very  droughty 
conditions  (which  cause  plants  to  wilt) 
without  specific  calibrations  for  these 
conditions. 

Sampling  in  the  second  stage  (in  which 
only  disk  meter  readings  are  taken) 
should  ideally  be  random.  However, 
random  location  of  disk  meter  readings  in 
a paddock  would  be  extremely  time 
consuming.  In  most  cases,  disk  meter 
readings  taken  in  several  transects 
across  a paddock  are  satisfactory.  Such 
transects  should  cut  across  any  obvious 
variation  in  forage  availability  within  a 
paddock. 


STATISTICAL  CONSIDERATIONS 

Primarily  we  will  look  at  the  statistical 
efficiency  of  the  two-stage  sampling 
procedure  based  on  the  disc  meter. 

1.  Description  of  the  Method 

Stage  1.  - The  calibration  stage. 

Site  a total  of  n^  (usually  about  10) 
sample  points  across  all  paddocks  in  the 
experiment.  This  would  normally  be  done 
systematically  in  an  effort  to  ensure  as 
wide  a range  of  conditions  as  possible. 

At  each  point: 

(a)  take  a disc  meter  reading  (h), 

(b)  mark  out,  as  accurately  as 
possible,  the  edges  of  the  disc  on 
the  ground  and  cut  off  the  forage 
within  the  delineated  circle,  dry 
the  cut  forage  and  record  the 
weight  (w), 

(c)  fit  a linear  regression  equation  to 
the  data  pooled  from  all  the 


60 


paddocks,  giving  a regression 
equation  of  the  form: 

A 

w = a + b h [1] 

where  w is  the  predicted  weight  of 
cut  forage  from  a point  with  disc 
meter  height  h. 

Also  conduct  an  auxilliary  Analysis 
of  Variance,  where  from  the  pooled 
data,  the  total  sum  of  squares  for 
weights  w is  subdivided  into 
regression  and  deviations  and  we 
eventually  calculate  S,  the 
deviations  mean  square. 

At  this  stage  from  step  1 we  need  to  have 
recorded  the  following: 

(i)  the  regression  equation  [1], 

o 

(ii)  the  deviations  mean  square  Sf 

(iii)  hi,  the  mean  disc  meter  height  from 
all  the  pooled  data, 

( iv)  SS(h)  = E (h  - h^) 2 , and 

(v)  n^  = total  first  stage  size. 


Stage  2.  - The  estimation  stage. 

Consider  the  question  now  of  estimating 
forage  availability  in  a specific  paddock 
of  an  experiment. 

(a)  Choose  n p (Usually  about  40  points) 
at  random  and  at  each  point  record 
the  disc  meter  height  h. 

(b)  Calculate  hp,  the  mean  disc  meter 
height  for  tne  paddock  and  , the 
variance  among  the  measured 
heights.  Now  calculate: 

A 

w = a + b h, 

where  a and  b come  from  equation 

[1]. 

This  calculated  value,  w,  is  the 
predicted  weight  of  forage,  per 
unit  disc  meter  area,  in  the 
paddock  of  interest. 

(c)  Make  the  following  calculation: 

Var(w)  = S2[l/n  + (h  - hi  ) 2 /SS] 

2 2 2 2 W 
+ b + S SpASS.np)  — [2] 

and  SE(w  ) = ,/var(w  ). 

Note  that  these  formulae  do  not 
correspond  to  the  standard 
regression  formulae  due  to  the 
random  nature  of  the  second  stage 
sampl ing. 


2.  Analysis  of  a Specific  Data  Set 

In  order  to  examine  the  efficiency  of 
this  scheme,  data  was  collected  from  a 
mixed  pasture  including  rye,  ryegrass  and 
crimson  clover. 

2. 1 Basic  Analysis 

A total  of  50  first  stage  units  were 
chosen  using  2 transects.  In  addition,  4 
points  were  chosen  where  the  pasture  was 
particularly  high.  A regression 
analysis  based  on  10  deliberately  chosen 
points  is  shown  in  table  1.  This 
typifies  the  type  of  regression  commonly 
found. 


A total  of  198  second  stage  units  were 
sited  on  a rectangular  grid  pattern.  The 
markedly  skew  frequency  distribution  of 
these  heights  is  illustrated  in  figure  1. 


Table  1.  Regression  of  first  stage 


Anal ysi 

s of  Var i ance 

Sum  of 

Mean 

Source  DF 

Squares 

Square 

F Value 

Model  1 

1944. <74055 

1944.94055 

52. 723 

Error  8 

295. 12045 

36. 89006 

C Total  9 

2240.06100 

Root  MSE 

6.07372 

R-Square 

0. 8683 

Dep  Mean 

22. 17000 

Adj  R-Sq 

0. 8518 

C.  V. 

27. 3961 1 

Parameter  Estimates 

Parameter 

Standard 

T for  MO: 

Variable  DF 

Est i mate 

Er  ror 

Parameter=0 

INTERCEP  l 

5.509196 

2. 99231734 

1.841 

DH  1 

0.946637 

0. 13037213 

7.  '‘61 

1 1 1 1 1 1 l l 1 122222222223333233 
3*55670901  2345670  9.0  1 234567090123456 

Disk  Height 


61 


Figure  2.  (a)  Histogram  of  predicted  values 


2.2  Analysis  to  examine  the  effects  of 
skewness 

The  following  two-stage  simulation  scheme 
was  carried  out.  In  stage  1 the 
regression  parameter  estimates  were 
randomly  generated  so  that  their  means 
corresponded  with  those  in  the  observed 
real  sample  of  size  10.  In  the  second 
stage,  a sub-sample  of  size  40  was 
randomly  drawn  from  the  real  observed  198 
units.  Then  the  predicted  forage  value 
was  calculated  from  its  regression 
equation  and  its  SE  using  equation  [2]. 
This  was  repeated  1000  times  and  two 
frequency  distributions  were  drawn  up. 

The  first  is  that  of  the  predicted  values 
w and  the  second  is  that  of 

t = (w  -^o-)/(SE(w  ) 

whereyu_is  the  mean  value  of  w overall. 

This  calculated  value  should 
approximately  follow  the  t distribution 
with  38  degrees  of  freedom.  As  the 
histograms  in  figure  2 show,  the 
distributions  are  nearly  symmetric  and 
further  analysis  of  the  calculated  t 
values  shows  that  they  are  very  close  to 
their  theoretical  expectations. 

2.3  Analysis  to  examine  optimal  choice  of 
sample  size 

Using  equation  [2],  one  can  calculate 
SE's  of  predicted  values  for  varying 
values  of  nl  and  n2.  Figure  3 
illustrates  these  values  for  nl  between  5 
and  25  and  for  n2  up  to  100. 

Figure  1.  Frequency  of  second  stage  disc  heights 

Frequency 


70  ♦ 


60  ♦ 


30  ♦ 


40  ♦ 


30  + •**»* 


10  -*■ 


10  15  20  25  30  35  40 

DH  MIDPOINT 


Hi s tog ram 


*-*-*-* 

* * *-*-*-**- 
*-*-#--*-* -*-***- 
**  * ■**-*--*-*-•* *--**  «-*•*-*# 

********************  ****  ***** 

*******  **** **  **  **  ****  ***  *** *********** 
************ ********** ************** 
**************  * *****  * * **************  * 
***-*************************** 
******************** 

************* 

******* 

**** 


14 
25 
■=;  a 


115 

151 

143 

146 

120 

78 

51 

28 

14 


8.  5+* 


* may  represent  up  to  4 counts 

Figure  2.  (b)  Histogram  of  t values 


Hi stogram 


****•*-**■**-****■*- 

*****************.***********-*-* 

-0.  25+  ***************-*■*-*-•*-*-■#■*-**• **-*■*  **"*-*-*-*■**--** 

. *****-*****************.**-**..*.*.*.*.x.* 
********-**■*****-■*-**-■*-*- 
*•■**--*"*  **--*-* 

*-*■**- 
■** 

•* 

-3.75+* 

+ F + + 4- h 4- F 

* may  represent  up  to  5 counts 
a 

Figure  3.  (SE(w  )for  varying  sample  sizes 


i 

4 

6 

14 
20 
18 

15 
96 
45 
19 

9 


:se(w 


"i-20 

H!-2S 


b b b 


b b b b b 


b b b b b 


d d d d d 


O lO  20  30  40  50  60  70  00  90  100 


62 


'vl  O O vj  >0  VJ  LH  hJ 


An  alternative  approach  is  to  determine 
an  optimal  sampling  strategy  which  will 
minimize  cost  to  achieve  a specific 
accuracy.  In  table  2 these  values  are 
given  for  cost  ratios,  being  the  relative 
costs  of  first  and  second  stage  units 
between  2.5  to  1 and  10  to  1 (the  cost 
being  considered  in  terms  of  time 
required  to  obtain  a first  and  second 
stage  sample). 

Table  2.  Optimal  sample  sizes  for  varying  cost  ratios. 


SE 

1 

: i 

2.  5 

* 

Cost  Ratios 

1:5  1:7 

5 

1 : 10 

: ni 

n2  ; 

n i 

"2  ; 

ni 

n2 

: 

"2 

1 . 0 

27 

72 

23 

86 

21 

97 

20 

107 

1 . 9 

23 

60 

19 

72 

18 

81 

17 

89 

2.  1 

19 

51 

16 

61 

15 

69 

14 

75 

2.3 

17 

43 

14 

52 

13 

59 

12 

65 

2.4 

14 

38 

12 

45 

1 1 

51 

1 1 

56 

2.6 

13 

33 

1 1 

40 

10 

45 

9 

49 

2.8 

11 

29 

9 

35 

9 

40 

8 

43 

2.9 

10 

26 

8 

31 

8 

35 

7 

39 

3.  1 

9 

23 

7 

28 

7 

32 

7 

35 

3.3 

8 

21 

7 

25 

6 

28 

6 

31 

3.  4 

7 

19 

6 

23 

6 

26 

5 

28 

3.6 

7 

17 

6 

21 

5 

23 

5 

26 

3.8 

6 

16 

5 

19 

5 

21 

4 

24 

4.0 

5 

14 

5 

17 

4 

20 

4 

22 

4.  1 

5 

13 

4 

16 

4 

18 

4 

20 

4.3 

5 

12 

4 

15 

4 

17 

3 

18 

4.5 

4 

11 

4 

14 

3 

16 

3 

17 

4.6 

4 

11 

3 

13 

3 

14 

3 

16 

4.8 

4 

10 

3 

12 

3 

13 

3 

15 

5.0 

4 

9 

3 

1 1 

3 

13 

3 

14 

5.  1 

3 

9 

•? 

