A Simple Procedure for Constructing Experiment Designs with Incomplete Blocks of Sizes 2 and 3

In order to maximize control of heterogeneity within complete blocks, an experimenter could use incomplete blocks of size k = 2 or 3. In certain situations, incomplete blocks of this nature would eliminate the need for such spatial types of analyses as nearest neighbor. The intrablock efficiency factors for such designs are relatively low. However, with recovery of interblock information, FEDERER and SPEED (1987) have presented measures of design efficiency factors which demonstrate that efficiency factors approach unity for certain ratios of the intrablock and interblock variance components. Hence with recovery of interblock information, even incomplete block designs with k = 2 or 3 have relatively high efficiency factors. The reduction in the intrablock error variance over the complete block error variance in many situations will provide designs with high efficiency. A simple procedure for constructing incomplete blocks of sizes 2 and 3 is presented. It is shown how to obtain additional zero-one association confounding arrangements when v = 4 t, t an integer, and for v = pk, k ≤ p. It is indicated how to do the statistical analysis for these designs.     

Increased gene coverage and Alu frequency in large linkage disequilibrium blocks of the human genome

The human genome has linkage disequilibrium (LD) blocks, within which single-nucleotide polymorphisms show strong association with each other. We examined data from the International HapMap Project to define LD blocks and to detect DNA sequence features inside of them. We used permutation tests to determine the empirical significance of the association of LD blocks with genes and Alu repeats. Very large LD blocks (>200 kb) have significantly higher gene coverage and Alu frequency than the outcome obtained from permutation-based simulation, whereas there was no significant positive correlation between gene density and block size. We also observed a reduced frequency of Alu repeats at the gaps between large LD blocks, indicating that their enrichment in large LD blocks does not introduce recombination hotspots that would cause these gaps.     

Middle Stone Age stratigraphy and excavations at Die Kelders Cave 1 (Western Cape Province, South Africa): the 1992, 1993, and 1995 field seasons

Die Kelders Cave 1, first excavated under the direction of Franz Schweitzer in 1969-1973, was re-excavated between 1992 and 1995 by a combined team from the South African Museum, SUNY at Stony Brook, and Stanford University. These renewed excavations enlarged the artefactual and faunal samples from the inadequately sampled and less intensively excavated lower Middle Stone Age (MSA) layers, increased our understanding of the complex site formation processes within the cave, enlarged the hominid sample from the MSA deposits, and generated ESR, TL, and OSL dates for the MSA layers. Importantly, these new excavations dramatically improved our comprehension of the vertical and lateral characteristics of the MSA stratigraphy. Surface plotting of the MSA layers has led to the identification of at least two major zones of subsidence that significantly warped the layers, draping some along the eroding surface contours of major blocks of fallen limestone roof rock. A third zone of subsidence is probably present in the older excavations. Dramatic roof falls of very large limestone blocks occurred at least twice-once in the middle of Layer 4/5 where the roof blocks were only slightly weathered after collapse, and at the top of Layer 6 where the blocks weathered heavily after collapse, producing a zone of decomposed rock around the blocks. Many of the sandy strata are cut by small and localized faults and slippages. All of the strata documented by Schweitzer's excavations are present throughout the exposed area to the west of his excavated area, where many of them thicken and become more complex. Layer 6, the thickest MSA layer, becomes less diagenetically altered and compressed to the west.     

Some Bounds on Parameters of Efficiency-Balanced Block Designs

Bounds on the eigenvalue of the C -matrix with respect to R and the number of blocks for efficiency-balanced block designs are given. AMS 1970 Subject Classification: 62K10     

Maximin D-optimal designs for longitudinal mixed effects models

In this article, the optimal selection and allocation of time points in repeated measures experiments is considered. D-optimal cohort designs are computed numerically for the first- and second-degree polynomial models with random intercept, random slope, and first-order autoregressive serial correlations. Because the optimal designs are locally optimal, it is proposed to use a maximin criterion. It is shown that, for a large class of symmetric designs, the smallest relative efficiency over the model parameter space is substantial.     

