Name: GLÊYCE PEREIRA SANTOS
Publication date: 08/12/2020
Advisor:
Name |
Role |
|---|---|
| EDILSON ROMAIS SCHMILDT | Advisor * |
Examining board:
| Name |
Role |
|---|---|
| EDILSON ROMAIS SCHMILDT | Advisor * |
| SARA DOUSSEAU ARANTES | Internal Examiner * |
Summary: SANTOS, Gleyce Pereira; M.Sc.; Federal University of Espírito Santo; December
2020; Biometrics in conilon coffe clones; Advisor: Edilson Romais Schmildt,
Coadvisor: Omar Schmildt.
In carrying out scientific experiments, statistics is an important point, being one of its
bases. Research involving statistical analysis can be used as a basis for other
research, this field being little explored in order to involve agricultural crops. For a
researcher who wants to follow the leaf development of a plant, the most used ways
to obtain the leaf area require a high expenditure of time and investment in expensive
equipment, making the experiment idle and difficult to perform. For agricultural
experimentation there is a need for the researcher to initially establish the design, the
number of treatments, repetitions and the size of the plots, to predict the physical space
and material expenditure. This planning is not always easy due to the lack of research
that points out the adequate plot size, mainly in experiments in the seedling phase.
The literature points to some contribution to this aspect with Arabica coffee seedlings,
however, not yet with Conilon coffee. For this, the biometric analysis on the coffee
plants of the Conilon clone LB1 can be used as a basis for the facilitation of future
research in the area. From determining the optimal plot size for an experiment, to
modeling the leaf area, scientific research is extremely useful. This work aims to
biometric studies with Conilon seedlings (Coffea canephora Pierre ex Froehner) to
cultivate LB1, divided into two chapters. In chapter 1, the determination of the optimalvii
plot size for experiments with seedlings was carried out, using the methodology
proposed by Hatheway in which the coefficient of variation values and the
heterogeneity index were obtained by bootstrap simulation with replacement. In the
installation of conilon clone LB1 coffee experiments, in randomized blocks with 7 to 40
treatments and three repetitions, plots containing nine seedlings are sufficient to
identify significant differences between treatment means on non-destructive
characters at 5% probability and difference between mean treatments 30% of the
general average of the experiment. For destructive characteristics, plots containing 14
seedlings are sufficient to identify significant differences between treatment means. In
chapter 2, the leaf area was modeled for coffee seedlings produced in two containers
(bags and tubes) from linear dimensions of the leaf surface. The length (C), the width
(L), the observed leaf area (AFO) and the product of multiplying the length by the width
(CL) of all leaves were measured. The covariance analysis was carried out using the
student's t test at 5% probability to verify the possibility of using a single equation model
that estimates the leaf area of the seedlings in both containers. The mean error (E),
mean absolute error (AME), root of the mean error square (RQME) and Willmott index
(d) were used as validation criteria. The leaf area of coffee seedlings cultivar LB1
produced in bags and tubes can be estimated by the equation AFE = 1.157364 +
0.646417 (CL).
Key words: Coffea canephora Pierre ex Froehner, experimental precision,
experimental design.