10 

3 

12 

2 

13 

5.3 

3 

8 

3 

10 

2 

1 1 

2 

12 

5.5 

3 

8 

2 

9 

2 

10 

2 

1 1 

5. 6 

3 

7 

2 

9 

2 

10 

2 

1 1 

5. a 

3 

7 

2 

8 

2 

9 

2 

10 

* 

Cost 

ratio  is  defined 

as 

the  ratio 

of  costs 

of  sampling 

one  second  stage  unit  to  one  -first  stage  unit. 


Finally  it  is  worth  commenting  that  in 
the  set  of  data  used,  in  order  to  achieve 
the  same  precision  using  quadrat  sampling 
only,  as  what  we  get  from  10  stage  1 and 
40  stage  2 units,  one  would  need  to 
sample  18  first  stage  quadrats. 

REFERENCES 

1. Bransby,  D.I.,  B.E.  Conrad,  H.M.  Dicks, 
and  J.W.  Drane.  1988.  Justification  for 
grazing  intensity  experiments:  analysing 
and  interpreting  grazing  data.  J.  Range 
Manage.  41:  274-279. 

2.  Bransby,  D.I.,  A. 6.  Matches  and  G.F. 
Krause.  1977.  Disk  meter  for  rapid 
estimation  of  herbage  yield  in  grazing 
trials.  Agron.  J.  69:  393-396. 

3.  Bransby,  D.I.  and  N.M.  Tainton.  1977. 

The  disk  meter:  possible  applications  in 
grazing  management.  Proc.  Grassl . Soc. 
Sth.  Afr.  12:  115-118. 

4.  Read,  J.C.  and  B.J.  Camp.  1986.  The 
effect  of  the  fungal  endophyte, 

Acreinonium  coenophial urn  in  tall  fescue  on 
animal  performance,  toxicity  and  stand 
maintenance.  Agron.  J.  78:  848-850. 


63 


EFFECT  OF  ENDOPHYTE  LEVEL  OF  TALL  FESCUE  ON 
SUBSEQUENT  FEEDLOT  .PERFORMANCE 
OF  STEERS- 


N.  Andy  Cole- 


37 


Introduction 


Approximately  80%  of  the  tall  fescue  (Festuca 
arundinacea  Schreb)  pastures  in  the  Southeast 
and  Midwest  are  Infested  with  the  endophytic 
fungus  Acremonium  coenophialum  (Daniels 
et  al.,  1985).  When  consuming  infested 
forage,  cattle  have  lower  feed  intakes,  poorer 
daily  gains,  and  lower  heat  tolerance  than 
cattle  consuming  uninfested  forage  pastures 
(Hemken  et  al . , 1981;  Stuedemann  and  Hoveland, 
1988).  Other  clinical  signs  of  endophyte 
toxicity  are  rough  haircoats,  rapid  breathing, 
elevated  body  temperature,  and  lowered  serum 
prolactin  concentrations  (Stuedemann  and 
Hoveland,  1988).  For  many  years,  the  cattle 
feeding  industry  reported  that  calves  from 
areas  with  predominately  fescue  pastures  had  a 
higher  incidence  of  health  problems 
[especially  bovine  respiratory  disease  (BRD)] 
than  calves  from  nonfescue  areas.  It  was 
generally  felt  this  high  incidence  of  BRD  was 
due  to  the  small  size  of  cattle  operations  and 
the  marketing  system  used.  In  the  past  few 
years,  the  possible  role  of  the  fescue 
endophyte  in  this  health  problem  has  een 
considered . 


Stuedemann  et  al.  (1985b)  reported  that  the 
adverse  effects  of  the  fescue  endophyte 
appeared  to  carry  over  for  4-8  weeks  in  calves 
switched  from  high-  to  low-endophyte 
pastures.  The  adverse  effects  appeared  to 
show  up  almost  immediately  in  calves  switched 
from  low-  to  high-endophyte  pastures.  In  the 
past  few  years,  several  trials  have  been 
conducted  to  determine  how  cattle  from 
endophyte  infested  pastures  perform  in  the 
feedlot  and  if  carry-over  effects  cause 
increased  health  problems. 

Animal  Health 

Diagnostic  Lab  Reports.  No  controlled  data 
are  available  to  indicate  that  a carry-over 


— Contribution  from  USDA,  Agricultural 
Research  Service,  Conservation  and  Production 
Research  Laboratory,  P.0.  Drawer  10,  Bushland, 
TX  79012. 

2/ 

For  presentation  at  the  44th  Annual 
Meeting  of  the  Southern  Pasture  and  Forage 
Crop  Improvement  Conference,  Lexington,  KY. 

May  10-12,  1988. 


effect  of  the  fescue  endophyte  causes 
increased  incidence  of  health  problems  in 
feeder  calves.  Perino  (1985)  reported  severe 
cases  of  heat  stroke  in  some  groups  of  feeder 
calves  during  the  first  3-5  days  in  the 
feedlot  in  July.  Death  losses  as  high  as  10% 
were  reported  in  some  lots.  Heat  sensitivity 
appeared  to  continue  for  2-4  weeks,  although 
all  groups  were  not  affected  to  the  same 
degree.  Cattle  in  the  affected  groups  were 
known  to  have  grazed  fescue  pastures  prior  to 
entering  the  feedlot  and  had  typical  symptoms 
of  summer  fescue  toxicosis. 

Sprowls  (1987)  reported  that  groups  of  cattle 
with  overt  signs  of  fescue  toxicosis  generally 
appeared  to  have  a high  incidence  of 
respiratory  disease.  However,  the  major 
health  problem  associated  with  cattle  from 
fescue  pastures  was  heat  stroke  during  the 
summer  months.  Other  problems  associated  with 
cattle  from  fescue  pastures  were  low  serum 
selenium  and  serum/liver  copper  levels.  Other 
studies  have  also  reported  low  serum  selenium 
(Lackey,  1985)  and  lowered  serum  and  liver 
copper  (Stoszek  et  al.,  1979)  in  cattle  from 
fescue  pastures. 

In  four  controlled  studies,  yearling  cattle 
that  had  grazed  pastures  containing  high  (59% 
infested),  moderate  (29%  infested),  or  low  (8% 
infested)  levels  of  endophyte  were  shipped 
from  Georgia  to  Texas  (Cole,  1987;  Cole 
et  al.,  1987).  Steers  on  the  four  trials 
arrived  in  July,  August,  September,  and 
October.  No  animals  in  any  group  required 
treatment  for  respiratory  disease;  however, 
steers  from  high-endophyte  pastures  tended  to 
have  higher  morbidity  scores  than  steers  from 
low-endophyte  pastures,  based  on  nasal 
discharge,  ocular  discharge,  and  elevated 
rectal  temperature.  Steers  from 
high-endophyte  pastures  tended  to  have  lower 
serum  complement  levels  than  steers  from 
bermudagrass  pastures  (Purdy  et  al . , 1987), 
suggesting  a suppressed  immune  response  in 
steers  from  high-endophyte  pastures. 

Marketing  and  Transit  Shrink 

The  effects  of  the  fescue  endophyte  on 
marketing-transport  shrink  of  feeder  steers 
are  equivocal.  Some  studies  have  reported 
greater  shrink  in  steers  from  highly  infested 
pastures  (Cole  et  al.,  1987),  while  others 
have  reported  less  shrink  in  calves  from 
infested  pastures  (Cole  et  al.,  1987;  Lusby, 
1988).  If  the  endophyte  does  affect 
marketing-transit  shrink,  the  results  are 
probably  dependent  upon  factors  such  as 
weighing  conditions,  the  forage  fed  prior  to 
loading,  and  weather  conditions. 


— Research  Animal  Scientist,  USDA, 
Agricultural  Research  Service,  Conservation 
and  Production  Research  Laboratory,  Bushland, 


TX  79012. 


Performance  in  the  Feedlot 

Missouri  Studies.  Hancock  and  Williams  (1985) 
compared  the  feedlot  performance  of  steers 


64 


from  fescue,  orchardgrass , and  bromegrass 
pastures  (Table  1).  During  the  early  portion 
of  the  feeding  period,  calves  from  fescue 
pastures  tended  to  have  the  poorest 
performance;  however,  by  the  end  of  the 
112-day  feeding  period,  calves  from  fescue 
pastures  had  faster  daily  gains  and  more 
efficient  feed  conversions  than  steers  from 
orchardgrass  and  bromegrass  pastures. 

Oklahoma  Studies.  In  one  study  (Lusby,  1988), 
steers  grazed  pastures  of  endophyte-infested 
fescue,  infested  fescue  + clover,  and 
endophyte-free  fescue  for  197  days  (Nov.  to 
May  1987)  (Table  2).  Steers  were  held  on 
ryegrass  pastures  for  6 days  and  then  moved 
about  450  km  to  a feedlot  for  finishing. 

Steers  from  infested  pastures  gained  about 
55  kg  less  than  the  remaining  two  treatments 
during  the  grazing  period.  Weight  gains  were 
slightly  higher  in  steers  from  infested 
pastures  than  in  steers  from  endophyte-free 
pastures  during  the  early  portion  of  the 
feeding  period  and  were  significantly  higher 
by  the  end  of  the  feeding  period.  There  were 
no  effects  on  carcass  traits. 