More Powerful Likelihood Ratio Tests for Isotonic Binomial Proportions

Binomial group testing involves pooling individuals into groups and observing a binary response on each group. Results from the group tests can then be used to draw inference about population proportions. Its use as an experimental design has received much attention in recent years, especially in public‐health screening experiments and vector‐transfer designs in plant pathology. We investigate the benefits of group testing in situations wherein one desires to test whether or not probabilities are increasingly ordered across the levels of an observed qualitative covariate, i.e., across strata of a population or among treatment levels. We use a known likelihood ratio test for individual testing, but extend its use to group‐testing situations to show the increases in power conferred by using group testing when operating in this constrained parameter space. We apply our methods to data from an HIV study involving male subjects classified as intraveneous drug users.     

The design and analysis of breeders' trials of sugar beet (Beta vulgaris) using two-dimensional blocking structures

A method for generating two-dimensional blocking designs to fit any shape and size of experimental area is presented. The method takes an alpha design appropriate to the experimental area and imposes additional blocking perpendicular to the alpha blocks. Improvements to the design are then made by repositioning entries within the alpha blocks. Although this method of construction is less sophisticated than for other two-dimensional designs, the designs are particularly suited to large scale breeders trials where no alternative two-dimensional design may exist. The designs have been used for 3 yr in a sugar beet breeding programme, and have given improvements in efficiency over one-dimensional alpha designs.     

Mitotic chromosomes are constrained by topoisomerase II–sensitive DNA entanglements

We have analyzed the topological organization of chromatin inside mitotic chromosomes. We show that mitotic chromatin is heavily self-entangled through experiments in which topoisomerase (topo) II is observed to reduce mitotic chromosome elastic stiffness. Single chromosomes were relaxed by 35% by exogenously added topo II in a manner that depends on hydrolysable adenosine triphosphate (ATP), whereas an inactive topo II cleavage mutant did not change chromosome stiffness. Moreover, experiments using type I topos produced much smaller relaxation effects than topo II, indicating that chromosome relaxation by topo II is caused by decatenation and/or unknotting of double-stranded DNA. In further experiments in which chromosomes are first exposed to protease to partially release protein constraints on chromatin, ATP alone relaxes mitotic chromosomes. The topo II–specific inhibitor ICRF-187 blocks this effect, indicating that it is caused by endogenous topo II bound to the chromosome. Our experiments show that DNA entanglements act in concert with protein-mediated compaction to fold chromatin into mitotic chromosomes.     

Structural alterations in nerve fibers produced by hypotonic and hypertonic solutions

In sections of KMnO(4)-fixed, developing mouse sciatic nerves, the central gap of mesaxons in myelinating fibers is normally closed with close apposition of the outside approximately 20 A dense strata of the two approximately 75 A Schwann cell membranes. The two combined outside strata make the intraperiod line bisecting each myelin lamella. The approximately 150 A mesaxon is elaborated spirally around the axon in either a right hand or left hand spiral, and its inside (cytoplasmic) approximately 20 A strata in apposition form the major dense lines of myelin. In hypotonic solutions the lamellae of adult frog sciatic myelinated fibers split apart along the outside membrane strata apposed at the intraperiod line throughout the spiral. Under similar conditions the inside (cytoplasmic) strata of the membranes, in apposition at the major dense lines, do not separate. The approximately 150 A membranous structure resulting from this is called an "internal compound membrane." The double membranes of normal and control frog sciatic unmyelinated fibers have a central gap approximately 100 to 150 A wide. After soaking in 4 to 10 times normal strength Ringer solution or 10 N sucrose-Ringer solution, this gap closes and a membranous structure approximately 150 A wide resembling developing mouse mesaxons results. This is designated by the term "external compound membrane." The latter membranes resemble internal compound membranes, but their central dense zones, each consisting of two apposed outside membrane strata, are less dense.     

The characteristics of chromosomal distribution and of the variability of the multiply repeating DNA fraction of the genome in Bos taurus L.]

The uniform distribution of satellite DNA II and IV has been revealed using in situ hybridization and differential staining in centromeric regions of autosomes. The sex chromosomes have not found such nucleotide blocks. There is only minor satellite IV block inside Y chromosome short arm. The Y chromosome has got some (TG)n enriched blocks distributed also among other parts of genome and one copy of sequences like human ZFY gene. The high repetitive fraction of bovine genomic DNA have not revealed RFLP. However, the difference has been found by blot hybridization between genomic organization of satellite IV in cattle and yak chromosomal DNA. Non-Mendelian distribution of some such nucleotide blocks has been obtained for interspecies crosses of cattle and yak.     