Arkansas  Studies.  In  a study  by  Piper  et  al. 
(1987),  steers  grazed  endophyte-infested  or 
endophyte-free  fescue  pastures  for  84  or 
168  days.  Steers  were  moved  to  drylot  pens  in 
October.  Steers  which  had  grazed  infested 
pastures  had  faster  feedlot  daily  gains  (Table  3). 
Serum  prolactin  concentrations  were  lower  in 
steers  from  infested  pastures  on  days  0 and  7 
in  the  feedlot  but  were  similar  by  day  14  in 
the  feedlot. 

In  a second  study  (Piper  et  al.,  1987),  steers 
grazed  infested  and  noninfested  pastures  and 
were  moved  to  the  feedlot  in  July.  Daily 
weight  gains  in  the  feedlot  were  not 
significantly  different;  however,  steers  from 
infested  pastures  tended  to  have  lower  daily 
gains  than  steers  from  noninfested  pastures. 

Serum  prolactin  concentrations  were  lower  in 
steers  from  infested  pastures  on  days  0,  7, 

14,  and  21  in  the  feedlot  but  were  similar  by 
day  28.  The  differences  in  performance  and 
serum  prolactin  concentrations  in  the  October 
and  July  studies  suggested  that  the  ambient 
conditions  during  the  feeding  period  could 
affect  the  time  required  for  calves  from 
endophyte-infested  pastures  to  recover  from 
the  adverse  effects  of  the  endophyte. 

Kentucky  Studies.  Smith  et  al.  (1986)  allowed 
steers  to  graze  infested  or  noninfested 
pastures  from  April  to  September,  when  they 
were  moved  to  drylot  for  59  days.  Initial 
weights  were  349  and  445  kg  for  the  infested 
and  noninfested  groups,  respectively 
(Table  4).  Steers  from  endophyte-free 
pastures  had  higher  feed  intakes,  lower  daily 
gains,  and  higher  feed/gain  ratios  than  steers 
from  high-endophyte  pastures.  Due  to  the 
heavy  starting  weights  of  the  steers  from  the 


low-endophyte  pastures,  it  could  not  be 
clearly  determined  if  these  effects  were  due 
solely  to  the  endophyte  or  were  partly  due  to 
differences  in  starting  weights  and  body 
condition  of  the  steers. 

USDA-ARS :Georgia~Texas  Studies.  Four 
cooperative  studies  have  been  conducted 
between  investigators  at  the  USDA-ARS, 
Watkinsville , GA,  and  Bushland , TX.  In  each 
study,  12  steers  grazed  pastures  containing 
high,  moderate,  or  low  endophyte  levels.  In 
the  first  two  studies,  all  animals  were  of 
Angus  breeding;  and  in  the  latter  two  studies, 
one-half  were  Angus  and  one-half  were 
Brahman-crossbred.  Steers  were  shipped  from 
Georgia  to  Texas  in  October  1985,  July  1986, 
August  1986,  and  September  1987.  Steers  were 
removed  from  pastures,  fasted  overnight,  and 
weighed.  After  2 days  in  a simulated 
orderbuyer  barn,  steers  were  transported  for 
26  hours  to  Texas.  Upon  arrival,  steers  were 
weighed,  body  temperature  recorded,  and 
assigned  to  pens  equipped  with  Pinpointers 
(Model  4000B,  UIS  Corp.,  Cookville,  TN)  for 
measurement  of  individual  feed  intake.  Cattle 
were  slaughtered  when  backfat  thickness  was 
estimated  to  be  12  mm.  In  trials  2,  3,  and  4, 
six  steers  from  bermudagrass  pastures  were 
also  included. 

In  no  trial  was  there  evidence  of  a 
significant  carry-over  effect  of  the  endophyte 
(Table  5).  In  all  trials,  steers  from 
high-endophyte  pastures  had  faster  daily  gains 
and  improved  feed/gain  ratios  than  steers  from 
low-endophyte  pastures  during  the  first 
28  days  on  feed.  Carcass  traits  were  not 
affected  in  any  trial. 

Serum  cholesterol  values  were  lower  in  steers 
from  high-endophyte  than  low-endophyte 
pastures  for  about  14  days,  suggesting  some 
carry-over  effect  of  the  endophyte  (Stuedemann 
et  al.,  1985a).  This  carry-over  effect, 
however,  did  not  affect  early  feedlot 
performance  or  health.  Subjective 
observations  indicated  that  in  the  summer 
trials,  calves  exhibited  some  heat  stress 
during  the  afternoon.  Respiration  rates  of 
steers  from  the  high-endophyte  pastures  tended 
to  be  higher  than  those  of  steers  from 
low-endophyte  or  bermudagrass  pastures; 
however,  all  steers  tended  to  have  elevated 
respiration  rates,  even  during  the  cool  of  the 
morning.  Whether  this  was  due  to  a rapid 
change  in  altitude  (from  about  130  m to 
1,700  m) , to  a subclinical  respiratory 
infection,  or  to  other  causes  was  not  clear. 

Conclusions 

The  available  data  and  observations  suggest 
that  calves  from  fescue  pastures  infested  with 
endophyte  may  have  more  health  problems  than 
calves  from  noninfested  pastures,  especially 
during  hot  weather.  If  they  remain  healthy, 


65 


calves  from  highly  infested  pastures  will 
likely  have  feedlot  performance  equal  to  or 
better  than  calves  from  noninfested  pastures. 
This  is  probably  the  result  of  compensatory 
gain.  Differences  in  feedlot  performance  of 
calves  from  high-  and  low-endophyte  pastures 
may  be  affected  by  environmental  conditions 
(Hemken  et  al.,  1981)  and  by  differences  in 
performance  during  the  grazing  period. 

Literature  Cited 

Cole,  N.  A.  1987.  Fescue  toxicosis: 

Carry-over  effects  in  the  feedlot.  In 
Proc.  Symp.  on  Fescue  Toxicosis  - From 
Cow-Calf  to  Slaughter.  Amarillo,  TX, 

April  15,  1987. 

Cole,  N.  A.,  J.  A.  Stuedemann,  C.  W.  Purdy, 
and  D.  P.  Hutcheson.  1987.  Influence  of 
endophyte  in  tall  fescue  pastures  on  the 
feedlot  performance  of  feeder  steers.  J. 
Anirn.  Sci  . 65(Suppl.  1 ) : 331 . 

Daniels,  L.  B.,  E.  L.  Piper,  B.  J.  Hankins, 

G.  Gee,  T.  S.  Nelson,  and  J.  Gergerich. 
1985.  Infestation  level  of  Acremonium 
coenophialum  in  fescue  in  Arkansas.  J. 
Anim.  Sci.  61(Suppl.  1 ) : 3 3 2 . 

Hancock,  D.  L.,  and  J.  E.  Williams.  1985. 

Effects  of  previous  forage  grazing  system 
on  performance,  body  composition,  carcass 
characteristics,  and  plasma  variables  of 
feedlot  cattle.  Missouri  Cattle  Feeders 
Seminar,  Anim.  Sci.  Rep.  #108:20-31. 

Hemken,  R.  W.,  J.  A.  Boling,  L.  S.  Bull,  R.  H. 
Hatton,  R.  C.  Buckner,  and  L.  P.  Bush. 
1981.  Interaction  of  environmental 
temperature  and  anti-quality  factors  on 
the  severity  of  summer  fescue  toxicosis. 

J.  Anim.  Sci.  52:710. 

Lackey,  J.  1985.  How  they  perform  in  my 

feedlot.  Univ.  Missouri  Beef  Cattle  Rep. 
Anim.  Sci.  Rep.  # 108:53-55. 

Lusby,  K.  1988.  KOMA  Beef  Cattle  Conf . 
Joplin,  MO,  January  1988. 

Perino,  L.  1985.  Possible  fescue  toxicosis 
in  incoming  feeder  cattle.  Proc.  Annual 
Conv . Am.  Acad.  Bovine  Practitioners, 
Stillwater,  OK.  pp . 139-140. 

Piper,  E.  L.,  K.  W.  Beers,  A.  L.  Goetsch,  and 
Z.  Johnson.  1987.  Effect  of  grazing 
endophyte  infected  fescue  on  subsequent 
feedlot  performance  of  beef  steers.  J. 
Anim.  Sci.  65(Suppl.  1):331. 

Purdy,  C.  W.,  N.  A.  Cole,  and  J.  A. 

Stuedemann.  1987.  The  effect  of  fescue 
toxicosis  on  classical  complement  in 
yearling  feedlot  cattle.  Am.  Acad.  Vet. 
Lab.  Diag . (In  press). 


Smith,  W.  L.,  N.  Gay,  J.  A.  Boling,  and  M.  W. 
Crowe.  1986.  Post-grazing  response  of 
steers  previously  consuming  high-  and 
low-endophyte  fescue.  J.  Anim.  Sci. 
63(Suppl . 1 ) : 296 . 

Sprowls,  R.  W.  1987.  Fescue  toxicosis: 

Potential  health  problems  for  stocker  and 
feeder  cattle.  In  Proc.  Symp.  on  Fescue 
Toxicosis  - From  Cow-Calf  to  Slaughter, 
Amarillo,  TX,  April  15,  1987. 

Stoszek,  M.  J.,  J.  E.  Oldfield,  G.  E.  Carter, 
and  P.  H.  Weswig.  1979.  Effect  of  tall 
fescue  and  quackgrass  on  copper  metabolisTi 
and  weight  gains  of  beef  cattle.  J.  Anim. 
Sci.  48:892. 

Stuedemann,  J.  A.,  and  C.  S.  Hoveland.  1988. 
Fescue  endophyte:  History  and  impact  on 
animal  agriculture.  J.  Prod.  Agric.  1:39. 

Stuedemann,  J.  A.,  T.  S.  Rumsy,  J.  Bond,  S.  R. 
Wilkinson,  L.  P.  Bush,  D.  J.  Williams,  and 
A.  B.  Caudle.  1985a.  Association  of 
blood  cholesterol  with  occurrence  of  fat 
necrosis  in  cows  and  tall  fescue  summer 
toxicosis  in  steers.  Am.  J.  Vet.  Res. 
46:1990. 

Stuedemann,  J.  A.,  S.  R.  Wilkinson,  D.  P. 
Belesky,  0.  J.  Devine,  D.  L.  Breedlove, 

F.  N.  Thompson,  C.  S.  Hoveland,  and 
H.  Ciordia.  1985b.  Residual  effects  of 
high  endophyte  infected  KY-31  tall  fescue 
on  steer  performance  and  behavior.  J. 
Anim.  Sci.  61(Suppl.  1):333. 


66 


TABLE  1.  EFFECT  OF  PREVIOUS  FORAGE  ON  FEEDLOT  PERFORMANCE  (HANCOCK  AND 
WILLIAMS,  1985). 


Item 

Fescue 

Orchardgrass 

Bromegrass 

Daily  weight  gain,  kg 

Days  0-28 

Days  0-112 

0.81 

1.38 

0.91 

1.29 

1.00 

1.25 

Dry  matter  intake,  kg/hd/d 

Days  0-28 

Days  0-112 

7.2 

8.5 

7.3 

8.5 

7.4 

8.6 

Feed/gain  ratio 

Days  0-28 

Days  0-112 

9.68 

6.60 

10.21 

6.90 

7.93 

7.06 

TABLE  2.  FEEDLOT  PERFORMANCE 
ENDOPHYTE-FREE  FESCUE  PASTURE 

OF  STEERS  FROM  ENDOPHYTE-INFESTED  AND 

IN  OKLAHOMA  (LUSBY,  1988). 

Item 

Infested 

Infested  + clover 

Endophyte- free 

No.  steers 

27 

19 

26 

Weight  off  pasture,  kg 

343a 

399b 

398b 

Weight  gains  6 days  on 
ryegrass 

17. 3a 

17. 7a 

5.9b 

Shipping  shrink,  kg 

13.6 

21.8 

16.8 

Daily  gains  in  feedlot,  kg 

Days  0-49 

Days  0 - 117 

Ship  to  117 

2.20a 

1.79 

1.66a 

1.96b 

1 • 72. 
1.52b 

2 . 10ab 

1.73. 

1.57b 

a.b 

Means  without  a common  superscript 

differ  (P  «=  .05). 

TABLE  3.  INFLUENCE  OF  GRAZING  ENDOPHYTE 
ON  FEEDLOT  PERFORMANCE  OF  STEERS  STARTED 
ARKANSAS  (PIPER  ET  AL . , 1987). 

-INFESTED  OR  ENDOPHYTE-FREE  FESCUE 
ON  FEED  IN  JULY  OR  OCTOBER  IN 

Item 

Infested  Endophyte-free 

Trial  1 - October 

Weight  gain  on  pastures,  kg  27 

Feedlot  daily  gain,  kg  1.26 

Serum  prolactin  day  0, 

ng/ml  8 


38 

1.05 


b 


22 


Trial  2 - July 

Feedlot  daily  gain,  kg  0.91  1.00 

Serum  prolactin  day  0, 

ng/ml  80 


137 


TABLE  4.  FIFTY-SIX-DAY  FEEDLOT  PERFORMANCE  OF  STEERS  FROM 
ENDOPHYTE-INFESTED  AND  ENDOPHYTE-FREE  PASTURES  IN  KENTUCKY  (SMITH  ET  AL . , 
1986)  . 


Item 

Infested 

Endophyte-free 

Pasture  daily  gains,  kg 

0.49 

0.93 

Feedlot  starting  wt . , kg 

349 

445 

Initial  rectal  temp.,  C 

41. 0a 

39.9 

Dry  matter  intake,  kg 

7.71 

8.83 

Daily  gain,  kg 

1.18 

!.°g 

Feed/gain  ratio 

6.6a 

8.1 

TABLE  5.  FEEDLOT  PERFORMANCE 

OF  ANGUS  STEERS 

FROM  LOW-,  MODERATE- 

, OR 

HIGH-ENDOPHYTE  PASTURES  SHIPPED  FROM  GEORGIA 

TO  TEXAS  (COLE  ET  AL . 

, 1987). 

Item  Low-fungus  Moderate-fungus  High-fungus 

Trial  1 (October  1985) 

Weight  off  pasture,  kg 

Daily  gain,  kg 

333a 

312b 

283C 

Days  0-28 

Days  0-finish 

0.66 

1 . 4 1 a 

0.98 

1.38a 

1 . 05, 
1.68b 

Feed/gain  ratio 

K 

Days  0-28 

Days  0-finish 

11.7  6a 

6.493 

7.35b 

6.45a 

6.94b 

5.46b 

Trial  2 (July  1986) 

Weight  off  pasture,  kg 

Daily  gain,  kg 

2853 

2763 

257b 

Days  0-28 

1.86 

2.14 

2.18 

Days  0-finish 

1.68 

1.78 

1.85 

Feed/gain  ratio 

Days  0-28 

4.54a 

3-7  9ab 

3 . 36b 

Days  0-finish 

5.78 

5.46 

5.32 

Trial  3 (August  1986) 

Weight  off  pasture,  kg 

Daily  gain,  kg 

321 

310 

302 

Days  0-28 

Days  0-finish 

1.89 

1.863 

2.54 

1.60b 

2.81  , 
1.73ab 

Feed/gain  ratio 

Days  0-28 

3.73 

2.65 

2.42 

Days  0-finish 

5.85 

6.17 

5.95 

Trial  4 (September  1987) 

Weight  off  pasture,  kg 
Daily  gain,  kg 

334 

317 

294 

Days  0-28  ^ 

1.08 

1.61 

2.02 

Days  0-finish 

Feed/gain  ratio 

Days  0-28  ^ 

1.50 

1.80 

1.92 

6.94 

5.58 

4.57 

Days  0-finish 

6.25 

5.75 

5.35 

Linear  effect  (P  < .05). 


68 


FORAGE  MANAGEMENT  IN  AN  INTEGRATED  BEEF- 
FORAGE  SYSTEM  IN  ARKANSAS,  A TOTAL  FARM 
MANAGEMENT  APPROACH 

B.  J.  Hankins^ 


INTRODUCTION 

The  major  objective  of  the  Arkansas  beef- 
forage  management  project  is  to  assemble  an 
integrated  set  of  forage  and  beef  cattle 
management  practices  on  a typical  beef  farm 
in  Arkansas  and  to  test  the  hypothesis  that 
adoption  of  such  a set  of  Extension  recom- 
mendations would  increase  profitability  in  a 
beef  cattle  enterprise.  The  project  began  in 
1984  and  will  continue  through  1989. 

A 425-acre  farm  in  Northwest  Arkansas  was 
chosen  for  the  project.  Nine  University  of 
Arkansas  Extension  and  Research  faculty 
members,  Mr.  Charles  Moreton  who  owns  the 
farm,  and  Mr.  Mike  Hamilton  and  Mr.  Merle 
Gross,  local  county  Extension  agents,  com- 
prise the  committee  that  annually  supervises 
the  project.  They  represent  expertise  in  the 
areas  of  soils,  forages,  animal  sciences, 
agricultural  engineering,  weeds,  and  entomology. 

The  Moreton  farm  is  comprised  of  the  homestead 
of  285  acres  and  a second  173-acre  farm 
located  four  miles  away.  In  1984,  each  of 
the  18  fields  on  the  farm  were  numbered  for 
record  keeping  and  a detailed  inventory 
taken  of  soils,  vegetative  cover,  animals, 
machinery,  fencing,  hay  storage  facilities, 
and  water  availability.  Bermudagrass  was 
found  to  be  the  dominant  tame  forage  species 
on  only  10  acres  of  the  farm's  386  acres  of 
pasture  land.  Tall  fescue  with  a 60  to 
100  percent  endophyte  infection  level  pre- 
dominated on  the  remaining  376  acres.  The 
inventory  also  showed  that  soils  were  fertile 
with  a slightly  acid  pH.  This  was  the  result 
of  15  years  or  more  of  continuous  poultry 
litter  use  for  fertilizer.  Ninety-two  percent 
of  the  acreage  had  an  available  phosphorus 
level  above  120  pounds  per  acre.  Sixty-eight 
percent  of  the  acreage  had  an  available 
potash  level  above  200  pounds  per  acre. 
Seventy-three  percent  of  the  acreage  had  pH 
values  of  greater  than  6.5. 

Mr.  Moreton  began  managing  the  farm  upon  the 
death  of  his  father-in-law  at  about  the  time 
this  project  began.  At  that  time,  it  supported 
228  crossbred  cows,  94  calves,  and  15  bulls. 
Cattle  typically  showed  summer  syndrome 
symptoms,  and  the  average  205-day  adjusted 
calf  weaning  weight  was  283  pounds. 


^•Extension  Agronomist  - Forages,  University 
of  Arkansas  Cooperative  Extension  Service, 
P.  0.  Box  391,  Little  Rock,  AR  72203. 


ACCOMPLISHMENTS 

Three  major  forage-related  goals  were  chosen 
at  the  beginning  of  the  project.  They  were: 
(1)  to  improve  hay  quality,  (2)  to  increase 
the  acreage  of  warm  season  forage,  and  (3)  to 
upgrade  the  cool  season  forage. 

Hax 

Hay  quality  was  improved  by  using  more  timely 
harvests  and  by  converting  from  low  to  high 
quality  forage  species.  Crude  protein  content 
in  large  round  bales  was  increased  as  much  as 
six  and  TDN  as  much  as  eight  percentage 
points  since  1984.  Hay  in  1986  tested 
11  percent  crude  protein  and  55.7  percent 
TDN.  Further  improvement  in  quality  and  a 
reduction  in  dry  matter  loss  are  anticipated 
as  new  hay  storage  facilities  are  constructed 
in  1988. 

Increasing  Warm  Season  Acreage 

In  Northwest  Arkansas,  warm  season  forage 
species  should  predominate  on  one-third  of 
the  acreage  of  beef  cattle  farms.  Bermuda- 
grass predominated  on  less  than  three  percent 
of  the  pasture  acreage  of  the  Moreton  farm  in 
1984.  In  1987  it  comprised  23  percent.  The 
conversion  from  cool  to  warm  season  forage 
species  (tall  fescue  to  bermudagrass)  was 
accomplished  by  two  methods. 