Missing Observations in Designed Experiments Revisited

Missing data in designed experiments always cause trouble. To overcome the difficulties, various methods have been developed within the last sixty years. In this note, interrelations between some approaches to the problem of missing observations are reconsidered in view of the etimability criteria.     

Ectomycorrhizal sporophore distributions in a southeastern Appalachian mixed hardwood/conifer forest with thickets of Rhododendron maximum

Sporophore abundance of putatively ectomycorrhizal fungi was compared in a mature mixed hardwood/conifer forest inside of (+) versus outside of (-) Rhododendron maximum thickets (RmT). Experimental blocks (1/4 ha) were established inside of (3) and outside of (3) RmT at the Coweeta Hydrologic Laboratory in Macon County, North Carolina, USA. Litter and organic layer substrates were removed, composited and redistributed among 90 2 × 2m plots within the blocks. Plots received either +RmT or -RmT litter, and either +RmT or -RmT organic layers, or were unmanipulated for controls. Sporophores of 67 putatively ectomycorrhizal species were collected from the blocks. Species diversity and overall community structure were similar inside of and outside of RmTs, and no grouping was detected by substrate type. Differences within the ectomycorrhizal fungus community were associated only weakly with environmental parameters, as indicated by ordination. In light of these results, recent observations of ectomycorrhizal suppression and strong shifts in the proportions of morphotypes on tree seedlings inside of RmT do not appear to be related to differences in sporophore distributions. The changes in seedling mycobiont dominance in relation to RmT and the influence this has on seedling health should be examined directly from root tips.     

The effect of substrate abundance in the vertical stratification of bromeliad epiphytes in a tropical dry forest (Mexico)

Diversity of epiphytes is associated with niche partitioning, through vertical strata and host preferences. However, abundance of substrate offered by hosts differs between vertical strata, misleading if epiphytes prefer a stratum or are randomly distributed. In a tropical dry forest of San Andres de la Cal Morelos, central Mexico, we tested the null hypothesis, that epiphytes follow the abundance of the substrate, rather than showing preference for a particular vertical stratum, and tested whether microclimatic variables, seed germination and seedling survival match with observed epiphyte distribution. Our data show that epiphytes could be randomly distributed inside some host; but in some host species, certain structures presented either a deficit or an excess of all, atmospheric, or tank epiphytes. In the hosts Bursera copallifera and Bursera glabrifolia, distribution of epiphytes was biased towards the upper strata, with a deficit of epiphytes in the lower strata. In Conzattia multiflora, Sapium macrocarpum and Ipomoea pauciflora, epiphyte distribution was biased towards the lower strata. Vertical gradation of light, seed germination and seedling survival did not generally match with epiphyte distribution and did not support the notion that the microclimatic gradient governs the vertical distribution of epiphytes. Our data indicate that vertical distribution of epiphytes in such tropical dry forests is mainly driven by the distribution of the structures, which apparently influence dispersion of the seeds and by the lifespan of branches, which allow the concentration of epiphytes in the stratum that optimizes seed capture and the clonal growth of epiphytes.     

Likelihood-based experimental design for estimation of ED50

In selecting the best dosage choice for the estimation of ED50, it is natural to try to minimize the length of the confidence intervals. In this presentation, the dose allocation that minimizes the length of the likelihood-based confidence intervals is presented and compared with alternative allocations that have been proposed based on the length of different types of confidence intervals, such as those based on the asymptotic variance or on Fieller's Theorem. Effective strategies to deal with the parameter dependence of these allocations are explored. A series of experiments to evaluate the effect of small doses per fraction on the radiation tolerance of the rat cervical spinal cord provide the motivation and an illustration for the proposed procedures.     