Forty-five  acres  of  tall  fescue-common 
bermudagrass  mixture  were  converted  to  acres 
dominated  by  common  bermudagrass  in  a two- 
year  project  by  annually  discriminating 
against  tall  fescue  and  favoring  bermudagrass. 
This  was  done  with  herbicide  applications  in 
late  March  followed  by  a fertilizer  treatment 
in  late  June.  In  1985  the  herbicide  treatment 
was  two  pounds  of  Atrazine  per  acre;  in  1986 
it  was  one  pound  of  Atrazine  mixed  with  one 
pint  of  Paraquat  per  acre.  In  1985  the 
fertilizer  treatment  was  1,500  gallons  of 
liquid  hen  litter  per  acre;  in  1986  it  was 
100  pounds  of  34-0-0  plus  100  pounds  of  0-0-60 
commercial  fertilizers  plus  2,600  gallons  of 
liquid  hen  litter  per  acre.  The  out-of- 
pocket  cost  for  these  materials  and  their 
applications  was  $21.76  per  acre  in  1985  and 
$38.70  in  1986. 

Hybrid  bermudagrass  establishment  was 
accomplished  in  1985-86  on  20  acres  of 
field  8 by  first  fall  plowing  to  kill  fescue. 
Wheat  was  planted  at  an  expense  of  $31.64  per 
acre.  Forage  valued  at  $35.55  per  acre  was 
grazed  by  replacement  heifers  from  the  wheat 
during  winter  and  spring.  Then  the  wheat 
stubble  was  plowed  and  Midland  bermudagrass 
sprigged  in  May  of  1985  at  a cost  of  $61.50 
per  acre. 

Upgrading  Tall  Fescue 

Renovation  of  97  acres  of  endophyte-infected 
fescue  has  been  accomplished  on  five  fields 
by  either  (1)  grazing  close,  then  overseeding 


69 


with  a mixture  of  red  and  white  clover, 

(2)  killing  the  fescue  and  reseeding  with 
endophyte-free  fescue  and  clover,  or 

(3)  killing  the  fescue  and  reseeding  with 
orchardgrass  and  clover.  A no-till  drill  was 
used  to  plant  the  new  forage  species  into 
dead  sod  without  conventional  seedbed 
preparation. 


A summary  of  the  practices  used  for  each 
field  is  abridged  in  Table  1 below.  The 
remaining  unrenovated  acreage  of  endophyte- 
infected  tall  fescue  on  the  farm  (lo9  acres) 
has  been  fertilized  according  to  soli  test 
recommendations  and  harvested  for  hay  and/or 
grazed  rotationally  in  accordance  with  approved 
Extension  recommendations. 


Table  1.  Major  Practices  and  Materials  Used  to  Renovate  Endophyte-Infected  Tall  Fescue  Pastures 


F ield 

Number 

and 

Acres 

Original 

Forage 

Cover 

Dominant 

Forage 

Cover  After 
Renovation 

Major  Materials  and  Practices 
Used  in  Pasture  Renovation/A 

Out-of-Pocket 

Costs 

Per  Acre 
($) 

Date 

Accomplished 

11 

100% 

50%  Forager 

1. 

1 qt.  Roundup 

fescue 

2. 

15  lbs.  Forager  fescue 

17  acres 

KY31  fescue 

40%  clover 

3. 

8 lbs.  Kenland  red  clover 

63.47 

Fall  1984 

and  weeds 

10%  other 

4. 

2 lbs.  Regal  Ladino  clover 

5. 

$4. 00/A  drill  rental 

6. 

$6. 29/A  interest 

3A 

100% 

Same  as 

1. 

Grazed  to  2"  stubble  height 

25.62 

Fall  1984 

13  acres 

KY31 

before 

2. 

8 lbs.  Redland  II  red  clover 

fescue  and 

+ 15% 

3. 

2 lbs.  Regal  Ladino  clover 

weeds 

clover 

4. 

$4. 00/A  drill  rental 

5. 

$2. 30/A  interest 

3B 

100% 

Same  as 

1. 

Grazed  to  2"  stubble  height 

41 .83--/ 

Fall  1984 

20  acres 

KY31 

before 

2. 

1.5  T/A  lime 

fescue  and 

+ 20% 

3. 

8 lbs.  Redland  II  red  clover 

weeds 

clover 

4. 

2 lbs.  Regal  Ladino  clover 

5. 

$4. 00/ A drill  charge 

6. 

$3. 77/A  interest 

7 

100% 

100% 

1. 

1 pint  Embark/A  (1985) 

Fall  1986 

30  acres 

KY31 

new 

2. 

2 one  pint  applications  of 

f escue 

fescue 

Paraquat 

and 

varieties 

3. 

9 varieties  of  tall  fescue 

42 . 90^/ 

weeds 

4. 

$4. 00/A  drill  rental 

5. 

$3. 90/A  interest 

12A 

100% 

70%  orchard- 

1. 

1 pint  Embark/A  (1986) 

17  acres 

KY31 

grass 

2. 

2 applications  of  lh  pt. 

fescue 

20%  clover 

Gramoxone/A 

78 . 65—/ 

Fall  1987 

and 

10%  other 

3. 

2T  lime/A 

weeds 

4. 

10  lb.  orchardgrass/A 

5. 

6 lbs.  red  clover/A 

6. 

$4. 00/A  drill  charge 

7. 

$7. 15/A  interest  charge 

Ij  Lime  charge  was  not  amortized. 

2/  Not  included  are  200  lbs. /A  34-0-0,  hay  harvest,  100  lbs.  wheat  seed,  3 lbs.  arrowleaf  clover  seed,  and 
drill  rental  for  overseeding  wheat  grown  in  1985-86  and  225  lbs. /A  34-0-0,  and  1 pt.  Weedmaster/A  in 
1986. 

_3/  Lime  charge  was  not  amortized. 


SUMMARY  OF  RESULTS 

A total  economic  analysis  that  includes  fixed 
costs  has  not  been  run  on  any  of  the  renovated 
fields.  However,  accurate  records  of  major 
variable  costs  were  kept  as  were  hay  yields 
and  the  number  of  days  grazing  for  each 
field.  A summary  of  these  results  is  shown 
as  comparisons  for  three  fields  in  Table  2. 


During  the  first  two  years  of  this  project, 

65  cows  were  culled  and  55  replacement  heifers 
purchased;  the  average  cow  weight  increased 
by  101  pounds;  the  calf  crop  was  increased 
from  75  to  91  percent  (based  on  cows  retained 
in  the  herd) ; the  calf  weaning  weight  was 
increased  by  about  100  pounds;  and  returns 
above  specified  costs  increased  by  almost 
$15,000.00  per  year. 


70 


Table  2.  Comparative  Annual  Returns  From  Three  Pastures 


Net  1987 


F ield 
Number 

Size 

(Acres) 

Current  Forage 
Species 

Out-of-Pocket 

Returns/A 

1986 

Field  Fertilization  Program 

Net 

Out-of-Pocket 

Returns 

3A 

12 

KY31  Tall  Fescue 

Plus  Clover 

$11.78 

200 

lbs.  34-0-0/A 

$19.31 

8A&B 

20 

Midland  Bermudagrass 

$86.50 

400 

0.7 

lbs.  34-0-0 

loads  poultry  litter/A 

$82.60 

11 

17 

Forager  Tall  Fescue 
and  Clover 

$82.60 

200 

lbs.  0-0-60/A 

$73.97 

CONCLUSION 

Net  profits  are  likely  to  be  substantially 
increased  when  an  integrated  system  of  forage 
and  livestock  management  practices  are  in- 


corporated into  a beef  cattle  enterprise. 
This  project  lends  credence  to  management 
practices  developed  by  University  research 
and  recommended  by  the  Cooperative  Extension 
Service. 


71 


GENERAL  BUSINESS  AND  INFORMATION  EXCHANGE 
GROUP  MEETINGS 

MINUTES  OF  BUSINESS  MEETING 
44th  SPFCIC 
May  12,  1988 
Lexington,  Kentucky 


Or.  Don  Ball,  Chairman,  called  the  meeting 
to  order  and  requested  a roll-call  of  states. 

Old  Business 

1.  The  secretary's  report  was  presented  but 
not  read  because  it  had  been  published  in 
proceedings  of  the  43rd  conference. 

2.  Ihe  treasurer's  report  was  accepted  as 
read,  which  indicated  a balance  of  $6,162.77. 
Motion  by  Jim  Kaiser  and  seconded  by  Jim 
Rice.  Report  attached. 

3.  A comment  by  J.  P.  Mueller  about  the 
current  balance  indicating  a need  for 
evaluating  whether  we  were  within  IRS 
guidelines.  Don  Ball  appointed  a committee  of 
J.  Matches  (chair),  Bill  Stringer  and  J.  Green 
to  make  recommendations  at  the  45th  conference 
about  the  "surplus"  balance. 

New  Business 

1.  B.  J.  Hankins  (Arkansas)  reminded  the 
group  that  we  would  meet  in  Little  Rock  in 
early  June  1989.  He  announced  he  is  seeking 
program  and  tour  ideas.  A preliminary 
indication  was  that  a one-half  day  tour  is 
desired . 

2.  Mark  Hussy  of  Texas  read  an  invitation 
from  Dr.  Clark,  Research  Director  TAM,  to  meet 
in  Texas  in  1990.  Motion  to  accept  by  J.  D. 
Burns  and  seconded  by  J.  P.  Mueller.  Motion 
passed . 

3.  H.  Lippke,  chair  of  nominating  committee, 
recommended  Ken  Quesenberry,  as  chairman-elect- 
elect.  Motion  by  J.  D.  Burns  that  he  be 
elected  by  acclamation.  Seconded  by  W. 
McMurphy.  Motion  passed. 


MINUTES  OF  EXECUTIVE  COMMITTEE 
44th  SPFCIC 
May  10,  1988 
Lexington,  Kentucky 


People  Present:  H.  Lippke,  D.  Ball,  J. 
Mosjidis,  R.  Kalmbacher,  T.  Johnson,  J. 
Stuedemann,  L.  Sollenberger,  J.  Green,  W.  Essig 


Discussion  Topics: 

1.  Don  Ball  requested  that  the  names  of 
elected  officers  for  the  work  groups  be 
obtained  and  shared  soon  after  this 
conference. 

2.  Don  Ball  mentioned  that  D.  Belesky  had 
agreed  to  continue  serving  as  Proceedings 
Coordinator  and  would  obtain  a new  bulk 
mail  permit  for  Beckley,  WV. 

3.  Don  Ball  will  appoint  a committee  to 
evaluate  what  to  do  with  the  money  that  is 
presently  in  the  account.  This  decision 
was  brought  about  by  the  fact  that  the 
account  has  over  $6,000  and  there  is  a 
need  to  keep  this  at  a somewhat  lesser 
balance . 

4.  Ball  named  the  nominating  committee 
consisting  of  W.  Faw,  C.  Dougherty,  H. 
Lippke.  They  are  to  select  a nominee  for 
Chairman  Elect-Elect  of  SPFCIC. 

5.  Ball  appointed  W.  Essig,  J.  Stuedemann, 
and  B.  Nelson  to  the  resolution  committee. 

6.  There  was  some  discussion  about  making 
sure  that  complete  programs  were  mailed  in 
advance  of  the  meeting.  Don  Ball  will 
take  the  suggested  timetable  assembled  by 
H.  Lippke  and  develop  an  official  outline 
of  responsibilities  and  calendar  of 
actions  to  be  followed  by  future 
officers.  There  is  a need  for  better 
coordination  among  work  group  program 
chairmen,  the  SPFCIC  program  chairman  and 
local  arrangement  chairman. 


4.  W.  Essig,  chair  of  resolution  committee, 
read  the  resolution  expressing  gratitude  to 
Kentucky  host  and  all  conference  workers.  See 
attached.  Motion  by  Essig  to  send  resolution 
to  appropriate  UK  administration.  Second  by 
W.  McMurphy.  Motion  passed. 

5.  Other  activities  included  passing  of  gavel 
by  Don  Ball  to  Werner  Essig  and  the 
presentation  cf  plaque  to  Don  Ball  by  H. 
Lippke. 

6.  The  meeting  was  adjourned. 

Respectfully  Submitted, 


James  T.  Green 
Secretary /Treasurer 


72 


RESOLUTION  ADOPTED  UNANIMOUSLY  BY  THE  44TH 
ANNUAL  SOUTHERN  PASTURE  AND 
FORAGE  CROP  IMPROVEMENT  CONFERENCE 


SOUTHERN  PASTURE  AND  FORAGE.  CROP 
IMPROVEMENT  CONFERENCE 
EXECUTIVE  COMMITTEE  1989 


WHEREAS,  the  membership  of  the  44th  annual 
Southern  Pasture  and  Forage  Crop  Improvement 
Conference  has  reaped  great  benefits  from  its 
participation  in  the  Conference,  and 
WHEREAS,  such  benefits  could  not  have  been 
attained  without  the  warm,  friendly, 
hospitable  and  concerted  efforts  of  the  staff 
and  administration  of  the  University  of 
Kentucky; 

BE  IT  RESOLVED:  That  this  44th  Conference 
express  its  grateful  appreciation  to  the  staff 
of  the  University  of  Kentucky  for  the  warm, 
friendly  welcome  extended  to  it  and  the  use  of 
the  superior  facilities  provided  during  the 
meeting  in  the  Radisson  Plaza  Hotel  in 
Lexington  and  the  field  trip  to  Spindletop 
Farm; 

That  the  University  of  Kentucky  and  its 
personnel  are  to  be  commended  for  their  aware- 
ness of  agricultural  problems,  particularly  in 
Grassland  agriculture,  and  for  their  leadership 
and  vision  in  attacking  and  solving  the 
problems  in  improvement,  management,  and 
utilization  of  forage  crops  as  we  move  toward 
the  challenges  of  the  future; 

That  special  recognition  is  extended  to 
Or.  Jack  Hiatt,  Chairman,  Department  of 
Agronomy,  Dr.  Virgil  Hays,  Chairman, 

Department  of  Animal  Science,  and  all 
individuals  serving  on  the  Local  Arrangements 
Committee  and  commercial  firms  for  making  our 
stay  in  Kentucky  so  pleasant. 

That  special  recognition  is  due  and 
extended  to  Conference  Arrangements  Chairman 
Or.  Norm  Taylor,  Conference  Chairman  Dr.  Don 
Ball,  Past  Chairman  and  Program  Chairman  Hagen 
Lippke,  and  to  our  esteemed  and  faithful 
Secretary/Treasurer,  Jim  Green. 

THEREFORE,  We  move  that  this  resolution  be 
adopted  by  unanimous  acclamation  and  recorded 
in  the  Minutes;  and  further,  that  a copy  of 
this  resolution  be  sent  to: 

Dr.  David  Rosell,  President,  University  of 

Kentucky 

Dr.  C.  E.  Barnhart,  Dean,  U.K.  College  of 

Agriculture 

Dr.  Milt  Shuffot,  Associate  Dean  for 

Research 

Dr.  Shirley  Phillips,  Associate  Dean  for 

Extension 


H.  W.  Essig 
John  Stuedemann 
Billy  Nelson 


Executive  Officers 

Warner  Essig 
Rob  Kalmbacher 
Kenneth  Quesenberry 
Don  Ball 

Jim  Green 

David  Ba 1 tensperger 
Lynn  Sollenberger 
Bruce  Pinkerton 
Lance  Tharel 
Dave  Belesky 


Cha i rman 
Cha i rman- elect 
Chai rman- elect  elect 
Immediate  Past 
Chairman  and  Program 
Chairman  for  1989 
(45th  Meeting) 
Secretary  Treasurer 

Breeders  Work  Group 
Cha i rman 

Ecology  Pnysiology 
Work  Group  Chairman 
Extension  Work  Group 
Chai rman 

Utilization  Work 
Group  Chairman 
Proceed i ngs 
Coordinator 


Officers  of  1989  Work  Groups 


Breeders  Work  Gro up 
David  Ba 1 tensperger 
Jorge  Mosjidis 

Ecology  and  Physiology 

Lynn  Sollenberger 
Richard  Joost 
Chuck  West 
Vivian  Allen 

Extension  Work  Gro up 

Bruce  Pinkerton 
B.  J.  Hankins 

Troy  Johnson 

Utilization  Work  Group 

Lance  Tharel 
Steve  Schmidt 
John  Stuedemann 


Cha i rman 
Secretary 
Past  Chairman  and 
Program  Director 

Work  Group 

Cha i rman 

Cha i rman  elect 

Secretary 

Past  Chairman  and 

Program  Director 


Cha i rman 

Secretary  and  Program 

Director 

Past  Chairman 


Chairman  and  Program 
Director 
Chai rman- elect 
Secretary  elect 
Past  Chairman 


73 


SPFCIC 

Breeders  Work  Group,  May  11,  1988 

Meeting  called  to  order  by  Jorge  Mosjidis 
11:30  a .m. 

Minutes  approved 

No  Old  Business 

Nomination  of  Mark  Hussey  as  Secretary 
Fleeted  by  acclamation 
- D.  Bal tensperger  moves  to  Chairman 
J.  Mosjidis  moves  to  Program  Coordinator 

Or.  Everett  F.mino,  Administrative  Advisor, 
announce  need  for  renewal  of  Regional 
Information  Exchange  Groups  by  1990. 

Dr.  J.  Preston  Jones,  our  CSRS  representative, 
said  Bob  Conger  did  renewal  last  time. 

Meeting  was  adjourned. 

Secretary, 

0.  0.  Baltensperger 


cc:  Director  of  Exp.  Sta.  Kentucky 

Administrative  Advisor  - Dr.  Emino 
CSRS  Rep.  Dr.  J.  P.  Jones 


ECOLOGY  AND  PHYSIOLOGY  WORK  GROUP 
BUSINESS  MEETING 

The  meeting  was  called  to  order  by  Dr.  Lynn 
Sollenberger  at  12  noon  on  5/11/88.  Dr. 
Sollenberger  announced  the  current  officers  of 
the  group:  Dr.  Vivian  Allen  (VPI,  Program 
Chair  for  the  Lexington  meeting).  Dr.  Lynn 
Sollenberger  (Univ.  of  Florida,  Chair),  and 
Dr.  Richard  Joost  (LSU,  Secretary).  The  chair 
then  opened  the  floor  for  nominations  for 
secretary  of  the  group  for  1988  89.  Chuck 
West  (Univ.  of  Arkansas)  nominated  Wilfred 
McMurphy  (0SU).  The  motion  was  seconded  by 
Rob  Kalmbacher  (Univ.  of  Florida).  Wilfred 
McMurphy  nominated  Chuck  West,  indicating  that 
the  idea  was  to  get  younger  scientists  involved 
in  the  administration  of  the  organization. 

The  motion  was  seconded  by  Rich  Joost.  Chuck 
West  was  elected  by  a majority  vote. 

Dr.  Sollenberger  then  requested 
suggestions  from  the  floor  regarding  topics 
for  next  year's  meeting.  The  following 
suggestions  were  offered: 

1.  Overview  of  soil  fertility/plant  nutrition 
as  it  relates  to  soil/plant/animal  mineral 
uti 1 i zation . 

2.  Forage  components  acting  as  feeding 
stimulants  and  feeding  deferents. 

3.  Possibility  of  discussing  the  present 
state  of  modeling  including  coverage  of 
expert  systems  approaches. 

4.  Coverage  of  the  tall  fescue  endophyte  from 
the  perspective  of  our  current  knowledge 
of  the  toxins  produced  and  management 
considerations  for  endophyte  + and  - 
stands.  This  could  include  the  basic 
physiology  and  ecology  of  the 
endophyte/plant  association. 