Fourier transform infrared microscopy and imaging: detection of fungi in wood

FTIR microscopy was used to detect and discriminate the two wood decaying fungi Trametes versicolor and Schizophyllum commune in experimentally infected beech wood blocks. The distribution of fungal mycelium in wood was locally resolved and semiquantitatively recorded using FTIR microscopy combined with a focal plane array detector and image analysis. Cluster analysis revealed major differences between FTIR spectra recorded from wood fibers and empty vessel lumina and spectra from mycelium of both fungal species, irrespective of whether the fungi were grown on the surface of wood or inside vessel lumina. Species-specific clustering of spectra of fungal mycelium grown on the wood surface and inside vessel lumina demonstrated the potential of FTIR microscopy to discriminate among fungal species decaying wood.     

Levene tests of homogeneity of variance for general block and treatment designs

This article develops a weighted least squares version of Levene's test of homogeneity of variance for a general design, available both for univariate and multivariate situations. When the design is balanced, the univariate and two common multivariate test statistics turn out to be proportional to the corresponding ordinary least squares test statistics obtained from an analysis of variance of the absolute values of the standardized mean-based residuals from the original analysis of the data. The constant of proportionality is simply a design-dependent multiplier (which does not necessarily tend to unity). Explicit results are presented for randomized block and Latin square designs and are illustrated for factorial treatment designs and split-plot experiments. The distribution of the univariate test statistic is close to a standard F-distribution, although it can be slightly underdispersed. For a complex design, the test assesses homogeneity of variance across blocks, treatments, or treatment factors and offers an objective interpretation of residual plots.     

Uniform magnetic fields and double-wrapped coil systems: improved techniques for the design of bioelectromagnetic experiments.

A common mistake in biomagnetic experimentation is the assumption that Helmholtz coils provide uniform magnetic fields; this is true only for a limited volume at their center. Substantial improvements on this design have been made during the past 140 years with systems of three, four, and five coils. Numerical comparisons of the field uniformity generated by these designs are made here, along with a table of construction details and recommendations for their use in experiments in which large volumes of uniform intensity magnetic exposures are needed. Double-wrapping, or systems of bifilar windings, can also help control for the non-magnetic effects of the electric coils used in many experiments. In this design, each coil is wrapped in parallel with two separate, adjacent strands of copper wire, rather than the single strand used normally. If currents are flowing in antiparallel directions, the magnetic fields generated by each strand will cancel and yield virtually no external magnetic field, whereas parallel currents will yield an external field. Both cases will produce similar non-magnetic effects of ohmic heating, and simple measures can reduce the small vibration and electric field differences. Control experiments can then be designed such that the only major difference between treated and untreated groups is the presence or absence of the magnetic field. Double-wrapped coils also facilitate the use of truly double-blind protocol, as the same apparatus can be used either for experimental or control groups.     

Some Third-Order Rotatable Designs

Some new third-order rotatable designs are obtained in 5, 6, 7 and 8 dimensions. Many of these designs are of a sequential type and some require fewer experiments than the known designs with minimal number of experiments.     

Effects of Adult Dreissena polymorpha on Settling Juveniles and Associated Macroinvertebrates

The impact of Dreissena polymorpha settlement on recruitment of juvenile mussels and density of other macroinvertebrates was studied in field experiments using blank concrete blocks and tiles (control), blocks and tiles with attached empty zebra mussel shells, and blocks and tiles with attached living mussels. On blocks, dominant invertebrate taxa showed colonization patterns coinciding with increased habitat complexity owing to zebra mussel settlement or the biodeposition of faeces and pseudofaeces. Adult and especially juvenile zebra mussels preferred blocks with empty shells to blank blocks and blocks with living mussels; this might possibly be caused by a chemical cue that induces gregarious settlement. Lower recruitment on blocks with attached living mussels compared to blocks with only shells could be the consequence of ingestion of larvae by adult mussels and of competition for food. On tiles, the sediments deposited and the organic content of the sediment were investigated. Sedimentation was significantly higher on shell‐only and live‐mussel tiles compared to blank tiles. Organic matter differed significantly between blank and live‐mussel tiles.     

Experiments with Three-Treatment Three-Period Crossover Design: Analysis Through the General Linear Model

In this paper we consider balanced three-treatment three-period crossover designs. Using a strategy similar to that applied in the analysis of split-plot experiments we describe both the within and the between sample units models, as well as the corresponding Analysis of Variance. We illustrate these procedures with a numerical example and discuss their implementation through computer programs designed for the analysis of the general linear model.