Joe  Burns  (NCSU)  made  the  suggestion  that  we 
not  spend  our  entire  time  on  modeling.  There 
was  general  agreement  with  this  comment.  Joe 
Fontenot  (VPI)  indicated  that  the  tall 
fescue/endophyte  association  would  be  a good 
topic  for  a joint  meeting  with  the  Forage 
Utilization  Work  Group.  Dr.  Sollenberger 
informed  the  group  that  they  could  submit  any 
further  ideas  by  mail  to  himself  or  Dr.  Joost 
at  any  time. 

There  was  no  additional  new  business. 

Dr.  Chuck  West  moved  that  the  meeting  be 
adjourned.  Dale  Wolf  (VPI)  seconded  and  the 
meeting  was  adjourned  at  12:30. 


74 


MINUTES  OF  THE  SPFCIC 

FORAGE  UTILIZATION  WORK  GROUP  BUSINESS  MEETING 

Lexington,  KY 
May  11  , 1988 

The  meeting  was  called  to  order  by  President 
John  Stuedemann  at  Lexington,  KY  on  May  11, 
1988.  Current  officers  of  the  Forage 
Utilization  Work  Group  were  introduced  and 
individuals  in  attendance  introduced 
themselves,  stating  their  location  and  area  of 
research  or  interest.  A questionnaire  was 
distributed  asking  for  suggestions  for  program 
topics  for  future  meetings. 

A motion  was  made  and  seconded  to  dispense 
with  the  reading  of  the  minutes  of  the  1987 
business  meeting  held  in  Clemson,  SC.  The 
minutes  were  approved  as  printed  in  the 
proceedings  of  the  43rd  SPFCIC.  The 
nominating  committee  chaired  by  Lance  Tharel 
presented  Dwight  Fisher,  Crop  Science 
Department,  North  Carolina  State  University  as 
Secretary-Elect.  He  was  unanimously  elected. 
Lance  Tharel  will  serve  as  president  for  the 
next  year. 

Nick  Hill  noted  that  a change  in  the 
manuscript  format  for  the  proceedings  needs  to 
be  considered.  The  photo-ready  copy  sheets 
that  have  been  used  in  the  past  do  not  work 
with  many  of  the  new  word  processors  and 
printers.  John  Stuedemann  agreed  to  bring 
this  up  with  the  executive  committee. 

The  motion  was  made  and  seconded  to  adjourn. 


The  following  is  a summary  of  responses  to  the 

questionnaire  on  program  topics: 

1.  Establishment/renovation  of  endophyte-free 
fescue  pastures,  i.e.  methods  advantages 
and  disadvantages  of  each. 

2.  Management  of  newly  established 
endophyte-free  fescue  pastures,  i.e. 
stocking  rate  and  grazing  pressure. 

3.  Should  fescue  (noninfected  or  infected)  be 
planted  separately  or  in  combination  with 
other  perennials?  (i.e.  bermudagrass . ) 

4.  Use  of  pasture  probe  or  pasture  meter  for 
prediction  of  DM/Ac,  and  use  of  these  data 
in  management  systems. 

5.  Methods/markers  of  measuring  intake  on 
warm  and  cool  season  forages, 
implications,  new  techniques  on  horizon. 

6.  Comparison  of  newer  techniques  for 
measuring  short-term  intake,  i.e.  bite 
count,  bolus  size  measurements  vs. 
electronic  measurement  or  bolus 
movement- bol us  size  techniques. 


7.  Ammoniated  hay  toxicity  update,  i.e. 
causes,  symptoms,  prevention. 

8.  Research  of  problems  associated  with 
feeding  ammoniated  hay,  i.e.  storage, 
risks,  are  risks  sufficient  to  warrant  not 
advocating  the  practice  - legal  aspects 
(liability). 

9.  Endophyte  relationships  in  tall  fescue, 
i.e.  insect,  disease  and  drought  tolerance. 

10.  Update:  of  latent  information  of  toxins  or 
chemicals  found  in  tall  fescue  and  their 
plant/animal  relationships. 

11.  Forage  availability  x animal  performance. 

12.  Use  of  microcomputers  and  associated 
software  to  aid  in  analysis  of  forage 
utilization  data. 

13.  New  techniques/ideas  of  data  recording  and 
handling  data. 

14.  Use  of  available  forage  as  a covariate  in 
grazing  research  analysis  is  this  the 
best  term?  Why?  What  other  options 
avai lable? 


Stephen  P.  Schmidt 
Secretary,  1988 


75 


EXTENSION  WORK  GROUP  MINUTES 

The  meeting  of  SPFCIC  Extension  Work  Group 
met  Wednesday,  May  11,  1988  in  Lexington, 
Kentucky.  The  meeting  was  called  to  order  by 
Troy  Johnson. 

Those  in  attendance  were: 

Keith  Edmisten,  Mississippi 
Jim  Woodruss,  South  Carolina 
Jim  Green,  North  Carolina 
Wade  F.  Faw,  Louisiana 
Harlan  E.  White,  Virginia 
Don  Ball,  Alabama 
Troy  Johnson,  Georgia 
B.  J.  Hankins,  Arkansas 
Joe  Burns,  Tennessee 
Monroe  Rasnake,  Kentucky 
J.  Paul  Mueller,  North  Carolina 
Warren  Thompson,  Kentucky 
Garry  Lacefield,  Kentucky 

Dr.  Johnson  reported  that  as  a result  of 
illness,  Dr.  Bruce  Pinkerton  (Program  Chairman) 
was  unable  to  attend. 

Chairman  Johnson  presented  a slate  of 
officers  for  1989  including:  B.  J.  Hankins, 
Secretary;  Garry  Lacefield,  Program  Chairman; 
and  Bruce  Pinkerton,  Chairman.  Motion  by  Paul 
Mueller,  second  by  Wade  Faw,  "that  officers  be 
elected  by  accl imation" . Motion  carried. 

Chairman  Johnson  led  discussion  on  role, 
need  and  direction  of  Extension  Work  Group. 
Those  in  attendance  agreed  that  the  Extension 
Work  Group  had  provided  an  important  and 
unique  opportunity  for  Extension  Forage 
Workers  over  the  years.  A strong  desire  was 
expressed  by  those  in  attendance  to  continue 
the  work  group  with  options  for  joint  session 
as  needed.  Work  Group  members  pledged  support 
in  assisting  Program  Chairman  and  encouraged 
greater  participation  in  getting  materials 
into  Proceedings. 

Motion  by  Monroe  Rasnake,  second  by  Don 
Ball  that  "secretary  and  program  chairman 
position  be  combined  and  that  position  be 
occupied  by  representative  from  host  state". 
Motion  carried. 

Meeting  adjourned  at  9:10. 


SOUTHERN  PASTURE  AND  FORAGE  CROP 
IMPROVEMENT  CONFERENCE 
1988  Financial  Statement 

Income  Expend e Balar :e 


04/04/87 

Balance  on  hand  at 
Wachovia  Bank  and  Trust 

Account  #6261  206760  3,738.27 

08/05/87 

Deposit  of  balance 

from  43rd  meeting 

in  South  Carolina  2,167.71 

04/06/88 
Interest  on  bank 
account  for  04/87- 
04/06/88  285.90 

04/28/88 

Plaque  for  Chairman 

(Tro  Craft  Studios)  29.11 


05/04/88  Balance  on  Hand  $6,162.77 

Respectfully  submitted  by  James  T.  Green,  Jr. 


Respectively  submitted, 


Garry  Lacefield 
Secretary 


76 


i 4th  SPECIE  REGISTRANTS 

Vivian  G.  Allen 
Agronomy  Department 
2236  Smyth  Hall 
Virginia  Tech 
Blacksburg,  VA  24061 

Don  Ball 
Extension  Hall 
Auburn  University 
Auburn,  AL  36849 

David  B.  Baltensperger 
Agronomy  Department 
304  Newell  Hall 
University  of  Florida 
Gainesville,  FL  32611 

Jose  De  Battista 
Department  of  Agronomy 
Plant  Sciences  Bldg.,  Rm.  311 
University  of  Georgia 
Athens,  GA  30602 

David  P.  Belesky 
Appalachian  Soil  & Water 

Conservation  Res.  Laboratory 
Airport  Road,  Box  867 
Beckley,  WV  25802-0867 

Roy  Blaser 
704  York  Drive 
Blacksburg,  VA  25061 

Paul  R.  Beuselinck 
USDA  ARS 
207  Waters 

University  of  Missouri 
Columbia,  M0  65211 

Joe  Bouton 

Agronomy  Department 

311  Miller  Plant  Sciences  Bldg. 

University  of  Georgia 

Athens,  GA  30602 

David  Bransby 
Agronomy  and  Soils 
Auburn  University 
Auburn,  AL  36849 

Wi  1 1 iam  A.  Brock 
Rt.  2,  Box  150 
Newton,  MD  39345 

Joe  D.  Burns 
P.0.  Box  1071 
University  of  Tennessee 
Knoxville,  TN  37901 

Joe  C 8urns 
3627  Gardner  Hall 
Box  7614 

North  Carolina  State  Univ. 

Raleigh,  NC  27695-7614 


Paul  B.  Burrus,  Jr. 

USDA  ARS 

Agronomy  Department 
University  of  Kentucky 
l.exington,  KY  405^6-0091 

Lowell  Bush 
Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546-0091 

Scott  Carr 

Animal  Science  Department 
Room  276 
Virginia  Tech 
Blacksburg,  VA  24061 

D.  S.  Chamblee 

Crop  Science  Department 

Box  7620 

North  Carolina  State  University 
Raleigh,  NC  27650 

Allan  Chestnut 

Department  of  Animal  Science 
P.0.  Box  1071 
University  of  Tennessee 
Knoxville,  TN  37901 

Peter  Clark 
Agronomy  and  Soils 
Auburn  University 
Auburn,  AL  36849 

Andy  Cole 

Bovine  Respiratory  Disease  Research 
P.0.  Drawer  10 
Bushland,  TX  79012 

Sam  Coleman 
USDA -ARS 
Box  1199 

El  Reno,  OK  73036 

Glenn  Collins 
Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546  0091 

Michael  Collins 
Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546  0091 

Mark  Dahmer 
Soil  and  Crop  Science 
Texas  A & M University 
College  Station,  TX  77843 

R.  L.  Dalrymple 
Nobel  Foundation 
P.0.  Box  2180 
Ardmore,  AR  73401 

Roy  Deason 

Rt.  1 , Box  1098 

Carnesville,  GA  30521 


Jim  Dobson 

Georgia  Mtn.  Branch  Station 
Rt.  1 , Box  45 
Blairsville,  GA  30512 


77 


Chuck  Dougherty 
Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546  0091 

Wanzer  Drane 

Research  Data  Analysis 

Auburn  University,  AL  36849 

Craig  Edminster 
International  Seeds,  Inc. 
P.0.  Box  168 
Halsey,  OR  97348 


Keith  Edmisten 
Room  190 

3825  Ridgewood  Road 
Jackson , MS  3921 1 

Georgia  C.  Eizenga 
USDA  ARS 

Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546-0091 

Evert  R.  Emino 
1022  McCarty  Hall 
University  of  Florida 
Ga i nes vi le , F L 3621 1 

Werner  Essig 
Animal  Research  Center 
Mississippi  State  University 
Mississippi  State,  MS  39762 

Jack  Fvans 

Pacific  Scientific  Co. 

P.0.  Box  144 
Onalaska,  W1  54650 

Wade  F . Faw 
255  Knapp  Hall 
Louisiana  State  University 
Baton  Rouge,  LA  70803  1900 

Dwight  Fisher 

Crop  Science  Department 

Box  7620 

North  Carolina  State  University 
Raleigh,  N.C.  27650 

J.  P.  Fontenot 

Department  of  Animal  Science 
Virginia  Tech 
Blacksburg,  VA  24061 

David  Forbes 
Ag  Engineering 
University  of  Kentucky 
Lexington,  KY  40546 

Walter  Graves 

Univ.  of  Calif.  Coop.  Ext. 

Bldg.  4,  5555 
Overland  Ave. 

San  Diego,  CA  92123 


James  T.  Green 

Crop  Science  Department 

Box  7620 

North  Carolina  State  Univ. 
Raleigh,  NC  27650 

Roger  N.  Gates 
P.0.  Box  466 
Iberia  Res.  Stn. 

Jeanerette,  LA  70544 

B.  J.  Hankens 
P.  0.  Box  391 
Littlerock,  AR  72203 

A.  J.  Hiatt 
Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546-0091 

Nick  Hill 

Agronomy  Department 
Miller  Plant  Sciences  Bldg. 
University  of  Georgia 
Athens,  GA  30602 

Bill  Holloway 
Texas  A & M 

Agricultural  Research  Center 
Uvalde,  TX  78801 

Carl  S.  Hoveland 
Agronomy  Department 
Miller  Plant  Sciences  Bldg. 
University  of  Georgia 
Athens,  GA  30602 

Mark  A.  Hussey 

Rm  430-Soil  snd  Crop  Sciences 
Texas  A&M  University 
College  Station,  TX  77843 

David  Hutcheson 

Texas  A & M Agr  Res  and  Ext  Ctr 
6500  Amarillo  Blvd.  West 
Amarillo,  TX  79106 

Troy  Johnson 
Agronomy  Department 
Miller  Plant  Sciences  Bldg. 
University  of  Georgia 
Athens,  GA  30602 

Richard  Joost 
Agronomy  Department 
104  M.B.  Sturgis 
Louisiana  State  University 
Baton  Rouge,  LA  70803 

J.  Preston  Jones 
UDSA-CSRS 
209  JSM  Building 
Washington,  D.C.  20250-2200 


Rob  Kalmbacher 

Univ.  of  Florida  - Ona  AREC 

Rt.  1 Box  62 

Ona,  FL  33865 


78 


J.  Kaiser 

Agronomy  Department 
University  of  Illinois 
Urbana.  IL  61801 

Garry  D.  Lacefield 

Univ.  of  Ky  Res.  & Ed.  Cntr. 

P.0.  Box  469 

Princeton,  KY  42445 

K.  T.  Leath 
USDA-ARS 

U.S.  Regional  Pasture  Res.  Lab. 
University  Park,  PA  16802 

Hagen  Lippke 

Texas  Agric.  Expt.  Sta. 

P.0.  Box  728 
Angleton,  TX  77515 

Gill  Lovell 
USDA  ARS 

Reg.  Plant  Introduction  Station 
Georgia  Experiment  Station 
Experiment,  GA  30212 

Jerry  Matches 

Plant  & Soil  Science  Dept. 

Texas  Tech  Univ. 

Lubbock,  TX  79409 

Wilfred  McMurphy 
Agronomy  Department 
Oklahoma  State  University 
Stillwater,  OK  74078 

Ronald  L.  Mitchell 
The  Nobel  Foundation 
P.0.  Box  2180 
Ardmore,  OK  73402 

Monty  Montgomery 
Animal  Science  Department 
University  of  Tennessee 
Knoxville,  TN  37901 


Jorge  A.  Mosjidis 
Agonomy  and  Soils  Department 
Auburn  University 
Auburn,  AL  36849 

J.  Paul  Mueller 
Crop  Science  Dept. 

Box  7620 

North  Carolina  State  University 
Raleigh,  NC  27695-7620 

Billy  Nelson 
Agronomy  Department 
Louisiana  State  University 
Baton  Rouge,  LA  70803 

Jeff  Pedersen 
USDA-ARS 

Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546-0091 


Mike  Phillips 
Agronomy  Department 
University  of  Arkansas 
Fayetteville,  AR  72701 

Bruce  W.  Pinkerton 
Agronomy  Department 
P & AS 

Clemson  University 
Clemson,  SC  29634  0359 

Ken  Quesenberry 
Agronomy  Department 
University  of  Florida 
Gainesville,  FL 

Monroe  Rasnake 

Univ.  of  Ky  Res.  & Ed.  Cntr. 

P.0.  Box  469 

Princeton,  KY  42445 

John  H.  Reynolds 

Plant  & Soil  Science  Dept. 

P.0.  Box  1071 
Univ.  of  Tennessee 
Knoxville,  TN  37901 

Harold  B.  Rice 
Univ.  of  Kentucky 
Robinson  Substation 
Quicksand,  KY  41363 

James  S.  Rice 
Agronomy  & Soils  Dept. 

Clemson  University 
Clemson,  SC  29634  0359 

Marvin  E.  Riewe 
Texas  Agric.  Exp.  Sta. 

P.0.  Box  728 
Angleton,  TX  77515 

E.  G.  Rhoden 
201  Farm  Mech  Bldg 
Tuskeegee  University 
Tuskeegee,  AL  36088 

Monte  Rouquette 

Texas  A & M Agr  Res  and  Ext  Ctr 
Overton,  TX  75684 

Tommy  G.  Sanders 
Costal  Plain  Branch  MAFFS 
Route  2,  Box  150 
Newton,  MS  39345 

Steve  Schmidt 

Animal  Science  Department 

Auburn  University,  AL  36849 

Dwight  Seman 
USDA  ARS 

Southern  Piedmont  Conservation 
Research  Lab 

Highway  53  P.0.  Box  555 

Watkinsvi 1 le,  GA  30677 


79 


Richard  Senft 

South  Central  Family  Farms  Res 
Rt.  2,  Box  144  A 
Highway  23  South 
Booneville,  AR  72929 

Malcolm  R.  Siegel 
Plant  Pathology  Dept. 

University  of  Kentucky 
Lexington,  KY  40546 

L.  E.  Sollenberger 
Aqronomy  Department 
Bldg  447,  I FAS  0681 
University  of  Florida 
Gainesville,  Florida  32611 

Richard  R.  Smith 
USDA-ARS 

U.S.  Dairy  Forage  Res.  Ctr. 

1925  Linden  Drive 
Madison,  WI  53706 

Harold  Stern 
Pacific  Scientific  Co. 

409  Shelburne  Dr. 

Carol  Stream,  IL  60188 

William  C . Stringer 
Agronomy  & Soils  Dept. 

204  P & A Bldg. 

Clemson  University 
Clemson,  SC  29634  0359 

John  Stuedemann 
USDA  ARS 

Southern  Piedmont  Conservation 
Research  Lab 

Highway  53,  P.0.  Box  555 
Watkinsvi 1 le,  GA  30677 

Lance  Tharel 

South  Central  Family  Farms  Res. 
Rt.  2,  Box  1 44  A 
Highway  23  South 
Booneville,  AR  72929 

Norm  Taylor 
Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546-0091 

T.  H.  Taylor 
Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546 

Thomas  Terri  1 

175  Buena  Vista  Ave. 

Athens,  GA  30601 

Bill  Templeton 
800  Brook  Hill  Drive 
Lexington,  KY  40502 

Warren  Thompson 
121  Dantzler  Ct. 

Lexington,  KY  40503 


Ann  Marie  Thro 
Agronomy  Department 
Louisiana  State  University 
Baton  Rouge,  LA  70803-2110 

Dan  Undersander 
Agronomy  & Soils  Dept. 

204  P & A Bldg. 

Clemson  University 
Clemson,  SC  29634  0359 

Chuck  West 
Altheimer  Lab 
Route  1 1 , Box  83 
University  of  Arkansas 
Fayetteville,  AR  72703 

Elizabeth  G.  Williams 
Agronomy  Department 
University  of  Kentucky 
Lexington,  KY  40546  0091 

Harlan  E.  White 
Agronomy  Department 
Virginia  Tech 
Blacksburg,  V A 24061 

Stanley  R.  Wilkinson 
USDA  ARS 
SPCRC , Box  555 
Watkinsvi 1 le , GA  30677 

Dale  Wolf 

Agronomy  Department 
Virginia  Tech 
Blacksburg,  VA  24060 


80